The effects of short-term high-fat feeding on exercise capacity: multi-tissue transcriptome changes by RNA sequencing analysis

Background The effects of short-term high fat diets on physiology are elusive and the molecular changes following fat overconsumption remain largely unknown. In this study, we aimed to evaluate exercise capacity in mice fed with a high fat diet (HFD) for 3 days and investigate the molecular mechanisms in the early response to high-fat feeding. Methods Exercise capacity was assessed by weight-loaded swimming test in mice fed a control diet (10 kcal% fat) or a HFD (60 kcal% fat) for 3 days. Global gene expression of ten important tissues (brain, heart, liver, spleen, lung, kidney, stomach, duodenum, skeletal muscle and blood) was analyzed using RNA Sequencing. Results A HFD for just 3 days can induce 71% decrease of exercise performance prior to substantial weight gain (P <0.01). Principle component analysis revealed that differential gene expression patterns existed in the ten tissues. Out of which, the brain, spleen and lung were demonstrated to have more pronounced transcriptional changes than other tissues. Biological process analysis for differentially expressed genes in the brain, spleen and lung showed that dysregulation of peripheral and central immune response had been implicated in the early stage of HFD exposure. Neurotransmission related genes and circulatory system process related genes were significantly down-regulated in the brain and lung, respectively. Conclusions Our findings provide new insights for the deleterious effects of high-fat feeding, especially revealing that the lung maybe as a new important target attacked by short-term high-fat feeding. Electronic supplementary material The online version of this article (doi:10.1186/s12944-017-0424-7) contains supplementary material, which is available to authorized users.


Background
High-fat and high-calorie diets along with a low physical activity lifestyle have contributed to the onset or development of type 2 diabetes, metabolic syndrome and cardiovascular disease [1]. There has been conflicting results on whether consumption of a high fat diet (HFD) is detrimental or beneficial for endurance performance. Studies in rats have demonstrated a beneficial effect of a fat-rich diet on exercise capacity via increasing the ability to oxidize fat and concomitantly sparing glycogen content [2][3][4]. In contrast, Murray et al. [5] reported that 9 days of high-fat feeding impaired energy production and physical performance associated with respiratory uncoupling in skeletal muscle mitochondria. In the present study, we aimed to evaluate exercise capacity in mice fed with a HFD for 3 days and investigate the molecular mechanisms in the early response to high-fat feeding.
Advances in genomic technologies may help to reveal the early molecular changes by enabling simultaneous analysis of thousands of genes in response to a HFD. The serial analysis of gene expression strategy identified 12 transcripts of hypothalamus which regulated by food intake in mice at 3 h after high-fat meal ingestion [6].
The transcriptomic analysis of duodenum mucosa after high-fat meal ingestion in C57BL/6 J mice found substantial changes of genes related to lipid metabolism [7]. Microarray analysis showed markedly changes of numerous genes involved in various biological processes including morphogenesis, fatty acid catabolism and amino acid metabolism following 3 days of high-fat feeding in the skeletal muscle of C57BL/6 J mice [8]. cDNA microarrays analysis of mRNA expression showed downregulation of genes related to fatty acid biosynthesis in the liver of one week HFD-fed BALB/c mice [9].
However, no study to date has simultaneously analyzed the systemic gene expression profile of multi-tissues in response to short-term HFD and it remains unknown that which tissue has the most pronounced changes of gene expression profile in the early stage after high-fat feeding. Recently, RNA sequencing (RNA-seq), as an attractive alternative to microarrays for transcriptome analysis, provides major advances in robustness, comparability and richness of expression profiling data [10]. Thus we utilized RNA-seq to investigate gene expression profile of ten tissues (brain, heart, liver, spleen, lung, kidney, stomach, duodenum, skeletal muscle and blood) in C57BL/6 J mice with 3 days of high-fat feeding, which may contribute to the understanding of molecular mechanisms of changes in exercise performance induced by short-term HFD.

