EPA and DHA elicit distinct transcriptional responses to high-fat feeding in skeletal muscle and liver

Abstract

Omega-3 polyunsaturated fatty acids (n-3 PUFAs) exert numerous beneficial biological effects and attenuate diet-induced insulin resistance in rodent models. In the present study, the independent, tissue-specific effects of two nutritionally relevant n-3 PUFAs, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), were characterized in the context of a high-fat diet (HFD). EPA and DHA supplementation (3.2% of total fat) in 6-mo-old male C57BL/6 mice fed an HFD (60% fat) partially mitigated reductions in insulin sensitivity. At 5 wk, the area above the curve below baseline glucose following an intraperitoneal insulin tolerance test was 54.5% lower in HFD than control, whereas HFD + EPA and HFD + DHA showed 27.6% and 17.1% reductions, respectively. At 10 wk, HFD increased mitochondrial oxidative capacity supported by lipid and carbohydrate-based substrates in both liver and skeletal muscle (P < 0.05), with little effect of EPA or DHA supplementation. Whole genome transcriptomic analyses revealed HFD-induced transcriptional changes indicative of inflammation and fibrosis in both liver and muscle. Gene set enrichment analyses indicated a downregulation of transcripts associated with extracellular matrix in muscle (family-wise error rate P < 0.01) and liver (P = 0.04) and in transcripts associated with inflammation in muscle (P = 0.03) in HFD + DHA compared with HFD alone. In contrast, EPA appeared to potentiate some proinflammatory effects of the HFD. In the skeletal muscle, DHA increased the expression of stress-responsive genes, whereas EPA upregulated the expression of transcripts related to cell cycle. Therefore, although both EPA and DHA supplementation during HFD partially preserve insulin signaling, they modulate distinct processes, highlighting their unique biological effects in the context of obesity.

INTRODUCTION

Diet-induced obesity contributes to the rising rates of insulin resistance and type 2 diabetes. Insulin resistance is accompanied by chronic inflammation and altered mitochondrial function in insulin-sensitive tissues such as skeletal muscle and liver (26), reflecting the complex relationships among these variables and potential involvement in the etiology of insulin resistance (10). For example, recruitment and accumulation of proinflammatory macrophages in adipose tissue have been implicated in insulin resistance (15), and inhibition of inflammatory pathways and promotion of anti-inflammatory signaling appear to improve insulin sensitivity (31, 57). Although debate continues about the exact role of mitochondria in the etiology of insulin resistance, it is evident that excessive reactive oxygen species production by these organelles contributes to diet-induced insulin resistance (2).

Therapeutic strategies targeting multiple components in the pathophysiology of insulin resistance could help reduce the comorbidities associated with rising prevalence of obesity and insulin resistance. Dietary omega-3 polyunsaturated fatty acids (n-3 PUFAs) attenuate high-fat diet (HFD)-induced insulin resistance in rodent models (27, 48), although more modest effects are observed in humans (25). The therapeutic benefits of n-3 PUFAs appear to go well beyond their established triglyceride-lowering effects. n-3 PUFAs exhibit potent anti-inflammatory effects, inhibiting both macrophage infiltration into the adipose tissue (50) and the activation of the NF-κB and JNK proinflammatory signaling pathways in these macrophages (39). Emerging evidence shows that dietary n-3 PUFAs influence mitochondria in beneficial ways. For example, fish oil administration attenuated HFD-induced reductions in muscle oxygen consumption, tricarboxylic acid cycle intermediates, and the expression of PGC1-α, a key transcriptional coactivator in mitochondrial biogenesis (27, 33), indicating beneficial effects of these fatty acids on mitochondrial maintenance and function in skeletal muscle. In the liver, n-3 PUFA-induced alterations in genes associated with enhanced lipid β-oxidation and decreased lipogenesis have been observed, concomitant with improvements in the homeostatic model assessment of insulin resistance index (11).

Despite an accumulating body of evidence highlighting the numerous potentially beneficial biological effects of n-3 PUFAs, there is need for a better understanding of how n-3 PUFAs influence skeletal muscle and liver, two key insulin-sensitive tissues, in the context of HFD-induced insulin resistance. A lingering issue in this field is that most studies of n-3 PUFAs involve a mixture of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), two very long-chain n-3 PUFAs that appear to have distinct biological effects and tissue-specific effects. Although both EPA and DHA have been found to reduce insulin resistance and inflammation in HFD-fed rodent models (42), we have shown that EPA better prevents age-related declines in mitochondrial function and protein quality (24), and others have shown that EPA improves glucose uptake in cultured myotubes (23), whereas DHA failed to exert such beneficial effects. The present study was conducted to further characterize the independent and tissue-specific effects of these two nutritionally relevant n-3 PUFAs (EPA vs. DHA) in the context of an HFD in mice. We determined the independent effects of EPA versus DHA on whole body glucose tolerance, insulin sensitivity, and mitochondrial physiology in two key insulin-sensitive tissues: skeletal muscle and liver. In an effort to better understand the independent mechanisms by which EPA and DHA influence these tissues, we explored their effects on the respective transcriptomes through RNA sequencing.

METHODS

Animals.

Six-month-old, male C57BL/6 mice (Jackson Laboratories, Bar Harbor, ME) were obtained and housed individually in a facility maintained at 20°C–22°C and 50%–60% humidity with a 12:12-h light-dark cycle. Following a 10-day acclimation period during which mice were fed standard chow and water ad libitum, mice were randomly assigned to 1 of 4 groups (n = 8 per group): standard chow diet (control; 20% kcal protein, 70% kcal carbohydrate, 10% kcal fat), HFD (20% kcal protein, 20% kcal carbohydrate, and 60% kcal fat), HFD with EPA supplementation (HFD + EPA; 20% kcal protein, 20% carbohydrate, 60% kcal fat enriched with 3.2% purified EPA), and HFD with DHA supplementation (HFD + DHA; 20% kcal protein, 20% carbohydrate, 60% kcal fat enriched with 3.2% purified DHA). The fats in the HFDs were composed exclusively of saturated fats (lard and corn oils), and all HFDs were matched for macronutrient distribution (Supplemental Table S1; All Supplemental Tables are available at https://doi.org/10.6084/m9.figshare.c.4524848). The 3.2% n-3 PUFA enrichment was selected to provide a target daily dose of 350 mg n-3 PUFA·kg⁻¹·day⁻¹; conversion of this dosage to the human equivalent dose based on allometric scaling (37) yields EPA or DHA doses similar to those in high-concentration n-3 PUFA supplements. All diets were supplied by Research Diets Inc. (New Brunswick, NJ). Following acclimation, mice were given ad libitum access to their respective diets for 10 wk, during which time body mass was recorded weekly. Body composition was measured by Echo-MRI (Echo Medical Systems, Houston, TX) before the diet administration and at the end of the 10-wk intervention. Following the 10-wk diet intervention, mice were euthanized with an intraperitoneal injection of pentobarbital sodium administered between 12:00 and 13:00 following an ~6-h fast in a clean cage with ad libitum access to water. To control for tissue harvest time, time of day, and duration of fasting, no more than two mice were euthanized per day. All experiments and procedures were reviewed and approved by the Mayo Clinic Institutional Animal Care and Use Committee (Protocol no. A51211).

