Data Mining and Validation of AMPK Pathway as a Novel Candidate Role Affecting Intramuscular Fat Content in Pigs

Simple Summary Intramuscular fat (IMF) is increasingly being recognized as a key meat trait in the modern pork industry. The aims of this research were to identify potential signaling pathways associated with IMF content in the longissimus dorsi (LD) muscle of different pig breeds and investigate the gene expression levels in the screened signaling pathways. Our results indicated that the AMPK signaling pathway may be related to IMF deposition in the LD muscle of pigs. The results of qRT-PCR analysis showed that the expression of ten key hub genes (AMPK, ADIPOR1, ADIPOR2, LKB1, CAMKKβ, CPT1A, CPT1B, PGC-1α, CD36, and ACC1) differed between the LD muscle of Min and Large White pigs. The protein expression levels of AMPK, LKB1, CaMKK2, CPT1A, and ACC1 were similar to the genes expression patterns in the LD muscle of Large White pigs. The results of this study provide novel insights into the regulatory function of the AMPK signaling pathway in relation to the IMF content in the LD muscle of different pigs. Abstract Intramuscular fat (IMF) is an important economic trait for pork quality and a complex quantitative trait regulated by multiple genes. The objective of this work was to investigate the novel transcriptional effects of a multigene pathway on IMF deposition in the longissimus dorsi (LD) muscles of pigs. Potential signaling pathways were screened by mining data from three gene expression profiles in the Gene Expression Omnibus (GEO) database. We designed quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) arrays for the candidate signaling pathways to verify the results in the LD muscles of two pig breeds with different IMF contents (Large White and Min). Western blot analysis was used to detect the expression levels of several candidate proteins. Our results showed that the AMPK signaling pathway was screened via bioinformatics analysis. Ten key hub genes of this signaling pathway (AMPK, ADIPOR1, ADIPOR2, LKB1, CAMKKβ, CPT1A, CPT1B, PGC-1α, CD36, and ACC1) were differentially expressed between the Large White and Min pigs. Western blot analysis further confirmed that LKB1/CaMKK2-AMPK-ACC1-CPT1A axis dominates the activity of AMPK signaling pathway. Statistical analyses revealed that AMPK signaling pathway activity clearly varied among the two pig breeds. Based on these results, we concluded that the activation of the AMPK signaling pathway plays a positive role in reducing IMF deposition in pigs.


Introduction
As one of the most important domesticated animals for agricultural production, pigs provide many meat products for humans [1]. In modern society, pork quality has had an increasing influence on consumer acceptance and initial purchasing decisions. Consumers are interested in several major pork quality traits, including meat color, pH value, water holding capacity, and intramuscular fat (IMF) content, which are becoming increasingly important from an economic perspective [2,3]. Skeletal muscle is a heterogeneous tissue that consists of different types of myofibers, connective tissue, vascular tissue, nervous tissue, and IMF [4]. IMF is a major meat quality trait in pigs, and its content is directly associated with the sensory qualities, flavor, juiciness, tenderness, and nutritional quality of pork [5,6]. In recent decades, several studies have focused on the relationship between IMF and pork quality [7][8][9][10].
Famous lean pig breeds, such as Large White, Landrace, and Duroc, have lower IMF contents and reduced meat quality due to the intensive selection processes used to improve pork productivity. However, many excellent indigenous breeds are distributed in China, such as the Jinhua, Laiwu, Meishan, and Min, and they have higher IMF contents and better meat quality than the lean breeds [11][12][13][14][15][16]. Thus, it will be beneficial to reveal the molecular mechanisms of IMF deposition by comparing gene expression between lean and indigenous Chinese pig breeds.
With the rapid development of microarray and RNA-seq technologies in the last few decades, researchers are now able to study many differentially expressed genes (DEGs) simultaneously in a given tissue. To date, many studies concerning meat quality traits and gene expression in pigs have been reported [17][18][19]. Fortunately, the relevant datasets have been deposited and stored in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database and are freely accessible to researchers worldwide. However, few studies have focused on integrating and reanalyzing these datasets, which contain valuable clues regarding important porcine economic traits. Thus, by integrating and reanalyzing these datasets, we can provide significant insights into the molecular changes associated with IMF deposition.
In this study, we integrated and reanalyzed three original expression profiles from the GEO database based on a current popular differential gene expression analysis method. We found that the AMPK pathway plays a critical role in IMF deposition. We further validated this pathway through the use of quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) arrays in the Large White and Min pig breeds.