Animals and study protocol
Animal experiments were approved by the Animal Care and Use Committee of Southern Medical University (Approval No.2013027). The methods were carried out in accordance with the approved guidelines. Forty male C57BL/6 J mice at the age of 8 weeks were obtained from Laboratory Animal Center of Southern Medical University (Approval No. SCXK (Yue) 2011-0015). All the animals were maintained in a temperaturecontrolled room (22-25°C; 35-55% humidity) with a twelve-hour light/dark cycle. Mice were randomly divided into two groups, where 20 mice were fed a control diet (CD, D12450B, 10 kcal% fat) and 20 mice were fed a high fat diet (HFD, D12492, 60 kcal% fat) for 3 days. Mice were allowed free access to food and water. The changes of body weight were observed after 3 days.

Assessment of exercise capacity
A weight-loaded swimming test has been commonly used for assessment of exercise capacity in murine [11,12]. After 3 days, 10 mice were taken out from each group for swimming exercise performance test which was conducted as previously described with some modifications [13]. The mice were not fasted and were loaded the constant weight (1.5 g tin wire, attached to the tail). The mice were dropped individually into a swimming pool (30 cm high, 25 cm in diameter) filled with water at 25 ± 1°C. It was considered that the mice were exhausted when they failed to return to the surface of water within a 10 s period. The swimming time to exhaustion was used as the index of exercise capacity.

Tissue Processing and RNA Isolation
Five mice of each group were selected randomly from the mice which did not perform weight-loaded swimming test for RNA sequencing. Tissue samples included the whole brain and heart, liver, spleen, lung, kidney, stomach, duodenum, skeletal muscle and blood. The samples were dissected and immediately immersed in RNA later solution (Ambion, California, USA). All samples were stored at -80°C before processing. Total RNA was extracted from all samples using Trizol reagent (Invitrogen, Carlsbad, CA). The RNA concentration was quantified using a spectrophotometer (NanoDrop echnologies, Wilmington, DE) and the integrity was evaluated by the Agilent Bioanalyzer 2100 (Agilent, Santa Clara, CA).

RNA sequencing and gene expression analysis
In the CD and HFD group, fixed quantities of RNA of five samples from the same kind of tissue were combined into a single sample. The cDNA library was conducted by Illumina Tru-Seq RNA Sample Prep Kits (Illumina, San Diego, CA) with Ribosomal RNA depletion following manufacturer's instructions. Samples were sequenced for 50 bp single read using the HiSeq2000 platform. Before alignment, reads with a low quality and adapters were screened by FastQC and removed. The remaining reads were mapped to the mice reference genome (UCSC mm10) with TopHat v2.0.9. The maximum number allowed for mismatch mapping was 2. Reads Per Kilobase of exon model per Million mapped reads (RPKM) was calculated to express the mRNA abundances. Analysis of differential expression was performed using edgeR, which could be used even with the most minimal levels of replication [14]. The read counts per gene were normalized to counts per million (CPM). CPM values were utilized for differential expression analysis, whereas RPKM values were used for principle component analysis (PCA) with the GeneSpring Gx 12.0 software (Agilent Technologies, Palo Alto, CA).

Biological process analysis
Molecule annotation system (MAS) is a set of web tools for function annotation based on integration of various public resources such as Gene Ontology, KEGG, Bio-Carta, GenMapp, UniGene, OMIM and more [15]. Biological process analysis for differentially expressed genes (DEGs) was performed using the CapitalBio MAS 3.0 software (CapitalBio Corporation, Beijing, China). Absolute fold change >2 with P < 0.05 was considered statistically significant in the RNA-seq analysis.

Real-time quantitative RT-PCR verification of RNA-seq data
To further confirm the findings from the RNA-seq analysis, we selectively examined 22 genes expression (8 genes in the brain, 9 genes in the spleen and 5 genes in the lung) using real-time quantitative RT-PCR (qRT-PCR) method. Five samples from the same kind of tissue of purified RNA in each group were used for qRT-PCR. Total RNA from the samples was first reverse-transcribed into cDNA templates with the PrimeScriptTM RT reagent Kit (TaKaRa, Otsu, Japan) according to the manufacturer's instruction. PCR was run on a ABI 7500 Real-Time PCR System (Applied Biosystems, Inc., Foster City, CA, USA) using the SYBR Premix Ex TaqTM II (Otsu-Shi, Shiga, Japan). The reaction volume was 20 μL and the PCR conditions were as follows: 30 s. at 95°C, 40 cycles of 5 s. at 95°C and 34 s. at 60°C, followed by a melting curve analysis step. Every sample was measured in duplicate, and relative quantification was determined by the comparative Ct method (2-ΔΔCT). β-actin was used as a housekeeping gene to normalize the expression data. The primers used for gene validation are listed in Additional file 1: Table S1.