Glucose tolerance tests.

At baseline and after 5 and 10 wk of the prescribed diet, mice underwent intraperitoneal glucose tolerance tests (IPGTTs) as previously described (17). Briefly, following a 6-h fast, blood samples were collected from the tail vein for the determination of fasting blood glucose concentration, measured using a glucose meter (AlphaTrak; Abbott Laboratories, Abbot Park, IL) and fasting insulin concentration, measured subsequently in triplicate by enzyme-linked immunosorbent assay (Crystal Chem, Downers Grove, IL) from frozen plasma samples. Following the fasting measurements, mice received 2 g glucose/kg body wt by intraperitoneal injection, and blood glucose concentrations were measured from tail vein blood samples 5, 10, 15, 30, 60, and 120 min following glucose administration. The area of the curve above the fasting glucose concentration was calculated using GraphPad Prism software (v.7, Prism, San Diego, CA).

Insulin tolerance test.

Intraperitoneal insulin tolerance tests (IPITTs) were performed, as previously described (17), 5 days after each glucose tolerance test. Following a 6-h fast, mice were injected intraperitoneally with 0.75 U insulin/kg body wt. Blood samples were collected from the tail vein for blood glucose measurements before insulin administration and 15, 30, 60, and 120 min following the insulin injection. The inverse area under the curve below baseline glucose was calculated using GraphPad Prism software (v.7, Prism).

Mitochondria isolation and functional assessment.

Approximately 1 wk after the final insulin tolerance test, mice were euthanized and livers and quadriceps were immediately collected from each mouse. Muscle and liver tissues were placed in BIOPS buffer (2.77 mM CaK₂EGTA, 7.23 mM K₂EGTA, 5.77 mM Na₂ATP, 6.56 mM MgCl₂ 6H₂O, 20 mM taurine, 15 mM Na₂phosphocreatine, 20 mM imidazole, 0.5 mM DTT, and 50 mM MES), and mitochondria were isolated as previously described (28). Briefly, the muscle and liver were transferred into homogenization buffer (100 mM KCl, 50 mM Tris, 5 mM MgCl₂, 1.8 mM ATP, and 1 mM EDTA, pH 7.2) where the tissue was minced, treated with protease medium (muscle only) for 2 min, and washed twice with homogenization buffer. The tissue was then homogenized for 10 min using a Potter-Elvehjem tissue grinder. Following homogenization, mitochondria were isolated by differential centrifugation of the homogenate, which included a low-speed centrifugation (720 g) to extract the myofibrillar components followed by a high-speed spin (10,000 g) of the resultant supernatant to generate a mitochondrial pellet. The supernatant from the second spin was discarded, and the mitochondrial pellets were resuspended in homogenization buffer for a final high-speed wash spin, after which the mitochondrial pellet was resuspended in Buffer B (225 mM sucrose, 44 mM KH₂PO₄, 12.5 mM Mg acetate, and 6 mM EDTA) at a volume of 4 µL buffer/mg tissue. All procedures were performed using chilled vials and glassware and ice-cold reagents.

High-resolution respirometry was performed to assess the bioenergetics of the isolated mitochondria (Orobos Instruments, Innsbruk, Austria), as previously described (27, 28). The serial addition of substrates, inhibitors, and activators assessed respiratory states 1–4 and isolated electron flow through the respiratory chain complexes. The isolated mitochondrial suspension (50 µL) was added to each oxygraph chamber containing 2 mL respiration buffer (MiRO5, 110 mM sucrose, 60 mM potassium lactobionate, 0.5 mM EGTA, 1 g/L BSA essentially fat free, 3 mM MgCl₂, 20 mM taurine, 10 mM KH₂PO₄, 20 mM HEPES, pH 7.1). Basal respiration in the absence of any substrates was measured (state 1). The addition of 10 mM glutamate and 2 mM malate was used to assess state 2 respiration through complex I. Subsequent addition of 2.5 mM ADP stimulated state 3 respiration with electron flow exclusively through complex I. The addition of cytochrome C (10 µM) was then used to assess mitochondrial membrane integrity. Succinate (10 mM) was then added to assess state 3 respiration through both complex I and II. Rotenone (0.5 µM), a complex I inhibitor, was added, allowing for assessment of state 3 respiration with electron flow through complex II alone. Oligomycin (2 µg/mL) was then added to inhibit ATP synthase activity and allow for the determination of state 4 respiration, an indicator of proton leak or uncoupled respiration. Finally, to ensure that the measured consumption of oxygen was indicative of mitochondrial respiration, 2.5 µM of the complex III inhibitor antimycin A was added. The addition of glutamate, malate, and succinate allowed for the assessment of respiration supported by carbohydrate-based substrates, and parallel experiments using palmitoyl-L-carnitine and malate were performed to assess mitochondrial lipid oxidation. State 2 was assessed with palmitoyl-L-carnitine (0.005 mM) and malate (2 mM) before the addition of 2.5 mM ADP to assess state 3 respiration supported by lipid substrates. Oligomycin (2 µg/mL) and antimycin A (2.5 µM) were then sequentially added to assess state 4 respiration and to verify the mitochondrial origin of the measured respiration, respectively.

The average oxygen flux rates for each condition were corrected for background oxygen flux using Datlab software. The flux rates were then normalized to wet tissue weight (mg) and to protein content within the mitochondrial suspension (µg), measured using the Pierce 660-nm Protein Assay. The respiratory control ratio (RCR) was calculated (state 3 respiration/state 4 respiration) as a measure of mitochondrial coupling efficiency.

RNA sequencing.

