Consuming a Linoleate-Rich Diet Increases Concentrations of Tetralinoleoyl Cardiolipin in Mouse Liver and Alters Hepatic Mitochondrial Respiration

Abstract

Background

Hepatic mitochondrial dysfunction is a major cause of fat accumulation in the liver. Individuals with fatty liver conditions have hepatic mitochondrial structural abnormalities and a switch in the side chain composition of the mitochondrial phospholipid, cardiolipin, from poly- to monounsaturated fatty acids. Linoleic acid (LA), an essential dietary fatty acid, is required to remodel nascent cardiolipin (CL) to its tetralinoleoyl cardiolipin (L₄CL, CL with 4 LA side chains) form, which is integral for mitochondrial membrane structure and function to promote fatty acid oxidation. It is unknown, however, whether increasing LA in the diet can increase hepatic L₄CL concentrations and improve mitochondrial respiration in the liver compared with a diet rich in monounsaturated and saturated fatty acids.

Objectives

The main aim of this study was to test the ability of a diet fortified with LA-rich safflower oil (SO), compared with the one fortified with lard (LD), to increase concentrations of L₄CL and improve mitochondrial respiration in the livers of mice.

Methods

Twenty-four (9-wk-old) C57 BL/J6 male mice were fed either the SO or LD diets for ∼100 d, whereas food intake and body weight, fasting glucose, and glucose tolerance tests were performed to determine any changes in glycemic control.

Results

Livers from mice fed SO diet had higher relative concentrations of hepatic L₄CL species compared with LD diet-fed mice (P value = 0.004). Uncoupled mitochondria of mice fed the SO diet, compared with LD diet, had an increased baseline oxygen consumption rate (OCR) and succinate-driven respiration (P values = 0.03 and 0.01). SO diet-fed mice had increased LA content in all phospholipid classes compared with LD-fed mice (P < 0.05).

Conclusions

Our findings reveal that maintaining or increasing hepatic L₄CL may result in increased OCR in uncoupled hepatic mitochondria in healthy mice whereas higher oleate content of CL reduced mitochondrial function shown by lower OCR in uncoupled mitochondria.

Introduction

Mitochondrial dysfunction plays a key role in the hepatic fat accumulation that can progress into nonalcoholic fatty liver disease (NAFLD) [1]. Disruptions of the inner mitochondrial membrane (IMM) permit accumulation of reactive oxygen species (ROS) that may bidirectionally worsen IMM damage from disruptions in the complex proteins and, in turn, causes dysregulated beta oxidation [[2], [3], [4], [5], [6], [7]]. This loss of efficient hepatic fatty acid oxidation leads to a constant flux of excess lipids into the mitochondria that causes an increase in ROS, which is common in individuals with NAFLD and nonalcoholic steatohepatitis (NASH) [8,9]. An adaptive increase of hepatic mitochondrial function that occurs early in the pathogenesis of NAFLD has been shown in individuals with obesity with insulin resistance and in ob/ob mice [8,9]. However, as NAFLD progresses, mitochondrial respiration becomes impaired leading to a reduction in efficient ATP production [8].

Cardiolipin (CL) is the signature phospholipid of the IMM that controls the efficiency of ATP synthesis. CL contains 3 glycerol molecules that are esterified to 4 fatty acyl chains with the most abundant species in the liver being tetralinoleoycardiolipin (L₄CL) [[9], [10], [11], [12], [13], [14], [15], [16], [17]]. This linoleic acid (LA)-rich CL species is crucial for creating an optimally functioning electron transport chain (ETC) and has been shown to promote energy utilization from fatty acid oxidation [14,[18], [19], [20]]. Remodeling of defective CL occurs when oxidized LA and other fatty acyl molecules are replaced with nonoxidized fatty acid molecules, usually LA in energetic tissues. However, in states of decreased LA, oxidative tissues can remodel these defective fatty acid side chains with varying fatty acids that can lead to changes in mitochondrial function [20]. For example, the replacement of LA by oleate (OA) has significant negative effects on both coupled respiration (ATP synthesis) and uncoupled respiration from fatty acids in isolated rat cardiac cells [20]. Having a low dietary fraction of LA promotes this substitution and may allow ROS to accumulate and result in damage to the IMM [13,17,19,21]. Compared with healthy controls, individuals with obesity-induced NASH have less LA content in CL despite an increase of total CL concentrations [18]. Although a shortage of LA-containing CL has been linked with NAFLD, we are unaware of any studies that have utilized dietary LA to restore hepatic LA content in CL and tested the effects on hepatic mitochondrial respiration.