GEO Data Collection
The gene expression profiles GSE24192, GSE75045, and GSE99092 [12,20,21] were downloaded from the GEO database. The GSE24192 dataset contained six samples, which included three Large White longissimus dorsi (LD) samples and three Northeastern Indigenous (Min) LD samples. The GSE75045 dataset contained six samples, which included three Large White LD samples and three Wannanhua LD samples. The GSE99092 dataset contained six samples, which included three Large White LD samples and three Wei LD samples. In the present study, Large White was set as the experimental group, the indigenous Chinese pig breeds were set as the control group.

Identification of DEGs
The limma package from Bioconductor and the online tool iDEP (https://github.com/gexijin/ iDEP) were used to identify DEGs for the selected gene expression profile datasets [22]. A p-value less than 0.05 and |log Fold Change(FC) | ≥1 were regarded as the cutoff thresholds for DEGs.

Signaling Pathway Enrichment Analysis of DEGs
To analyze the functions of the DEGs, we performed a Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of the DEGs using the online tool Kobas 3.0 [23]. A p-value less than 0.05 was considered statistically significant. ClusterProfiler was used for the statistical analysis and visualization of the functional profiles of the DEGs in the GEO datasets and qRT-PCR arrays [24]. TBtools was used to construct the Venn diagrams of the KEGG pathways for the DEGs in the three GEO datasets.

Animals and Tissue Collection
Three sows each from the Min and Large White breeds were used in this study. The Min pig is an excellent indigenous breed from northeastern China, and it has an IMF content higher than that of Large White pigs [12]. The pigs used in this study were obtained from the Institute of Animal Husbandry Research, Heilongjiang Academy of Agricultural Sciences (Harbin, China). The pigs were raised for 180 days under the same conditions. When the pigs were slaughtered, the LD muscle was collected between the 10th and 12th ribs from the carcasses. All tissue samples were divided into two groups; one group was quickly frozen immediately after collection and stored at −80 • C until use in qRT-PCR arrays, and the other group was stored at 4 • C for the determination of IMF content. All animal procedures were performed according to the University Committee on the Use and Care of Animals at Jilin University (approval ID: 201706030).

Determination of IMF Content
In the present study, the IMF content was measured in each LD sample by the Soxhlet extraction method with petroleum ether [25].

RNA Extraction and Quantitative Real-Time Reverse Transcription-Polymerase Chain Reaction (Qrt-PCR) Arrays
Total RNA from the LD was isolated from approximately 200 mg of frozen tissue using TRIzol-A + (TIANGEN, Beijing, China) following the manufacturer's instructions. The BioRT cDNA First Strand Synthesis Kit (Bioer Technology, Hangzhou, China) was used to synthesize first-strand cDNA. Subsequently, the expression levels of the target genes were analyzed on an iQ™5 real-time PCR detection system (Bio-Rad, USA). A BioEasy SYBR Green I Real Time PCR kit (Bioer Technology) was used according to the manufacturer's instructions to detect each sample in triplicate. The primers used for the qRT-PCR arrays are listed in Table S1. The gene IDs from the selected pathways were obtained from the KEGG database (Table S2).

Statistical Analysis of the Qrt-PCR Array Results
In the present study, the Large White pigs were set as the experimental group and the Min pigs were set as the control group. GraphPad Prism 6.01 (GraphPad Software, San Diego, CA, USA) was used for analyzing our results. Student's t-tests were used to compare the control and experimental groups. For all comparisons, * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001 were considered significant differences. The 2 −∆CT method was used to calculate the Ct values from the qRT-PCR array data. The limma package was used to identify the significant DEGs of the qRT-PCR arrays between different groups [22]. TBtools was used to draw heatmaps of the qRT-PCR arrays (https://github.com/CJ-Chen/TBtools). The online tool KEGG Mapper was used to draw the colored map of DEGs (https://www.kegg.jp/kegg/mapper.html). The STRING database was used to predict protein interactions and construct the network for DEGs [26]. The protein-protein interaction (PPI) network was visualized by Cytoscape [27].