RNA sequencing was performed as previously described (24). Whole transcriptome RNA-Seq technology was utilized to assess differences in muscle and liver gene expression between the four diet intervention groups (n = 4 per group). Total RNA was isolated from quadriceps and liver sections via the TruSeq method. RNA libraries were prepared using the TruSeq RNA Sample Prep Kit v2 (Illumina, San Diego, CA), according to manufacturer’s instructions, and loaded onto paired-end flow cells, according to the standard protocol for the Illumina cBot and cBot Paired-End Cluster Kit v3. An Illumina HiSeq 2000 using the TruSeq SBS sequencing kit version 3 and the HCS v2.0.12 data collection software was used to sequence the flow cells as 51 × 2 paired-end reads. Illumina RTA version 1.17.21.3 was used for base calling, and the RNA-Seq data were analyzed by the Mayo Bioinformatics Core pipeline using MAP-RSEQ v.1.2.1, which consists of alignment with TOPHAT 2.0.6 against the hg19 genome build and gene counts with HTSEQ software 0.5.3p9 using Illumina gene annotation files. EdgeR software (version 2.6.2) was used to normalize the gene expression data and assess their differential expression between the diet intervention groups, using the standard recommended protocol. Differential expression between groups was defined by a false discovery rate (FDR)-corrected P value ≤0.01 and an absolute log2 fold change ≥1 (where a value of 0.0 signifies no change). With the low dispersion in the expression of genes (computed as 0.02453 from samples) and n = 4, the power to detect a fold change of >1.0 was 0.8 at an alpha value of 0.05. Ingenuity Pathway Analysis (IPA) and MetaCore pathway analysis were performed on genes defined as differentially expressed. Venn diagrams were generated using BioVenn (19).

Statistical analysis.

Data were tested for the assumptions of normality and homogeneity of variance, and any data failing to meet the assumptions were logarithmically transformed. All data are presented in their original metrics. Changes in body composition, insulin and glucose tolerance, and fasting glucose and insulin levels over the 10-wk diet intervention were analyzed using a mixed-design repeated measures ANOVA. The main effects of time, the between-group main effects of the prescribed diet, and the interaction between time and diet, allowing for the determination of between-group differences in changes in each variable over the course of the intervention, were assessed. The same analysis was used for assessing the raw glucose values obtained during each insulin and glucose tolerance test, with time representing the time in minutes after insulin or glucose administration. Between-group differences in mitochondrial function were analyzed using a univariate ANOVA assessing the main effects of group for each respiration state. Two ANOVAs for each parameter were performed: one in which all four groups were compared and a second that specifically tested the effects of n-3 PUFA supplementation on the HFD by comparing only the HFD + n-3 PUFA groups to the HFD. Results of both analyses are presented in the corresponding figures. When significant main or interaction effects were observed, post hoc pairwise comparisons with Bonferroni corrections were performed. Statistical significance was set a priori at P < 0.05, and analyses were performed using Statistical Package for the Social Sciences (IBM, Armonk, NY).

RESULTS

n-3 PUFA supplementation does not prevent increases in body and fat mass elicited by an HFD.

Mice fed an HFD for 10 wk showed significantly greater increases in body mass compared with mice fed a normal chow diet (group × time interaction P < 0.001), and the HFD-induced increases in body mass were not mitigated by EPA or DHA supplementation (n = 8 per group; Fig. 1A). By week three of the HFD, all three HFD groups had significantly (P < 0.05) greater body masses than control, and no differences in body weight were observed between HFD, HFD + EPA, and HFD + DHA at any time point. Although all groups exhibited significant increases in body mass, fat mass, and percent fat mass after 10 wk, these changes were greater in HFD and HFD + n-3 PUFA groups (main effects of group P < 0.001; group × time interaction P < 0.001), and n-3 PUFA supplementation did not influence body composition (Fig. 1, B and D). Small but significant elevations in lean mass were observed in the HFD and HFD + EPA groups compared with control (group × time interaction P = 0.023; post hoc analyses P < 0.05, Fig. 1C).

Fig. 1.

Ten weeks on a high-fat diet (HFD) increased body mass and fat mass in 6-mo-old male mice, regardless of supplementation with eicosapentaenoic acid (EPA) or docosahexaenoic acid (DHA). Body mass was measured weekly (A), and fat mass (B), lean mass (C), and percent fat mass (D) were measured by echo MRI before and following the diet intervention. n = 8 per group. Data are presented as means and SD. ^λHFD and HFD + EPA differed significantly (P < 0.05) from control within a time point. ^ΦHFD, HFD + EPA, and HFD + DHA significantly differed from control within a time point. *Significant difference from control within a time point. ^#Significant difference from within group baseline values.

n-3 PUFA supplementation improves HFD-induced insulin intolerance without impacting glucose tolerance.

Ten weeks of HFD significantly (P < 0.05) increased fasting glucose and insulin compared with within-group baseline values and to the normal chow control group (Fig. 2, A and B). These changes were evident at week five and were not mitigated by supplementation with EPA or DHA, even after accounting for differences in baseline values (n = 8 per group). Insulin sensitivity, determined from the IPITT inverse area under the glucose curve, was similar at baseline in all groups (Fig. 2C, n = 8 per group) and decreased in all 3 HFD groups at 5 wk and 10 wk (Fig. 2, D and E). Both EPA and DHA attenuated the effects of HFD on insulin sensitivity at 5 wk and 10 wk, with HFD + EPA and HFD + DHA not significantly different from control (main effects of group P = 0.031; group × time interaction P = 0.009) (Fig. 2, C–E); however, the HFD + n-3 PUFA groups did not significantly differ from the HFD at either time point. Whole body glucose tolerance, determined from IPGTT area above baseline, decreased in all 3 HFD groups at 5 and 10 wk, with no effects of EPA or DHA supplementation (Fig. 2, F–H; n = 8 per group).

Fig. 2.

Eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) mitigated decreases in insulin sensitivity induced by 10 wk of a high-fat diet (HFD) but failed to improve HFD-induced reductions in glucose tolerance and elevations in fasting insulin and glucose. Fasting plasma glucose (A) and insulin (B) were measured, and intraperitoneal insulin tolerance tests (C–E) and glucose tolerance tests (F–H) were performed at baseline and after 5 and 10 wk of the study diet. The area above the curve below the baseline glucose value (AACBB) was calculated from the glucose kinetics curve following a 0.75-U/kg bolus of insulin (C–E). The area above baseline (AAB) was calculated from the kinetics curve generated from plasma glucose concentrations measured periodically after a 2 g/kg bolus glucose injection (F–H). *Significantly different from control within a time point. ^Significantly different from HFD within a time point. #Significantly different from within-group baseline value. ^ΦAll HFD groups significantly different from control within a time point. ^λHFD significantly different from HFD + DHA and HFD + EPA within a time point. ^πHFD + EPA and HFD + DHA significantly different from control. n = 8 per group. Data are expressed as means and SD.

HFD increases oxidative capacity in liver and skeletal muscle.

