A high-fat diet increases body fat mass and up-regulates expression of genes related to adipogenesis and inflammation in a genetically lean pig^{<a href="#FN01" rid="FN01" class=" fn">*</a>} ^{<a href="#FN02" rid="FN02" class=" fn">#</a>}

1. Introduction

A high-fat (HF) diet is accepted as a critical factor leading to the obesity epidemic (Schrauwen and Westerterp, 2000; Bray et al., 2004). Considerable interest has been shown in using pigs as models for human obesity research because of the physiological and pathophysiological similarities between pigs and humans (Rocha and Plastow, 2006). Most pig models are minipigs that are genetically predisposed to obesity and associated disorders (Christoffersen et al., 2013; Toedebusch et al., 2014). However, little information exists regarding the relevant characteristics of a polygenic model of diet-induced obesity in pigs. Domestic pigs, especially the fast-growing lean strains that have undergone intense selection for production traits, may provide a useful alternative for the investigation of the pathogenesis of diet-induced obesity (Rocha and Plastow, 2006). In the present study, we investigated whether an HF diet could induce domestic lean pigs into obesity and thus body fat accumulation. We determined the transcriptome of subcutaneous adipose tissue and liver tissue in this polygenic model of diet-induced obesity.

2. Materials and methods

Experimental procedures for this study were approved by the Animal Experimental Committee of the Institute of Animal Science, Guangdong Academy of Agricultural Sciences, China.

2.1. Animals and diets

One hundred and twelve castrated male pigs (Duroc-Landrace-Yorkshire) of similar age ((55±3) d) and body weight (BW, (18.06±0.22) kg) were randomly assigned to either a control (4% fat, 13.38 MJ/kg feed) or HF (11.4% fat, 15.55 MJ/kg feed) diet for three months. Animals undergoing each treatment were maintained in seven replicate pens, each pen housing eight pigs. The concentrations of proteins, essential amino acids, and other nutrients contained in both diets were adjusted to meet the required National Research Council (NRC; 1998) minimal levels (Table ​(Table1)1) for 20 to 60 kg pigs.

2.2. Experimental design and tissue sampling

The pigs were provided with unlimited access to their respective experimental diets and water. Feed intake on a per pen basis and individual BW were recorded weekly. One week before the end of the experiment, feed was withheld for 12 h before heparinized blood was collected from the anterior vena cava using vacuum tubes (with anticoagulant). The blood was then centrifuged for 10 min at 3000g at 4 °C, and obtained plasma was stored at −80 °C until further analysis.

At the end of the experiment, one pig with BW closest to the average BW of each pen was killed for sample collection (n=7 for both groups). Pigs were electrically stunned, exsanguinated, and processed following standard commercial procedures. The eviscerated carcass was split longitudinally through the midline, and the left side was weighed and dissected into lean mass, major white fat mass, bone, and skin.

These tissues were weighed and calculated as proportions of the left half-carcass. Samples of the dorsal white adipose tissue (WAT) and longissimus muscle between the 9th and 10th ribs were obtained from the right half-carcass and immediately stored in 10% formaldehyde solution for subsequent histological analysis. Samples of the dorsal WAT and liver tissue were also snap-frozen in liquid nitrogen and stored at −80 °C, pending RNA isolation.

2.3. Tissue histology

The histological examination of the dorsal WAT and muscle samples was conducted following procedures described previously (Ma et al., 2015), and the diameter, volume, and density of adipocytes and muscle fibers were determined.

2.4. Plasma biochemical analysis

The plasma concentrations of total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and glucose were measured with an automatic biochemical analyzer (Beckman CX5, Beckman Coulter Inc., Brea, CA, USA) using Beckman assay kits (TC#467825, TG#445850, HDL-C#650207, LDL-C#969706, and Glucose#442640, respectively). The plasma concentrations of insulin (H203) and leptin (H174) were measured using commercial radioimmunoassay kits (NBI, Beijing, China).