Body weight and blood plasma metabolites
As shown in Table 1, initial and final body weights showed no significantly differences among the groups. HFD feeding tended to increase weight gain, but this effect failed to reach statistical significance (P = 0.094). Plasma glucose levels were 67% higher in HFD-fed mice than CD-fed mice (P < 0.001). Although no differences were seen in plasma LDL-C, FFA and ApoE, the cholesterol level was significantly greater in the HFD-fed mice (P < 0.001). The increase in HDL-C (P < 0.001) and decrease in plasma triglycerides (P = 0.005) were also observed in HFD-fed mice. Plasma levels of CRP, SOD, HCY, ALT, AST, ALP, TP, ALB, GLB, ALB/GLB, TBIL, DBIL, IBIL, TBA, UA, Cr, Urea, CysC, LDH, HBDH, K and Ca were unchanged between the CD and HFD groups except for the CK, Na and Cl (Additional file 2: Table S2).

Exercise capacity in a weight-loaded swimming test
The swimming time indicated the exercise capacity. Both groups of mice swam the same mean time at baseline (CD = 687.3 ± 93.9 s, HFD = 669.0 ± 87.8 s). After 3 days of feeding, CD-fed mice maintained a similar swimming time of 748.9 ± 77.4 s, whereas HFD-fed mice swam 213.7 ± 44.4 s on average, 71% less far than the CD-fed mice (P <0.001) (Fig. 1).

Summary of sequencing data and global gene expression profiles
A range of 28.1 to 67.4 million raw reads were generated among samples. After removing reads with a low quality, an average of 32.3 million clean reads per sample was obtained (range, 21.3 to 51.5 million reads). Approximately 98.64% of clean reads per sample were mapped to the mice reference genome among samples ( Table 2). Totally 33151 unique genes among all samples was detected. To assess the effect of sequencing depth on RNA-seq data, we conducted sequencing saturation analysis. In the beginning of the RNA-seq, with increase of the counts of reads, the number of identified genes in each tissue was increasing. However, when the counts of reads rose to approximately 30 million, the growth rate of identified genes flattened which indicated that the number of identified genes tended to saturation.
To investigate the global gene expression profiles of ten tissues and identify the tissue with the most pronounced transcriptional changes after short-term highfat feeding, we performed PCA on all samples (Fig. 2). The results showed differential gene expression patterns in the ten tissues. Each sphere represented an individual sample. The sphere representing liver, heart, kidney, skeletal muscle and blood in CD and HFD group overlapped, indicating that gene expression patterns of HFDfed mice in the five tissues were almost not changed. The sphere representing stomach and duodenum in CD and HFD group were closely to each other, indicating that gene expression patterns of CD-fed and HFD-fed mice in the two tissues were nearly similar. Gene expression patterns of brain, spleen and lung in HFD group showed substantial differences as compared with CD group.
Genes and the related biological processes altered in the brain of HFD-fed mice According to the results of PCA, we found that the brain, spleen and lung had more pronounced transcriptional changes than other tissues following 3 days of HFD intervention. Consequently, we focused on the analysis of the genes and related biological processes altered in the brain, spleen and lung of HFD-fed mice. We found 145 DEGs in the brain, of which less than half of the genes were annotated with known function from the Ensembl database (Table 3). To gain insight into the possible biologic functions of the genes affected by highfat feeding, enrichment analysis of Gene Ontology for the DEGs was conducted. After 3 days of HFD exposure, the overrepresented biological processes in the brain were mainly enriched in neurological system process and immune response (Fig. 3). In the neurological system process related group, LIM homeobox transcription factor 1 beta (Lmx1b), and NK2 homeobox 1(Nkx2-1), which involved in neuron migration and development, were down-regulated to 8.11-fold and 10.56-fold respectively in the HFD-fed mice. Genes related to central nervous system morphogenesis were significantly downregulated, such as homeobox D11 (Hoxd11) (12.21-fold) and UNC homeobox (Uncx) (5.39-fold). Inflammatory/ immune related processes were altered as well. The mRNA levels of chemokine (C-C motif ) receptor 1(Ccr1) was up-regulated to 13.18-fold. The rest immune-related genes including chemokine (C-C motif ) receptor 4 (Ccr4), CD200 receptor 3 (Cd200r3), CD274 molecule (Cd274), CD300 antigen like family member G (Cd300lg) and transcription factor AP-2, alpha (Tfap2a)  Genes and the related biological processes altered in the spleen of HFD-fed mice 61 genes were markedly changed in the spleen, however, half of which were largely unknown. The overrepresented biological processes in the spleen were mainly related to acute-phase response and immune system, with a significant change in the expression of immune-related genes (Fig. 3). As shown in Genes and the related biological processes altered in the lung of HFD-fed mice In the lung, 83 genes were significantly altered. The overrepresented biological processes were mainly enriched in immune-related processes, including T cell and B cell mediated immune response (Fig. 3). As shown in Table 5, Immunoglobulin heavy constant epsilon   Verification of RNA-seq data qRT-PCR was used to validate the expression levels measured by RNA-seq for 22 selected genes (8 genes in the brain, 9 genes in the spleen and 5 genes in the lung) from the list of differently expressed genes. As demonstrated in Fig. 4, qRT-PCR showed significant alterations in the expression of the 17 genes in correspondence with the findings from the RNA-sequencing analysis, while no obvious differential expression was detected for the 5 genes of Trap2a, Cd274, Ighv1-84, Igkv5-39 and Traj37 by qRT-PCR.  Fig. 3 Enriched biological process significantly altered in HFD-fed mice compared to CD-fed mice in the brain, spleen and lung (CD, control diet; HFD, high fat diet)