Respiratory capacity supported by carbohydrate- and lipid-based substrates was measured in isolated mitochondria from liver (Fig. 3, A and B; n = 8 per group) and skeletal muscle (Fig. 3, C and D; n = 8 per group). The function of individual components of the electron transport chain and respiration at multiple energetic states was evaluated by the serial addition of these substrates and of various inhibitors. The addition of cytochrome C, used to assess mitochondrial integrity, did not augment respiration rates, indicating that mitochondrial membranes remained intact during the isolation procedures. Liver mitochondrial oxidative capacity was significantly increased in all three HFD groups, evident from increased state 3 respiration under conditions of complex II-supported respiration (succinate + rotenone) (Fig. 3A). State 3 respiration supported by complex I substrates (glutamate + malate) was similar across all groups. Modest differences between groups were observed in the presence of substrates supporting β oxidation (palmitoyl-L-carnitine + malate), with HFD + EPA demonstrating significantly higher rates of lipid oxidation compared with controls, whereas the other HFD groups did not reach statistical significance compared with controls (Fig. 3B). No significant differences in state 3 respiration were observed between the HFD + n-3 PUFA groups and HFD alone. Uncoupled respiration (state 4) in the presence of the ATP synthase inhibitor oligomycin was elevated in both n-3 PUFA-supplemented groups compared with control; however, there were no significant group differences in the RCR (state 3/state 4), indicating that the overall coupling efficiency was not markedly different between groups (Fig. 3, A and B).

Fig. 3.

Ten weeks of a high-fat diet (HFD) with or without omega-3 polyunsaturated fatty acid supplementation enhanced state 3 respiration supported by carbohydrate-based substrates glutamate, malate, and succinate (A and C) and lipid-based substrates palmitoyl-L-carnitine and malate (B and D) in mitochondria isolated from the liver (A and B) and quadriceps (C and D). Mitochondrial oxidative capacity was assessed using high-resolution respirometry. The serial addition of substrates and inhibitors allowed for the assessment of oxygen consumption at multiple energetic states and through distinct electron transport chain complexes (CI, CII). Respiration was normalized to tissue weight. Mitochondrial efficiency was evaluated using the respiratory control ratio (RCR, state 3/state 4). n = 8 per group. Data are expressed as means and SD. *Significantly different from control within a state. +Significantly different from HFD within a state. DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid.

Muscle oxidative capacity was elevated in mice fed a high-fat diet compared with control, with no effect of EPA or DHA supplementation (Fig. 3, C and D). The effect of an HFD on state 3 respiration was evident under conditions in which respiration was supported by carbohydrate-based substrates (Fig. 3C) and lipid-based substrates (Fig. 3D). Similar to liver mitochondria, state 4 respiration was elevated with EPA and DHA with both lipid- and carbohydrate-based substrates. The respiratory control ratios were elevated in the presence of lipid substrates, but only HFD alone was significantly higher than control (HFD P = 0.002; HFD + EPA P = 0.104; HFD + DHA P = 0.209; Fig. 3, C and D). In direct comparisons between the three HFD groups, state 4 respiration supported by lipid-based substrates was significantly elevated in HFD + DHA compared with HFD alone, but state 3 respiration and RCR were not significantly different between HFD and either HFD + n-3 PUFA group. Similar results were observed when respiration was normalized to tissue weight or mitochondrial protein content (data not shown).

HFD induces transcriptomic changes related to fibrosis and inflammation in liver and skeletal muscle.

The HFD induced significant changes in the liver [539 differentially expressed genes (DEGs); 348 up, 191 down] and skeletal muscle (406 DEGs; 347 up, 59 down) transcriptomes when compared with control. IPA and MetaCore pathway analysis of DEGs in the liver indicated patterns of gene expression consistent with the development of hepatic fibrosis (Fig. 4, B and C). With the differential expression of 28 genes related to “Hepatic Fibrosis/Hepatic Stellate Cell Activation,” this canonical pathway was the most significantly affected by the HFD, as identified by IPA (P = 2.94E-16). IPA also predicted significant upregulation in glycoprotein VI (GP6) signaling (P = 5.01E-11), significant alterations in atherosclerosis signaling (P = 2.65E-06), and predicted a significant downregulation in matrix metalloprotease inhibition (P = 1.42E-05). Similarly, pathway analysis using MetaCore software identified extracellular matrix (ECM) remodeling as the most significantly enriched pathway in the HFD-control comparison (FDR = 8.186E-11), and a number of the enriched process networks were related to ECM remodeling and cell recruitment and adhesion, which may contribute to the development of fibrosis. IPA analysis of diseases and biological processes affected by the HFD showed significant enrichment in broad categories related to hepatic fibrosis, including organismal injury and abnormalities (432 DEGs, P values 2.1E-20 to 5.97E-07), connective tissue disorders (138 DEGs, P values 1.21E-19 to 1.87E-07), cellular movement (145 DEGs, P values 2.95E-16 to 3.78E-07), and cell-to-cell signaling and interaction (112 DEGs, P values 2.19E-15 to 2.38E-07). The dysregulated biological functions also pointed to significant liver inflammation, with inflammatory disease (121 DEGs, P values 1.31E-15 to 2.8E-07), inflammatory response (143 DEGs, P values 1.96E-14 to 3.78E-07), and immune cell trafficking (88 DEGs, P values 4.21E-14 to 3.78E-07) all enriched in the HFD group (data not shown). Changes in the categories of endocrine system disorders (182 DEGs, P values 4.73E-17 to 5.67E-07), metabolic disease (128 DEGs, P values 4.73E-17 to 3.64E-07), and lipid metabolism (99 DEGs, P values 1.47E-13 to 5.84E-07) were also observed in the IPA analysis (data not shown).

Fig. 4.

Ten weeks of a high-fat diet (HFD) resulted in liver transcriptional changes indicative of inflammation and fibrosis. Supplementation with docosahexaenoic acid (DHA) during the HFD mitigated some of these transcriptional changes. A: comparison of differentially expressed genes when HFD, HFD + DHA, and HFD + eicosapentaenoic acid (EPA) were compared with the control. B: top 20 enriched pathways for each HFD group compared with the control, as identified by Ingenuity Pathway Analysis. Inflection point reflects an enrichment P value of 0.05. C: all significantly enriched process networks for each HFD group compared with the control, as identified by MetaCore. n = 4 per group. Numbers in parentheses reflect the total number of genes in each network. *Significant enrichment (false discovery rate < 0.05) in the HFD group. ^Significant enrichment in the respective omega-3 polyunsaturated fatty acid (n-3 PUFA)-supplemented groups. ECM, extracellular matrix.