2.5. Real-time PCR validation

Total RNA was isolated from dorsal adipose and liver tissue samples using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Total RNA (1 μg) was synthesized into complementary DNAs (cDNAs) using a reverse transcription kit (TaKaRa, Tokyo, Japan). The synthesized cDNA was diluted (1:10, v/v) and real-time quantitative polymerase chain reaction (qPCR) amplification was determined with SYBR green I (TaKaRa) and specific primers for porcine messenger RNA (mRNA) sequences (Table ​(Table2).2). Conditions for qPCR were an initial denaturation at 95 °C for 180 s, followed by 40 cycles at 95 °C for 15 s and 58 °C for 30 s, with a final elongation at 72 °C for 30 s. The expression of the target genes relative to the β-actin was evaluated by the 2^{C method: Δ}C_T=C_{T (target gene)}−C_{T (β-actin)} and ΔΔC_T=ΔC_{T (HF diet pigs)}−ΔC_{T (control pigs)}.

2.6. Porcine microarray analysis

To explore the molecular basis for the effects of an HF diet on adiposity and obesity, cDNA microarray analysis was used to profile transcripts in dorsal WAT. GeneChip Porcine Genome Arrays (Affymetrix, Santa Clara, CA, USA), containing 23 937 probe sets that interrogated approximately 23 256 transcripts from twenty 201 Sus scrofa genes, were used. Total RNA was extracted from the dorsal WAT of individual pigs using TRIzol reagent (Invitrogen). The integrity and concentration of RNA were evaluated with Agilent 2100 Bioanalyzer (Agilent, Palo Alto, CA, USA), and the RNA was further purified using an RNA cleanup kit (Qiagen, Dusseldorf, Germany). Equal amounts of total RNA from the seven individual pigs per treatment were pooled. Approximately 6 μg of RNA from each treatment group was amplified and labeled using the Affymetrix 3' IVT labeling kit (Lot No. 900228) and then hybridized with the GeneChip arrays for 16 h. After hybridization, the arrays were washed and stained using an Affymetrix fluidics station 450 and then scanned using an Affymetrix 3000 7G Plus scanner. The genes that met the criteria of a false discover rate value of ≤0.05 and an average fold change of at least 1.5 in either direction were considered to be significant. The GenMAPP-CS software package (http://www.genmapp.org) was used for gene ontology (GO) and pathway analyses.

2.7. Statistical analysis

Parameters including Insig-1, Insig-2, HMG-CoA reductase (HMGCR), sterol regulatory element binding protein 1c (SREBP-1c), fatty acid synthase 1 (FASN1), hepatocyte nuclear factor-4α (HNF-4α), diacylglycerol O-acyltransferase 1 (DGAT1), DGAT2, fatty acid binding protein 4 (FABP4), carnitine palmitoyl transferase 1 (CPT1), CPT2, proliferator-activated receptor (PPAR)-γ, and PPAR-α were performed using GraphPad Prism Version 5 (GraphPad software), and data were presented as mean±standard error of mean (SEM). Significance was evaluated using a two-tailed Student’s t-test. A value of P<0.05 was considered a statistically significant difference. Data (BW at baseline, BW at the end of experiment, average daily feed intake, average daily energy intake, percentage fat mass, adipocyte diameter, adipocyte volume, adipocyte density, percentage lean mass, muscle fiber diameter, muscle fiber density, TC, TG, HDL-C, LDL-C, glucose, insulin, and leptin) were analyzed using the general linear model (GLM) procedure of SAS 9.2 (SAS Institute, Cary, NC, USA), and presented as mean±SEM. P value of <0.05 was considered significant.

3. Results

3.2. Body composition and histological analyses of the dorsal WAT

The lean mass, dorsal muscle fiber diameter, and density were lower in the control group than in the HF-fed pigs, although with no significant statistics (Table ​(Table4,4, Figs. 1c and 1d). By contrast, the percentage of fat mass in the HF pigs was 37.2% higher than that in the control animals (P<0.05). The histological characteristics of the dorsal WAT are shown in Table ​Table44 and Figs. 1a and 1b. The diameter (P<0.05) and volume (P<0.05) of adipocytes in the dorsal WAT were significantly increased in the HF pigs compared with the control pigs. The density of adipocytes (number per unit sectioned area) in WAT was 64.6% higher in the HF pigs than in the control pigs (P<0.05).