Discussion
Consistent with previous reports [6,16], our results showed that the consumption of a high-fat diet for 3 days significantly increased plasma glucose level. The high-fat diet also increased plasma cholesterol and HDL cholesterol concentrations. Plasma triglycerides concentration decreased significantly after short-term high-fat feeding. Indeed, decreased triglycerides level was previously reported as early as 3 days after beginning a high-fat diet in a study that involved mice fed a chronic high-fat diet [17]. A randomized, double-blind, crossover study in 12 healthy subjects reported that plasma triglycerides concentration was significantly lower after a 3-d high-fat  diet [18]. The low plasma triglycerides could be due to increased liver triglyceride content, possibly resulting in triglycerides being stored in the liver [19].
To our knowledge, we first reported that 3 days of high-fat feeding can induce exercise performance decrease in mice prior to substantial weight gain. In contrast to previous studies which conducted microarray analysis in a single tissue [6][7][8][9], we used RNA-seq to investigate the global gene expression profiles of ten tissues in the early response to fat intake. Interestingly, our study showed that differential gene expression patterns existed in the ten tissues. Previous researches on the effect of HFD mainly focused on the liver, skeletal muscle, intestine and heart, which were thought to have significant responses to consumption of high fat diets. However, in our results, the brain, spleen and lung were demonstrated to have more pronounced transcriptional changes than other tissues following 3 days of high-fat feeding. The simultaneous analysis of multi-tissues by RNA sequencing yielded information which had not been revealed by previous analyses of a single tissue.
In the brain of HFD-fed mice, dopamine neurons differentiation related genes including Lmx1b, Nkx2-1 and Uncx were found to be significantly down-regulated. Lmx1b, a key transcription factor for the specification of dopaminergic cell fate, has been reported to increase midbrain size and allocation of dopamine progenitors by promoting Wnt1/Wnt signaling [20]. Deficiency of Nkx2-1 in mice would lead to a remarkable abnormality in the trajectory of the ascending dopamine pathway [21]. Uncx, also known as Uncx4.1, is involved in the development of midbrain dopaminergic neurons [22]. Therefore, the down-regulation of these genes may contribute to impaired exercise capacity, which is supported by the observation of a link between dopamine and levels of physical exercise [23]. Moreover, Lmx1b and Nkx2-1 were demonstrated to regulate the migration of the superficial dorsal horn neurons and interneurons to the striatum or cortex, respectively [24,25]. The decreased expression of the two genes suggested disturbed neuronal migration in HFD-fed mice, which may result in abnormal development of the nervous system. Inflammation is considered to be one of the important factors for deterioration of physical performance [26]. Similarly, our results also suggested that many immune-related genes were significantly altered in the brain of HFD-fed mice. Ccr1, which involved in the host response to pathogens and several inflammatory conditions [27], were significantly up-regulated. On the other hand, Ccr4 and Cd200r3 showed decreased expression in HFD-fed mice. Ccr4 were found to be functionally expressed on peripheral blood CD4 + CD25 + regulatory T (Treg) cells [28]. CD200 imparts an immunoregulatory signal through the receptor for CD200, leading to the suppression of Tcell-mediated immune responses [29]. The upregulation of Ccr1, coupled with the down-regulation of Ccr4 and Cd200r3, suggested the up-regulation of an inflammatory response toward the high fat diet in the brain. Consistently, it was reported that consumption of Fig. 4 Confirmation of differential gene expression via qRT-PCR analysis (CD, control diet; HFD, high fat diet; *P < 0.05 vs. CD group). The expression value was normalized to the β-actin expression level. Values are expressed as average mRNA expression ± SEM bars a HFD with 1 to 3 days induced hypothalamic inflammatory in both rats and mice [30]. However, the genes related to hypothalamic inflammatory signaling in the above study were not found to be significantly altered in our results. This discrepancy may reflect a difference between specific brain region and the whole brain.