Identification of DEGs in GEO Datasets
According to the cutoff threshold (p < 0.05 and |log FC | ≥ 1), in GSE24192, 1237 DEGs were identified in the LD of Large White pigs when compared with the indigenous Chinese breeds, and they included 877 upregulated genes and 360 downregulated genes. In GSE75045, a total of 2582 DEGs were identified in the LD of Large White pigs, and they included 1096 upregulated genes and 1486 downregulated genes. Finally, in GSE99092, a total of 1822 DEGs were identified in the LD of Large White pigs, and they included 809 upregulated genes and 1013 downregulated genes.

Identification of DEGs in GEO Datasets
According to the cutoff threshold (p < 0.05 and |logFC| ≥ 1), in GSE24192, 1237 DEGs were identified in the LD of Large White pigs when compared with the indigenous Chinese breeds, and they included 877 upregulated genes and 360 downregulated genes. In GSE75045, a total of 2582 DEGs were identified in the LD of Large White pigs, and they included 1096 upregulated genes and 1486 downregulated genes. Finally, in GSE99092, a total of 1822 DEGs were identified in the LD of Large White pigs, and they included 809 upregulated genes and 1013 downregulated genes.

IMF Content of LD Muscles in the Two Pigs
As shown in Figure 2, the Min pigs had significantly higher IMF content than the Large White pigs. Valine, leucine, and isoleucine degradation 5 4.77 × 10 −2 ssc00330 Arginine and proline metabolism 5 4.77 × 10 −2

IMF content of LD Muscles in the Two Pigs
As shown in Figure 2, the Min pigs had significantly higher IMF content than the Large White pigs.

Validation of the AMPK Signaling Pathway in the LD Muscles of the Two Pig Breeds
The AMPK signaling pathway (ssc04152) consists of 117 genes (Table S2). In this study, 114 of these genes were validated via qRT-PCR array. Of them, 40 genes were differentially expressed in the Large White LD; 22 were upregulated and 18 were downregulated. The qRT-PCR results for the AMPK signaling pathway are shown in Figures 3-5, Tables 4 and 5. A heatmap of the AMPK signaling pathway is presented in Figure 3A, indicating that the expression patterns of the DEGs in the AMPK signaling pathway among the Large White and Min breeds show significant differences. Figure 3B shows a colored map of the AMPK signaling pathway in the LD of Large White pigs. Figure  4 shows a PPI network of 14 upregulated genes in the LD muscle of Large White pigs, and it contains 14 nodes and 36 edges. The most significant 10 node degree genes are SLC2A4, PGC-1α, LKB1, AMPK, LIPE, CD36, CaMKK2, FOXO3, FOXO1, and CPT1B (Table 5). Moreover, based on the results of the colored map and PPI network, 10 genes in the AMPK signaling pathway (AMPK, ADIPOR1, ADIPOR2, LKB1, CaMKKβ, CPT1A, CPT1B, PGC-1α, CD36, and ACC1) were selected as key hub genes, and their expression patterns are presented in Figure 5. These hub genes are associated with fatty acid oxidation, lipid oxidation, and fatty acid metabolic processes. In addition, the protein expression levels of AMPK, LKB1, CaMKK2, CPT1A, and ACC1 were examined in the LD muscle of the Min and Large White pigs by Western blot analysis. The protein expression levels of AMPK, LKB1, CaMKK2, and CPT1A were higher in the Large White group than that the Min group, ACC1 showed low expression level in the LD muscle of Large White pigs ( Figure 6). Taken together, these