Fig. 1

Hematoxylin and eosin staining of dorsal white adipose tissue

and dorsal muscle tissue Representative images of dorsal white adipose tissue from control pigs (a) and high-fat diet-induced obese pigs (b); dorsal muscle tissue from control pigs (c) and high-fat diet-induced obese pigs (d)

Table 4

Body composition and dorsal white adipose tissue histological results

Group	Percentage fat mass (%)	Adipocyte diameter (μm)		Adipocyte volume (×10 μm)		Adipocyte density (amount/g)
Control (n=7)	12.13±0.93	94.91±1.61		0.45±0.02		2.46±0.12
High fat (n=7)	16.64±0.18*	110.67±4.66*		0.72±0.09*		1.59±0.20*
Group	Percentage lean mass (%)		Muscle fiber diameter (μm)		Muscle fiber density (amount/mm)
Control (n=7)	64.71±2.41		48.99±2.24		0.90±0.10
High fat (n=7)	61.53±1.01		46.35±2.76		1.07±0.07

Data were analyzed using the GLM procedure of SAS 9.2, and data are presented as mean±SEM.

P<0.05, vs. the control group

3.4. Effect of HF diet on adipose expression of genes responsible for cholesterol biosynthesis and lipid metabolism in adipose and liver tissues

The relative abundances of Insig-1 and Insig-2 transcripts in the backfat tissue of pigs fed the HF diet increased (3.15-and 2.5-fold, respectively; Fig. ​Fig.2a,2a, P<0.05). Of the selected genes involved in lipid metabolism, relative transcript abundances of SREBP-1c, FASN1, and DGAT2 all increased, while those of PPAR-α and CPT2 decreased; the increase in FABP4 transcripts (2.87-fold) was quite substantial. In liver tissue, HF feeding significantly elevated the expression of SREBP-1c, FASN1, DGAT2, and HNF-4α mRNA (P<0.05) and somewhat elevated mRNA expression of HMGCR and Insig-1 genes, although without a statistically significant difference between the two treatment pools; the amounts of FABP4, Insig-2, DAGT1, PPAR-α, and CPT2 in liver tissue were not affected (Fig. ​(Fig.2b2b).

Fig. 2

Relative transcript abundance of genes involved in cholesterol synthesis and lipid metabolism in backfat tissue (a) and in liver tissue (b) of control pigs and HF-diet pigs

Data were performed using GraphPad Prism Version and expressed as mean±SEM (n=7). P<0.05, P<0.01, P<0.001, vs. control. HMGCR, HMG-CoA reductase; SREBP-1c, sterol regulatory element binding protein 1c; FASN1, fatty acid synthase 1; HNF-4α, hepatocyte nuclear factor-4α; DGAT1, diacylglycerol O-acyltransferase 1; DGAT2, diacylglycerol O-acyltransferase 2; FABP4, fatty acid binding protein 4; CPT1, carnitine palmitoyl transferase 1; CPT2, carnitine palmitoyl transferase 2; PPAR-γ, proliferator-activated receptor-γ; PPAR-α, proliferator-activated receptor α; HF: high fat

3.1. Growth performance

No difference was detected in BW between the two dietary groups of pigs, either at the start (baseline) or at the end of the experiment (Table ​(Table3).3). The average daily feed intake in the HF diet group was significantly less than that in the control diet group (P<0.05), whereas energy intake did not differ between the groups.

3.2. Body composition and histological analyses of the dorsal WAT

The lean mass, dorsal muscle fiber diameter, and density were lower in the control group than in the HF-fed pigs, although with no significant statistics (Table ​(Table4,4, Figs. 1c and 1d). By contrast, the percentage of fat mass in the HF pigs was 37.2% higher than that in the control animals (P<0.05). The histological characteristics of the dorsal WAT are shown in Table ​Table44 and Figs. 1a and 1b. The diameter (P<0.05) and volume (P<0.05) of adipocytes in the dorsal WAT were significantly increased in the HF pigs compared with the control pigs. The density of adipocytes (number per unit sectioned area) in WAT was 64.6% higher in the HF pigs than in the control pigs (P<0.05).