The immune system has been considered to be affected by HFD exposure over a period of several weeks [31]. However, in the present study, we found that 3 days of HFD feeding had induced disturbed immune response in spleen, which is a major organ involved in B-cell maturation. Fcrls, which belongs to Fc receptor-like family possessing inhibitory and/or activating signaling motifs in B cell differentiation [32], were significantly downregulated in response to HFD feeding. Similarly, microarray profiling carried out by Cui et al. in the spleen of C57BL/6 mice fed on a HFD also showed a decreased expression of Fc receptor [33]. Meanwhile, we observed abnormal expression of a few genes involved in immunoglobulin/B cell receptor signaling. Igkj4 showed significant up-regulation in HFD-fed mice, whereas the expression of Igkj1 were significantly decreased. Additionally, Reg3a and Reg3b, which play critical roles in acute-phase response [34,35], were significantly downregulated. The findings suggested that immune dysfunction was implicated in the spleen in response to 3 days of HFD feeding. Interestingly, neurotransmitter uptake related genes were significantly changed, such as solute carrier family 17 (sodium-dependent inorganic phosphate cotransporter), member 6(Slc17a6) which plays a key role in the transport of glutamate into synaptic vesicles before exocytotic release and the regulation of glutamate signaling [36].
Evidence has increasingly shown that a HFD regarded as a primary cause for numerous diseases including diabetes, hypertension, and steatohepatitis. However, few studies have been carried out to examine the effect of a HFD on the lung. To our surprise, the PCA results showed pronounced transcriptional changes in the lung and this is the first report to investigate the lung transcriptome profile after 3 days of HFD feeding. As the lung is a major site of immune regulation, our results revealed that many immune-related genes were significantly altered. Chemokine receptor Ccr9, which have proved to be important in the Treg cells mediated self-tolerance [37], was markedly down-regulated in HFD-fed mice. Sit1, a critical negative regulator of TCR-mediated signaling [38], showed significant down-regulation as well. The decreased expression of Ccr9 and Sit1 suggested an activated inflammatory response affected by HFD in the lung, which was in agreement with the previous study of the involvement of a HFD on lung inflammation [39]. In addition, we discovered decreased expression of CD8a in HFD-fed mice, which is important in cell-mediated immune defense and T-cell development [40]. Moreover, genes involved in T cell receptor signaling and immunoglobulin/ B cell receptor signaling were found to be significant altered. This study also revealed that circulatory system related genes were down-regulated by HFD in the lung. Nppa and Nppb are the precursor of atrial natriuretic peptide (ANP) and b-type natriuretic peptide (BNP), which have important physiological functions in the regulation of vascular tone and plasma volume [41]. ANP exhibits a protective role in the lung function in acute lung injury apart from its vasodilatory and natriuretic effects [42]. Intake of a HFD had been proved to make a slower pulmonary O 2 uptake kinetics and attenuate microvascular blood flow and O 2 delivery during the transition to moderate intensity exercise [43]. Therefore, we speculated that the decreased expression of Nppa and Nppb may contribute to the impaired exercise capacity in HFD-fed mice.