Validation of the AMPK Signaling Pathway in the LD Muscles of the Two Pig Breeds
The AMPK signaling pathway (ssc04152) consists of 117 genes (Table S2). In this study, 114 of these genes were validated via qRT-PCR array. Of them, 40 genes were differentially expressed in the Large White LD; 22 were upregulated and 18 were downregulated. The qRT-PCR results for the AMPK signaling pathway are shown in Figures 3-5, Tables 4 and 5. A heatmap of the AMPK signaling pathway is presented in Figure 3A, indicating that the expression patterns of the DEGs in the AMPK signaling pathway among the Large White and Min breeds show significant differences. Figure 3B shows a colored map of the AMPK signaling pathway in the LD of Large White pigs. Figure 4 shows a PPI network of 14 upregulated genes in the LD muscle of Large White pigs, and it contains 14 nodes and 36 edges. The most significant 10 node degree genes are SLC2A4, PGC-1α, LKB1, AMPK, LIPE, CD36, CaMKK2, FOXO3, FOXO1, and CPT1B (Table 5). Moreover, based on the results of the colored map and PPI network, 10 genes in the AMPK signaling pathway (AMPK, ADIPOR1, ADIPOR2, LKB1, CaMKKβ, CPT1A, CPT1B, PGC-1α, CD36, and ACC1) were selected as key hub genes, and their expression patterns are presented in Figure 5. These hub genes are associated with fatty acid oxidation, lipid oxidation, and fatty acid metabolic processes. In addition, the protein expression levels of AMPK, LKB1, CaMKK2, CPT1A, and ACC1 were examined in the LD muscle of the Min and Large White pigs by Western blot analysis. The protein expression levels of AMPK, LKB1, CaMKK2, and CPT1A were higher in the Large White group than that the Min group, ACC1 showed low expression level in the LD muscle of Large White pigs ( Figure 6). Taken together, these results show that the AMPK signaling pathway is more active in the Large White breed than the Min breeds.          Figure 5. Relative expression levels of ten key hub genes in the AMPK signaling pathway in two pig breeds. All data are shown as the means ± SEM (n = 3), *** p < 0.001.  Table 5. Summary for the PPI network of 14 upregulated genes in the AMPK signaling pathway.

Gene Symbol Degree
Top ten degree genes are shown in bold.
Animals 2019, 9, x 9 of 16 Figure 5. Relative expression levels of ten key hub genes in the AMPK signaling pathway in two pig breeds. All data are shown as the means ± SEM (n = 3), *** p < 0.001.

GO Enrichment of DEGs in Qrt-PCR Arrays
The biological processes encoded by upregulated genes were involved in fatty acid oxidation, lipid oxidation, and fatty acid metabolic processes, while the biological processes encoded by the downregulated genes were involved in carbohydrate metabolic processes, including glucose, hexose, and monosaccharide metabolism, as well as hexose and monosaccharide catabolism (Figure 7). These results indicate that compared with the LD of Min pigs, the LD of Large White pigs consumes more fat for energy metabolism than carbohydrates.