Fig. 1

Hematoxylin and eosin staining of dorsal white adipose tissue

and dorsal muscle tissue Representative images of dorsal white adipose tissue from control pigs (a) and high-fat diet-induced obese pigs (b); dorsal muscle tissue from control pigs (c) and high-fat diet-induced obese pigs (d)

3.3. Increased levels of TC, TG, HDL-C, LDL-C, insulin, and leptin in HF-diet pigs

The plasma concentrations of TC, TG, HDL-C, LDL-C, and glucose were identical in the HF and control pigs at the beginning of experiment (data not shown), but after feeding for three months, the concentrations of TG, TC, HDL-C, and LDL-C were higher in HF pigs than in controls (all P<0.05, Table ​Table5).5). The HF pigs also showed higher plasma concentrations of insulin and leptin (both P<0.05) than controls, but there was no difference in plasma concentration of glucose (Table ​(Table55).

3.4. Effect of HF diet on adipose expression of genes responsible for cholesterol biosynthesis and lipid metabolism in adipose and liver tissues

The relative abundances of Insig-1 and Insig-2 transcripts in the backfat tissue of pigs fed the HF diet increased (3.15-and 2.5-fold, respectively; Fig. ​Fig.2a,2a, P<0.05). Of the selected genes involved in lipid metabolism, relative transcript abundances of SREBP-1c, FASN1, and DGAT2 all increased, while those of PPAR-α and CPT2 decreased; the increase in FABP4 transcripts (2.87-fold) was quite substantial. In liver tissue, HF feeding significantly elevated the expression of SREBP-1c, FASN1, DGAT2, and HNF-4α mRNA (P<0.05) and somewhat elevated mRNA expression of HMGCR and Insig-1 genes, although without a statistically significant difference between the two treatment pools; the amounts of FABP4, Insig-2, DAGT1, PPAR-α, and CPT2 in liver tissue were not affected (Fig. ​(Fig.2b2b).

Fig. 2

Relative transcript abundance of genes involved in cholesterol synthesis and lipid metabolism in backfat tissue (a) and in liver tissue (b) of control pigs and HF-diet pigs

Data were performed using GraphPad Prism Version and expressed as mean±SEM (n=7). P<0.05, P<0.01, P<0.001, vs. control. HMGCR, HMG-CoA reductase; SREBP-1c, sterol regulatory element binding protein 1c; FASN1, fatty acid synthase 1; HNF-4α, hepatocyte nuclear factor-4α; DGAT1, diacylglycerol O-acyltransferase 1; DGAT2, diacylglycerol O-acyltransferase 2; FABP4, fatty acid binding protein 4; CPT1, carnitine palmitoyl transferase 1; CPT2, carnitine palmitoyl transferase 2; PPAR-γ, proliferator-activated receptor-γ; PPAR-α, proliferator-activated receptor α; HF: high fat

3.5. Microarray analysis

A total of 576 genes differed between the two dietary pools (fold change >1.5 in either direction), of which 194 genes have detailed annotations and are accessible in the S. scrofa public database. Of these 194, 73 were up-regulated and 121 were down-regulated. These HF diet-responsive genes were classified using GO analysis into 15 categories of biological processes, including muscle contraction and development, ion transport and homeostasis, immune response, response to external stimulus, energy metabolism, and cell adhesion, etc. (Fig. S1).