As with the liver, HFD influenced the expression of genes related to inflammation and fibrosis in skeletal muscle (Fig. 5). Of the 406 DEGs (Fig. 5A), IPA identified 152 related to the inflammatory response (P values 1.72E-26 to 8.73E-06), 110 to immune cell trafficking (P values 1.12E-31 to 8.73E-06), 138 to cellular movement (P values 1.12E-31 to 8.73E-06), and 120 to cell-to-cell signaling and interaction (P values 1.42E-22 to 8.73E-06), indicating significant immune cell recruitment into the skeletal muscle. Many of the identified dysregulated canonical pathways, including complement system, phagosome formation, acute-phase response signaling, neuroinflammatory signaling, and granulocyte and agranulocyte adhesion and diapedesis, were also indicative of significant skeletal muscle inflammation and leukocyte infiltration (Fig. 5B). Furthermore, IPA-identified upstream regulators of the observed HFD-induced changes in the transcriptome included lipopolysaccharide (P = 9.63E-24), TNF (P = 3.14E-20), IL-6 (P = 3.02E-17), IFNγ (P = 1.79E-16), and IL-1β (P = 8.00E-16), potent stimulators of inflammation (data not shown). As in the liver, pathways related to fibrosis, GP6 signaling, and the inhibition of matrix metalloproteases were also significantly enriched in the skeletal muscle of the HFD group because of the differential expression of multiple collagen, C-C motif chemokine ligands and receptors, and matrix metallopeptidase genes, indicative of skeletal muscle fibrosis. These findings were supported by the MetaCore analysis, in which many of the enriched process networks were related to immune cell recruitment, ECM interactions and remodeling, and inflammation (Fig. 5C). IPA also demonstrated a significant enrichment in type II diabetes mellitus signaling (P = 2.47E-4) and identified many genes related to metabolic disease (96 genes, P values 3.74E-20 to 2.45E-06), endocrine system disorders (156 genes, P values 3.8E-15 to 8.46E-06), and lipid metabolism (76 genes, P values 1.04E-11 to 6.58E-06).

Fig. 5.

Ten weeks of a high-fat diet (HFD) resulted in muscle transcriptional changes indicative of fibrosis and immune cell recruitment. Supplementation with docosahexaenoic acid (DHA) during the HFD mitigated some of these transcriptional changes, and supplementation with eicosapentaenoic acid (EPA) enhanced the expression of genes related to the cell cycle. A: comparison of differentially expressed genes when HFD, HFD + DHA, and HFD + EPA were compared with the control. B: top 20 enriched pathways for each HFD group compared with the control, as identified by Ingenuity Pathway Analysis. Inflection point reflects an enrichment P value of 0.05. C: all significantly enriched process networks for each HFD group compared with the control, as identified by MetaCore (reproduction: progesterone signaling was significantly enriched in the HFD + EPA group but not included in the figure, as it was not identified in either HFD or HFD + DHA). n = 4 per group. Numbers in parentheses reflect the total number of genes in each network. *Significant enrichment (false discovery rate < 0.05) in the HFD group. ^Significant enrichment in the respective omega-3 polyunsaturated fatty acid (n-3 PUFA)-supplemented groups. ECM, extracellular matrix.

DHA supplementation mitigates transcriptomic changes induced by HFD.

To determine if EPA or DHA mitigated any of the transcriptional changes induced by HFD in liver or influenced expression of unique liver gene transcripts, the transcriptional profile of the HFD + EPA and HFD + DHA groups were also compared with that of the control group. HFD + DHA influenced the expression of 525 (314 up, 211 down) genes, and HFD + EPA influenced the expression of 711 genes (493 up, 218 down) (Fig. 4A). A total of 317 genes were differentially expressed in all 3 HFD groups, and many of the same pathways and functions observed to be enriched in the HFD group were also enriched in the HFD + n-3 PUFA groups, reflecting the strong transcriptional influence of HFD in liver regardless of whether or not mice were supplemented with EPA or DHA. Whereas many of these enriched pathways remained significantly enriched in the presence of DHA, IPA pathway enrichment was somewhat attenuated (Fig. 4B), as was the number of DEGs in many ECM- and inflammation-related Metacore process networks (Fig. 4C). Functional classification of those transcripts that were differentially expressed with HFD but not with HFD + DHA still included enrichment of pathways related to hepatic fibrosis, whereas the enriched pathways for transcripts differentially expressed in HFD + DHA but not HFD alone related primarily to immune signaling (Supplemental Table S2). In contrast, EPA seemed to amplify many of the HFD-induced changes, with greater numbers of DEGs (Fig. 4C) and greater pathway enrichment in many of the IPA canonical pathways enriched by HFD (Fig. 4B). Transcripts for a number of integrins and chemokines were differentially expressed in the HFD + EPA group but not the HFD group and were involved in pathways relating to immune activation. As was observed with DHA supplementation, EPA prevented the downregulation of genes involved in stress signaling, including Fgfr1, Hspa1a, and Hspa1b (Supplemental Tables S2 and S3).

DHA supplementation also appeared to attenuate the HFD-induced differential expression of skeletal muscle genes, with the differential expression of only 321 genes (271 up, 50 down) compared with the 406 genes altered by HFD alone and the 546 (487 up, 59 down) altered by HFD + EPA (Fig. 5A). MetaCore pathway analyses showed DHA-induced reductions in the enrichment of process networks related to cell chemotaxis and adhesion (Fig. 5C) and lower enrichment of IPA pathways related to fibrosis and ECM remodeling, including hepatic fibrosis and stellate cell activation and GP6 signaling pathways (Fig. 5B). DHA prevented the upregulation of a number of transcripts related to immune signaling, as observed by an analysis of the 152 transcripts differentially expressed in the HFD group but not the HFD + DHA group when compared with the control (Supplemental Table S4). DHA supplementation also altered the expression of genes involved in stress signaling (Fos, Hspa1a, and Hspa1b, Srebf1, Hsp90aa1), changes that were not observed with HFD alone (Supplemental Table S4). As was observed in the liver, HFD + EPA seemed to amplify many of the changes observed in the group receiving HFD alone, with greater numbers of DEGs and more significantly enriched pathways (Fig. 5, A–C). EPA supplementation also significantly increased the expression of genes related to the cell cycle, as identified in the MetaCore analysis (Fig. 5C). Pathway analysis of the 216 transcripts that were differentially expressed with HFD + EPA but not with HFD revealed enrichment in pathways related to the cell cycle and immune signaling (Supplemental Table S5). HFD + EPA prevented the differential expression of genes involved in adipogenesis and the downregulation of Fasn, a key enzyme in de novo lipogenesis (Supplemental Table 5).