Discussion
In the modern pork industry, the IMF content is an important trait that is positively associated with pork quality and in demand by consumers. As a complex meat trait, IMF deposition in the LD muscle is regulated by multiple genes and pathways. In this study, by integrating and reanalyzing three gene expression profiles, we compared the pathways related to IMF deposition in the LD muscle of Large White pigs with those of indigenous breeds. Several candidate signaling pathways were found, and the expression patterns of genes in the AMPK pathway in pigs were validated by qRT-PCR arrays in subsequent experiments.
The AMPK signaling pathway plays critical roles in controlling both glucose and lipid metabolism. AMPK is the central gene of the AMPK signaling pathway and a heterotrimeric enzyme with α, β, and γ subunits. Once activated, AMPK promotes lipid oxidation and glucose uptake, inhibits lipid synthesis and decreases IMF contents [28]. Accordingly, the activity of AMPK is inversely correlated with IMF accumulation. In the present study, AMPK was highly expressed in the LD muscle of Large White pigs, which has a lower IMF content than Min pigs. This result is consistent with several previous reports demonstrating that the expression levels of AMPK are higher in the low-IMF-content skeletal muscle of cattle and sheep [29][30][31], suggesting that AMPK plays a positive role in reducing the IMF content in pigs.
As a member of the adipocytokines, adiponectin plays crucial roles in whole-body energy homeostasis by stimulating AMPK. ADIPOR1 and ADIPOR2 are two major receptors for adiponectin and play key roles in metabolic pathways that regulate glucose and lipid metabolism, inflammation, and oxidative stress [32]. ADIPOR1 and ADIPOR2 mediate the metabolic actions of adiponectin by activating AMPK and PPARα, respectively. This activation leads to increased fatty acid oxidation and glucose uptake in mice [33][34][35]. Moreover, muscle-specific disruption of ADIPOR1 inhibits the increase in intracellular Ca 2+ concentration and reduces the activation of calmodulin-dependent kinase β (CaMKKβ) and AMPK by adiponectin. Consistent with these previously reported results, ADIPOR1 and ADIPOR2 were both upregulated in the LD muscle of Large White pigs in the present study, suggesting that these two genes may account for the low IMF accumulation in pigs. Interestingly, two upstream kinases of AMPK, the tumor suppressor LKB1 and Ca 2+ /CaMKKβ, which participate in the phosphorylation and activation of AMPK, were simultaneously highly expressed in the LD muscle of Large White pigs. AMPK can be activated via two distinct mechanisms, LKB1 encodes a serine-threonine kinase that directly phosphorylates and activates AMPK, and CaMKKβ can form a complex with and activate AMPK through their kinase domains in skeletal muscle [36][37][38][39]. Consistent with these previous findings, our results suggest that LKB1 and CaMKKβ play critical roles in reducing the IMF content in pigs by activating the expression of AMPK.
Furthermore, AMPK plays a central role in controlling lipid metabolism by regulating the downstream acetyl-CoA carboxylase (ACC1) and carnitine palmitoyltransferase 1 (CPT1) pathways. CPT1, a rate-limiting enzyme of mitochondrial fatty acid β-oxidation, is closely associated with fat deposition. Additionally, CPT1A and CPT1B, two common CPT1 subtypes in mammals, play prominent roles in fatty acid oxidation and lipid accumulation in humans, chickens, and pigs [40][41][42]. Moreover, according to the KEGG database, CPT1A and CPT1B are involved in the AMPK signaling pathway, implying that these genes participate in the mediation of fatty acid oxidation. In addition, the ACC1 gene encodes acetyl-CoA carboxylase (ACC), which is the rate-limiting enzyme responsible for the de novo synthesis of fatty acids [43]. As a target gene of the AMPK signaling pathway, the activity of ACC1 is inhibited by AMPK. Similarly, an increase in activity of ACC1 can also inhibit the expression of CPT1 and fatty acid oxidation in skeletal muscles [44][45][46]. Consistent with these previous results, both CPT1A and CPT1B were highly expressed and ACC1 was significantly downregulated in the LD muscle of Large White pigs, suggesting that the AMPK-ACC1-CPT1 pathway was positively associated with a decrease in fatty acid synthesis and IMF deposition in pigs.
In addition to CPT1, the fatty acid transporter fatty acid translocase/cluster of differentiation 36 (FAT/CD36) has also been found to regulate FA oxidation in skeletal muscle of human and mice [47,48]. FAT/CD36 has been identified as contributing to fatty acid transport and oxidation in mice [49][50][51]. In our results, FAT/CD36 was highly expressed in the LD muscle of Large White pigs, implying that the expression of this gene has a negative effect on IMF deposition in Large White pigs. Interestingly, the peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PPARGC1A, PGC-1α) was also involved in the activation of the CPT1 gene [52]. PGC-1α plays an important role in glucose and fatty acid metabolism and has a negative relationship with the IMF content in pigs [53][54][55]. In this study, PGC-1α presented a high expression level in the LD of Large White, suggesting that this gene exerts a negative effect on IMF deposition in pigs. Additionally, the protein expression of AMPK, LKB1, CaMKK2, CPT1A, and ACC1 were further analyzed with Western blot. In our results, four proteins (AMPK, LKB1, CaMKK2, and CPT1A) showed higher expression in the LD muscle of Large White pigs than that in Min pigs, and ACC1 had a low protein expression level in the Large White group, indicating that the AMPK signaling pathway is more active at the protein level and that the oxidative degradation of fatty acids is stronger in the Large White group. In summary, the above results suggest that the activation of the AMPK signaling pathway reduces the IMF content in the LD muscle of Large White pigs.

Conclusions
In conclusion, this study illustrates that the accumulation of IMF in Large White pigs is related to activation of the AMPK signaling pathway. The relatively high expression of genes in the AMPK pathway may represent one of the more significant features of pigs with artificially lean meat. Our results are also helpful for interpreting the different molecular mechanisms of IMF deposition between lean and fat pig breeds.
Author Contributions: H.Y. and H.O. conceived and designed the experiments; C.Y. and D.P. performed the experiments; C.L., A.X. and P.H. collected the samples; and C.Y. and D.P. wrote the paper.