Further analysis of the pathway maps showed that 51 up-regulated genes were significantly involved in 11 pathways (Table S1), including those associated with amyotrophic lateral sclerosis, amoebiasis, cytokine-cytokine receptor interaction, drug metabolism, insulin signaling, extracellular matrix (ECM)-receptor interaction, retinol metabolism, prion diseases, Chagas disease, the notch signaling pathway, and colorectal cancer. Of the down-regulated genes, 57 were significantly involved in the following 10 pathways (Table S2): glycolysis/gluconeogenesis, pyruvate metabolism, cardiac muscle contraction, dilated cardiomyopathy, hypertrophic cardiomyopathy, metabolic pathways, steroid hormone biosynthesis, steroid biosynthesis, propanoate metabolism, and bladder cancer. Among these 108 genes (51 up-regulated plus 57 down-regulated) that were significantly involved in various pathways, 20 genes were associated with adiposity (Table ​(Table6),6), including ECM remodeling (9 genes) and lipid metabolism (11 genes). Another 26 genes were involved in immune response (Table ​(Table7),7), including cytokine-cytokine receptor interaction (12 genes) and other immune responses (14 genes). Among these inflammation-related genes, 22 were up-regulated in pigs fed the HF diet. qPCR was further used to validate several selected gene expression results including CES1, CES3, COMT, SC4MOL, CCL2, and CXCL12 obtained from microarray analysis. We observed that all data had the same direction of change by both methods, except FASN1 (Table ​(Table88).

4. Discussion

HF diets can lead to increased BW and body fat mass in humans and animal models (Schrauwen and Westerterp, 2000). When domestic pigs (75% Danish Landrace×25% Yorkshire; aged 3 months) were fed a high-energy diet (containing 10% (0.1 g/ml) sugar and 10% (0.1 g/ml) soy oil) ad libitum (obese group) or at 60% intake (lean group) for 5–6 months, the BW of the obese pigs was significantly higher than that of the lean pigs (Rødgaard et al., 2013). When feed intake was equivalent between the HF and control groups of minipigs, however, no difference was observed in BW (Azorín-Ortuño et al., 2012). Because feed was provided ad libitum in the present study, the daily feed intake of pigs consuming the HF diet was lower than that of pigs fed the control diet, causing no difference in energy intake or BW between the two groups (Table ​(Table33).

Body fat mass is believed to be more accurate for detecting true adiposity than the frequently used body mass index (Rocha and Plastow, 2006). In the present study, although the increased percentage of body fat mass induced by an HF diet in domestic lean pigs was not as large as that in the minipigs (Ludvigsen et al., 2015), anatomical and histological analyses showed a significantly higher percentage of body fat mass and larger adipocytes in the HF pigs compared with controls, providing further evidence for the adiposity status of these HF pigs. No significant difference was observed in the percentage of lean mass or muscle fiber size between the animals in the two groups, suggesting an independent effect of dietary fat on the expansion of fat mass.

HF diets have been widely shown to increase levels of blood lipids in animal models (Yaqoob et al., 1995; Azorín-Ortuño et al., 2012; Choi et al., 2012). Consistent with that, the present study also found that the HF diet increased plasma concentrations of TC, TG, HDL-C, and LDL-C, indicating mild hypercholesterolemia and hypertriglyceridemia in the HF pigs. Although some reports have shown significant increases in fasting serum glucose in rats subjected to long-term (5–10 months) HF diets (Chalkley et al., 2002; Choi et al., 2012), this index did not differ here, indicating that an HF diet for 3 months may not be sufficient to alter glucose homeostasis in pigs (Kim and Park, 2010; Azorín-Ortuño et al., 2012). Our result was consistent with previous studies in both swine and rat models that an HF diet leads to plasma insulin up-regulation (Guilford et al., 2017; Liu et al., 2017; Malbert et al., 2017; Yang et al., 2017), and possibly insulin resistance. The adipose-derived hormone leptin is considered to be an adiposity signal (Munzberg, 2010), and blood leptin levels are proportional to WAT stores (Gimeno and Klaman, 2005). Plasma leptin was significantly increased (45%) with fat mass expansion in the present HF pigs, confirming the increase in adiposity.