Direct comparisons of muscle and liver transcriptomes between the three HFD groups were also performed to further evaluate unique biological and tissue-specific effects of EPA and DHA. A relatively small number of genes were differentially expressed in these comparisons (Fig. 6), and pathway analysis yielded few enriched canonical pathways (Supplemental Tables S6 and S7). Nevertheless, when compared with the HFD alone, HFD + DHA resulted in a greater number of DEGs in both liver and skeletal muscle than did HFD + EPA, and HFD + EPA more closely resembled the HFD alone group. Assessment of the IPA analysis for the three HFD group comparisons also indicated greater impacts of DHA. DHA supplementation modulated inflammatory gene responses (skeletal muscle: 10 DEGs, P values 3.75E-07 to 1.14E-02; liver: 12 DEGs, P values 4.02E-05 to 8.64E-03), cell death and survival (muscle: 17 DEGs, P values 1.54E-06 to 1.65E-02; liver: 11 DEGs, P values 6.26E-05 to 8.75E-03), and organismal injury and abnormalities (muscle: 30 DEGs, P values 1.77E-09 to 1.71E-02; liver: 33 DEGs, P values 1.6E-04 to 8.75E-03) in both the skeletal muscle and the liver (data not shown).

Fig. 6.

The liver (A) and skeletal muscle (B) transcriptomic profiles of mice fed a high-fat diet (HFD) for 10 wk were compared with mice fed an HFD and supplemented with either eicosapentaenoic acid (EPA) or docosahexaenoic acid (DHA) (n = 4 per group). HFD and HFD + EPA exhibited similar transcriptional profiles, whereas HFD + DHA exhibited a greater number of differentially expressed genes when compared with HFD alone or HFD + EPA.

To explore hypothesized mechanisms by which n-3 PUFAs influence the liver and skeletal muscle, gene set enrichment analysis was performed using the MitoCarta2.0 mitochondrial protein gene set (8), as well as curated gene sets relating to inflammation, macrophages, the ECM, fatty acid oxidation, insulin signaling, and glycolysis (Table 1). Mitochondrial and fatty acid oxidation gene sets were enriched in the liver of HFD-fed mice supplemented with DHA. In the muscle, mitochondrial gene sets were also enriched in the HFD + DHA group, whereas the expression of genes related to the ECM and inflammation were downregulated compared with HFD alone. In contrast, when compared with the HFD alone, inflammation, macrophage, and ECM gene sets were enriched in the livers of the HFD + EPA group, and inflammation and macrophage gene sets were enriched in the skeletal muscle, observations that suggest that DHA attenuates whereas EPA potentiates the proinflammatory effects of HFD.

Table 1.

Gene set enrichment analyses comparing the effects of 10 wk of an HFD to an HFD supplemented with either EPA or DHA on the liver and skeletal muscle transcriptomes

		Liver			Skeletal Muscle
Gene set	Comparison	NES	FDR Q value	FWER P value	NES	FDR Q value	FWER P value
Extracellular matrix	HFD vs. HFD + DHA	1.26	0.06	0.04	2.08	<0.001	<0.001
	HFD vs. HFD + EPA	−1.49	0.002	0.002	−1.13	0.19	0.16
Mitochondria (MitoCarta2.0)	HFD vs. HFD + DHA	−1.58	<0.001	<0.001	−2.15	<0.001	<0.001
	HFD vs. HFD + EPA	1.45	<0.001	<0.001	0.99	0.49	0.017
Insulin signaling	HFD vs. HFD + DHA	1.27	0.13	0.09	−1.31	0.12	0.05
	HFD vs. HFD + EPA	1.29	0.10	0.03	−1.30	0.09	0.06
Glycolysis	HFD vs. HFD + DHA	1.20	0.17	0.12	−1.25	0.13	0.05
	HFD vs. HFD + EPA	0.97	0.47	0.13	−1.08	0.32	0.24
Inflammation	HFD vs. HFD + DHA	1.06	0.30	0.25	1.23	0.04	0.03
	HFD vs. HFD + EPA	−1.90	<0.001	<0.001	−1.53	<0.001	<0.001
Macrophages	HFD vs. HFD + DHA	1.47	0.008	0.006	1.41	0.02	0.01
	HFD vs. HFD + EPA	−1.71	<0.001	<0.001	−1.51	0.005	0.004
Fatty acid oxidation	HFD vs. HFD + DHA	−1.76	0.005	0.002	−1.21	0.17	0.07
	HFD vs. HFD + EPA	−1.03	0.42	0.29	−1.30	0.10	0.07

Study revealed opposing effects of EPA and DHA on extracellular matrix, inflammation, and mitochondrial gene sets. Transcriptomic analysis was performed on n = 4 for each group. A positive NES indicates enrichment in the HFD group, and a negative NES indicates enrichment in the respective HFD + n-3 PUFA group. The MitoCarta2.0 was used to assess mitochondrial gene expression. Additional gene sets were curated from literature and from the Kegg, Biocarta, and Reactome pathways. DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; FDR, false discovery rate; FWER, family-wise error rate; NES, normalized enrichment score.

DISCUSSION

This study shows that although many of the phenotypic changes characteristic of an HFD were unchanged by EPA or DHA supplementation (e.g., body mass, body composition, mitochondrial oxidative capacity), both EPA and DHA supplementation partially mitigated HFD-induced impairments in insulin sensitivity and modestly influenced mitochondrial coupling efficiency in liver and skeletal muscle. We also show that EPA and DHA elicit distinct transcriptional changes in liver and skeletal muscle when supplemented against the background of an HFD in mice. Specifically, HFD alone triggered transcriptional patterns in liver and muscle consistent with increased fibrosis, inflammation, and immune cell trafficking. These transcriptional patterns were attenuated by DHA but amplified by EPA in both tissues, highlighting the fact that EPA and DHA are nutritionally important n-3 PUFAs with distinct biological effects on insulin-sensitive tissues.