Adipose tissue is one of the most important organs for storing energy in the form of lipids and it is a major endocrine organ producing adipokines, hormones, and cytokines (Kershaw and Flier, 2004). In pigs, it was reported that adipose tissue was the central organ for fat synthesis and deposition (O'Hea and Leveille, 1969). Fat accumulation in animals depends on the levels of TG synthesis and storage and on lipid mobilization, fatty acid oxidation, and cholesterol storage (Bernlohr et al., 2002). Genetically modified animal models have highlighted that the expression of several adipogenic and lipogenic genes including FASN1, DGAT, HNF-4α, SREBP-1c, and FABP, plays a central role in fatty acid synthesis, TG synthesis, and lipid storage (Liu et al., 2008; Qiu et al., 2017). Consistent with these results, the present study showed that the HF diet significantly increased the relative expression of FASN1, DGAT2, SREBP-1c, and FABP4 in backfat tissue and FASN1, DGAT2, SREBP-1c, and HNF-4α in the liver tissue. This study also showed that the HF diet reduced the relative expression of CPT2 and PPAR-α in backfat tissue, in accordance with Li et al. (2007), who demonstrated that these genes are related to increased fatty acid oxidation. Similarly, we found that insulin-induced genes, such as Insig-1 and Insig-2, involved in cholesterol synthesis, were up-regulated by the HF diet. Coordinated regulation of pathways involving these genes in the backfat and liver tissues of HF pigs resulted in increased serum TG and cholesterol levels (15% to 23%) and increased propensity for obesity; the content of carcass fat increased by 37%.

A microarray analysis, applied to provide a transcriptional basis for changes in the WAT of HF pigs, determined that 108 genes affected by the HF diet were significantly involved in 21 pathways. Of these 108 genes, 46 (42.6%) were associated with pathways related to adiposity development or immune response.

The expression of collagen (types 1A2, 4A5, 5A1, 5A3, and 6A1) as well as fibronectin 1, laminin 2, and secreted phosphoprotein 1, was up-regulated in the WAT of HF pigs. These genes encode proteins that participate in ECM remodeling, which is indispensable during the differentiation of fibroblast-like preadipocytes into lipid-laden adipocytes (Gregoire et al., 1998). Levels of ECM proteins (type I to VI collagens, laminin, and fibronectin) are increased during adipocyte differentiation (Nakajima et al., 2002). Others also found that expression of type IV collagen is increased during this process (Weiner et al., 1989). In obese humans, adipose expression of type VI collagen (Pasarica et al., 2009) and secreted phosphoprotein 1 (encoding osteopontin protein) (Leitner et al., 2015) is up-regulated. The results obtained in our study are consistent with those reports. Some studies, however, showed decreased expression of type I collagen and fibronectin (Bortell et al., 1994) as well as of type III collagen (Weiner et al., 1989) during adipocyte development.

Abnormalities in lipid metabolism may lead to an excessive expansion of fat mass. The expression levels of numerous genes involved in lipid metabolism are reportedly up-regulated in some animal models of obesity (Kim and Park, 2010; Choi et al., 2012). Other studies reported that expression levels of cytochrome P450 (Campión et al., 2006), fatty acid transporter protein, and glycerol-3-phosphate dehydrogenase (Soukas et al., 2000) are down-regulated in obese subjects. In the present study, all the genes significantly involved in lipid metabolism were down-regulated by the HF diet. Most of these genes, including adiponutrin, sterol-C4-methyl oxidase-like, carboxylesterase (CES1 and CES3), and 3α-hydroxysteroid dehydrogenase, are involved in lipolysis. Taken together, the increased expression of ECM genes along with the decreased expression of lipolysis genes probably accounts for the observed fat mass expansion in the HF pigs.

The over-expression of genes related to immune response is a key characteristic of WAT in obese subjects (Kim and Park, 2010). The excessive expansion of fat mass induces chronic inflammation in adipose tissue. Evidence shows that inflammatory pathways are activated in adipose tissue from obese subjects (Gimeno and Klaman, 2005; Choi et al., 2012). Consistent with these results, we found that 85% of the genes significantly involved in adipose inflammation were up-regulated in the HF pigs.

5. Conclusions

Obesity can be induced in genetically lean pigs by feeding an HF diet for three months with substantial expansion of body fat mass and transcriptional evidence of adipose-related inflammation. This genetically lean pig as a model for the study of diet-induced obesity may provide a useful and novel tool for research into metabolic diseases.

List of electronic supplementary materials