Neither EPA nor DHA supplementation attenuated HFD-induced weight gain or adiposity. Whereas others report that n-3 PUFA supplementation can mitigate HFD-induced increases in body and adipose mass (30, 42, 45), the findings in the present study are consistent with previous data from our laboratory (27) and others (12, 49) that fail to show an effect of n-3 PUFAs on HFD-induced weight gain. These discrepancies may be related to the n-3 PUFA dosage in relation to the severity and duration of the diet; in the present study, the HFD contained a high proportion of fat (35% wt/wt), and the n-3 PUFA dosage, selected to be representative of n-3 PUFA supplementation in humans, was relatively low. Although the n-3 PUFA-supplemented mice did not differ from the HFD mice in body or adipose mass, both EPA and DHA supplementation partially mitigated impairments in insulin sensitivity during HFD. This modest protective effect of EPA and DHA on whole body insulin sensitivity is consistent with other preclinical animal studies demonstrating insulin-sensitizing benefits of n-3 PUFA supplementation, as recently reviewed (26); however, in their review, Lalia and Lanza (26) make the important point that translational studies in humans are more ambiguous, with multiple n-3 PUFA interventions failing to show improvements in insulin sensitivity. The current study adds to this body of preclinical evidence, largely represented by studies using EPA and DHA in combination, by showing that EPA and DHA exhibit similar protective effects on insulin sensitivity even when given separately. In addition, the failure of the n-3 PUFAs to improve responses during the IPGTT while improving responses to the IPITT indicates a direct effect of n-3 PUFAs on insulin sensitivity, as the IPITT is a direct test of basal insulin sensitivity, particularly in the skeletal muscle (17). In contrast, the glucose response during the IPGTT reflects the combined effects of insulin secretion, insulin action, and “glucose effectiveness” (3).

Since altered mitochondrial physiology has been implicated in the development of insulin resistance (40, 41), and n-3 PUFAs exert beneficial influence on mitochondria, we determined how EPA and DHA influenced mitochondrial function in skeletal muscle and liver tissues. We observed elevated state 3 respiration in skeletal muscle mitochondria of HFD-fed mice. These findings are consistent with multiple studies reporting enhanced mitochondrial function with high-fat feeding (16, 27, 51), including previous work from our laboratory in which HFD-fed, insulin-resistant mice exhibited elevated skeletal muscle oxidative capacity, as well as a lower respiratory exchange ratio, indicating greater reliance on lipids as a fuel source (27). As discussed in our previous work, although individuals with insulin resistance or type 2 diabetes exhibit mitochondrial dysfunction (40, 41) and short-term high-fat feeding can impair mitochondrial gene expression and function (20, 47), a growing body of literature challenges the role of mitochondrial dysfunction in the pathology of insulin resistance (18) and indicates that animals fed an HFD may adapt to the lipid overload by increasing lipid oxidative capacity. Despite precedent literature demonstrating that n-3 PUFAs induce the expression of key genes involved in mitochondrial function and biogenesis (22, 27) and improve mitochondrial function (33), we did not observe any additional increases in muscle state 3 respiration by EPA or DHA beyond what was observed with HFD alone. In the liver mitochondria, we observed significant increases in carbohydrate-supported state 3 respiration with HFD, regardless of whether EPA or DHA were included. Although modest HFD-induced increases in lipid-supported state 3 respiration in liver mitochondria were observed, significant increases compared with control were noted only in mice supplemented with EPA, not DHA or HFD alone. Although HFD-induced enhancements in mitochondrial respiration may be a compensatory mechanism to support increased substrate availability, the further increased lipid oxidation in the HFD + EPA group may be advantageous, given that liver mitochondrial impairments, including reduced fatty acid oxidation and CPT1 activity precede the development of nonalcoholic fatty liver disease in overfed rats (44). Recently, both EPA and DHA were found to reduce HFD-induced histological changes indicative of hepatic steatosis; although no differences in genes involved in β oxidation were observed between the two n-3 PUFAs, EPA exhibited stronger triglyceride-lowering effects (49). Based on the present data, it is tempting to speculate that EPA may protect the liver from subsequent damage by enhancing lipid oxidation and attenuating the build-up of ectopic lipid, but it is important to note that EPA did not significantly increase lipid oxidation compared with HFD alone.

Transcriptomic analyses of liver and skeletal muscle were performed to broadly interrogate the tissue-specific effects of EPA and DHA and identify novel potential mechanisms behind their biological effects. Despite the robust HFD stimulus that dominated the differential expression of genes in all three HFD groups, DHA supplementation appeared to mitigate some of the transcriptional changes associated with HFD. In liver, HFD elicited significant upregulation of genes related to ECM remodeling, hepatic stellate cell activation, epithelial-to-mesenchymal cell transition, and hypoxia, all indicative of the development of liver fibrosis and the endothelial dysfunction that accompanies it (55). Whereas EPA appeared to exacerbate some of the HFD-induced changes, DHA seemed to provide protection against hepatosteatosis, mitigating the differential expression of several fibrosis-related genes, including Cd14, Plau, Apoa4, and Sparc, which have been found to be predictive of fibrosis severity (5, 52). Furthermore, DHA reduced the expression of genes involved in cell-cell and cell-matrix interactions, interactions that can result in increased liver stiffness, which may precede fibrosis (53, 54). In addition, macrophage infiltration, which is considered a cause and characteristic of hepatosteatosis (35), was significantly enriched in the HFD group when compared with HFD + DHA using gene set enrichment analysis. Interestingly, supplementation with DHA prevented the downregulation of the expression of the epidermal growth factor receptor gene observed in the HFD group. Epidermal growth factor receptor has recently been shown to be downregulated in both humans and animals with hepatosteatosis, and its restoration in diabetic obese mice reduced liver injury and improved liver regeneration (58). Taken together, these findings support previous work showing that DHA offers greater protection against HFD-induced hepatic fibrosis than EPA and better attenuates liver inflammation and oxidative stress (12, 32). The whole genome transcriptomic analysis implicates a number of potential mechanisms by which DHA may exert its beneficial effects in the liver that may warrant further exploration and validation by additional biochemical analyses.

Both EPA and DHA improved insulin signaling as measured by the IPITT, despite what appeared to be indications of exacerbated liver fibrosis and inflammation in the EPA-supplemented group. Fibroblast growth factor (FGF) signaling in the liver has been linked to the regulation of glucose and lipid metabolism and has been shown to improve insulin sensitivity, lower blood glucose, increase liver energy expenditure, decrease triglycerides, and reduce hepatic steatosis and inflammation (29, 56). As the preferential receptor for FGF21, FGFR1 expression is vital for its signaling (29), and despite elevated levels of tissue and circulating FGF21, obese mice show reduced FGF21 signaling responses, as well as reduced mRNA expression of Fgfr1 in the liver, indicating that obese mice are FGF21-resistant (14). In the current study, we show that HFD reduced the expression of Fgfr1, but both EPA and DHA supplementation preserved its expression despite the background of HFD. This finding suggests that the protective effects of n-3 PUFA on diet-induced insulin resistance may be mediated, in part, by increased liver FGFR1 expression, and future work examining liver FGF signaling in n-3 PUFA-supplemented mice may be warranted.

In skeletal muscle, HFD was associated with differential expression of genes involved with cell adhesion and ECM remodeling, including a number of leukocyte-related integrins, matrix metalloproteases, and collagens, mirroring transcriptional changes previously observed in the skeletal muscle of rhesus macaques fed a Western-style diet for 6 mo (34). Local inflammation and the infiltration of proinflammatory macrophages may be responsible for this ECM remodeling and contribute to skeletal muscle fibrosis (55). Consistent with this, we observed increased transcript levels of a number of macrophage-associated genes (Itgax, Ccl2, Emr1, Cd68) in mice fed an HFD alone. Many of the HFD-related transcriptional patterns were still evident following n-3 PUFA supplementation; however, DHA, but not EPA, attenuated the effects of the HFD on ECM and inflammation-related genes. Given prior associations between skeletal muscle insulin resistance and upregulation of ECM-related genes and increased collagen content (55), it is reasonable to hypothesize that, similar to liver, DHA may preserve insulin signaling in the skeletal muscle by inhibiting HFD-induced fibrosis.

Several genes upregulated in skeletal muscle with HFD + DHA compared with HFD alone include stress-responsive genes (Fosb, Fos, Hspa1a, Hspa1b, Dnajb1, Atf3). The heat shock protein (HSP)70 genes (Hspa1a, Hspa1b), which can be induced by reactive oxygen species or inflammation, play a role in muscle size and function and appear to have potent anti-inflammatory effects, interfering in inflammatory signaling cascades and exerting protective effects in inflammatory conditions (6, 21). The differential expression of these heat shock proteins also contributed to significant enrichment of the unfolded protein response (UPR), an adaptive stress response that helps maintain endoplasmic reticulum homeostasis. Fos, Fosb, and Atf3 are immediate early genes that respond to an array of stressors and are components of activator protein 1 transcription factor, a diverse group of heterodimers composed of FOS, JUN, and ATF proteins that have been implicated in a multitude of processes including cell proliferation, apoptosis, and mitochondrial gene expression (9, 46). Skeletal muscle expression of FOS and ATF3 are transiently upregulated following exercise and are postulated to mediate some of the beneficial effects of exercise, including increased expression of mitochondrial proteins and antioxidants and blunted expression of inflammation-related genes, respectively (13, 38). Furthermore, one of the MetaCore pathway maps uniquely upregulated in the HFD + DHA group compared with the control was the sirtuin 6 pathway, which responds to oxidative stress and plays a role in maintaining DNA integrity (4). The activation of these stress-responsive genes may therefore be beneficial and indicative of enhanced cytoprotection and the activation of cellular defense mechanisms in the DHA-supplemented group. Similar defense mechanisms also appeared to be activated by DHA in the liver, evident from upregulated Hspa1a and Hspa1b in HFD + DHA compared with HFD and enrichment of the sirtuin 6 and UPR pathways.

Unlike DHA, EPA failed to mitigate HFD-induced transcriptomic changes in the skeletal muscle or liver. In skeletal muscle, EPA supplementation upregulated genes associated with the cell cycle, findings that have been observed in stimulated macrophages exposed to high doses of EPA (1) and in peripheral blood mononuclear cells from humans supplemented with fish oil for 7 wk (36). In these studies, pathways related to apoptosis, inflammation, oxidative stress, endoplasmic reticulum stress, and the UPR were also enriched with n-3 PUFA supplementation, consistent with observations in the present study. By upregulating genes related to cellular and oxidative stress, n-3 PUFAs may function to prime the processes vital for cell homeostasis and maintenance of cell function, particularly in response to redox stress (1, 36). Interestingly, in the present study, it appears that EPA and DHA may modulate different aspects of these processes, as EPA increased the expression of genes related to cell cycle, whereas DHA appeared to upregulate genes related to apoptosis and cellular stress responses.

The broad exploration of the effects of EPA and DHA supplementation on genome-wide transcriptomic changes in the liver and skeletal muscle of mice fed an HFD provides a strong basis for the future examination of mechanisms by which DHA and EPA exert their beneficial effects on insulin sensitivity and in the case of DHA, fibrosis, and inflammation. These findings also highlight the importance of examining the tissue-specific effects of these individual n-3 PUFAs when translating these findings to human studies. As reviewed by Lalia and Lanza (26), whereas the insulin-sensitizing effects of n-3 PUFAs observed in animal models have largely failed to translate to humans, studies in which participants had high levels of inflammation, n-3 PUFA supplementation improved insulin sensitivity (7, 43), even in the absence of reductions in markers in systemic inflammation. Therefore, the beneficial effects of n-3 PUFAs may be mediated through their anti-inflammatory effects, particularly in the tissue. Although providing a basis for future human studies, the study is not without its limitations. The discussion of the effects of DHA and EPA on skeletal muscle and liver changes based on transcriptomic changes is admittedly speculative, as additional biochemical analyses were not performed, nor were phenotypic characterizations of the muscle and liver tissue. Furthermore, the inclusion of a group that received both EPA and DHA may have provided interesting insight into the interplay between these two n-3 PUFAs, particularly as they are often administered in concert. In addition, though the dosage for EPA and DHA was the same, we did not measure DHA and EPA levels in blood or tissue and are therefore unable to comment on their bioavailability or the interconversion of EPA and DHA.

In summary, 10 wk of supplementation with either EPA or DHA partially mitigated HFD-induced reductions in insulin sensitivity, without influencing body composition. Interestingly, EPA and DHA appeared to have distinct effects on skeletal muscle and liver. In both tissues, DHA better mitigated HFD-induced transcriptional changes related to inflammation and fibrosis. In contrast, HFD-induced transcriptional changes were largely unaffected by EPA, but EPA modestly increased liver mitochondrial lipid oxidative capacity, potentially a mechanism for lessening the impact of diet-induced lipid overload. Furthermore, transcriptional profiles revealed that although EPA and DHA both have beneficial effects on maintaining cell function, they appear to do so through distinct processes. The results of this study highlight the unique biological effects of EPA and DHA on the liver and skeletal muscle and will inform future studies seeking to utilize n-3 PUFA supplementation to combat obesity and its associated comorbidities.

GRANTS

This work was supported by a grant from the Mayo Clinic Center for Cell Signaling in Gastroenterology and the National Institute of Diabetes and Digestive and Kidney Diseases (P30-DK-084567) and by Grant UL1-TR-000135. H. E. Kunz was supported by T32-AR-056950.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

I.R.L. conceived and designed research; I.R.L. performed experiments; H.E.K., S.D., and I.R.L. analyzed data; H.E.K. and I.R.L. interpreted results of experiments; H.E.K. and I.R.L. prepared figures; H.E.K. and I.R.L. drafted manuscript; H.E.K., S.D., and I.R.L. edited and revised manuscript; H.E.K., S.D., and I.R.L. approved final version of manuscript.