Dietary LA influences the mitochondrial capacity in cardiac tissue in a preclinical model [22]. Although previous studies in rats have shown that dietary fat does influence hepatic content of L₄CL [21,23], whether dietary LA impacts mitochondrial function in liver is unknown. To determine the effect of dietary LA on hepatic mitochondrial activity and CL fatty acid side chain composition, mice were fed a diet made with a high-LA safflower oil (SO) or a diet made with lard (LD), which has a similar fat composition to a western style diet. Our goal was to determine if dietary LA under the metabolic stress of excess lipids would be able to maintain or increase mitochondrial function by maintaining hepatic L₄CL concentrations. The objectives of this study were to evaluate whether a diet rich in LA: 1) increases L₄CL and other LA-rich CL species, 2) improves mitochondrial function, and 3) impacts phospholipid fatty acid composition in liver tissue. We hypothesized that mice consuming the LA-rich diet, compared with mice fed a mixed fatty acid-composed diet, will have increased incorporation of LA into CL phospholipid, species, and higher mitochondrial respiration.

Methods

Animals, diets, and experimental design

The housing conditions for the mice were described previously [24]. Diet composition is provided in Supplemental Table 1, and the fatty acid composition of the diets in Supplemental Table 2. Body weight and food intake were measured every other day, except for 3 wk (Days 19–35) during the COVID-19 pandemic when food intake and body weight were measured every 3–4 days.

To measure mitochondrial function, a subset of mice (N = 9 per diet group) were killed by cervical dislocation on Days 101–130. The remaining mice (N = 3 per diet group) underwent cardiac puncture under isoflurane anesthesia before cervical dislocation on study Day 130. At necropsy, livers were collected and were flash-frozen in liquid nitrogen. Livers used in the mitochondrial assays were not frozen. Frozen tissues were weighed on dry ice and stored at −80°C until further analysis. All procedures were approved by The Ohio State University Institutional Animal Care and Use Committee (protocol # 2013A00000036-R2) and have therefore been performed in accordance with the ethical standards of the 1964 Declaration of Helsinki and its later amendments.

EchoMRI for body composition

To assess time-course changes in body composition of mice, Echo MRI (Houston,TX, USA) was used at 3 time points during the feeding period: Days 0, 65, and 100.

Glucose tolerance test

A glucose tolerance test (Day 91) was administered by moving mice to a quiet environment at 0700 h. Mice were transferred to new clean cages without food or nesting material, then deprived of food for 6 h before the procedure. A 2 g/kg glucose solution was administered via intraperitoneal injection and then whole blood was collected from the tail tip at 0, 15, 30, 45, 60, 90, and 120 min post IP injection using a portable glucometer (OneTouch Ultra Lifescan, Inc.).

Mitochondria isolation

Mitochondria were isolated from livers obtained after euthanasia as described previously [25]. In brief, liver was homogenized using an electronic tissue grinder fitted with a porcelain pestle (Eberach Co.) in samples at 4°C. Homogenates were centrifuged at 800 × g at 4°C for 10 min. The supernatant was filtered through multilayered cheesecloth and centrifuged at 8000 × g at 4° C for an additional 10 min. The resulting pellet was resuspended and centrifuged at 8000 × g at 4°C for 10 min and then resuspended a second time and centrifuged at 8000 × g at 4°C for 5 min. The relative quantity of mitochondria present in each sample was determined by Bradford assay (Bio-Rad Laboratories, Inc.).

Mitochondrial CL species

Lipids were extracted from the isolated hepatic mitochondria samples by the modified Bligh and Dyer method according to previously published methods with 1000 nmol tetramyristal CL as an internal standard (Avanti Polar Lipids) [26]. CL was quantified using previously published methods [27] with normal phase liquid chromatography coupled to electrospray ionization mass spectrometry in an API 4000 mass spectrometer (Sciex) [28]. Total CL was calculated by adding all the concentrations of CL species detected together.

Mitochondrial function

Respiration was measured from isolated hepatic mitochondria using electron coupling assay (EC) and electron flow assay (EF) (XFe96 Extracellular Flux Analyzer; Agilent Seahorse) as previously described [25,29,30]. Briefly, tissue-specific mitochondria were resuspended at optimal concentrations in EC or EF buffer comprised of MAS1 buffer (70 mM sucrose, 220 mM mannitol, 5 mM KH₂PO₄, 5 mM MgCl₂, 2 mM HEPES, 1 mM EGTA, pH 7.2). An EC buffer was supplemented with 10 mM succinate, 2 μM rotenone, and 0.5% bovine serum albumin (BSA), pH 7.2. An EF buffer comprised of MAS1 buffer was supplemented with 10 mM pyruvate, 2 mM malate, 4 μM FCCP, and 0.5% BSA, pH 7.2. Resuspended mitochondria (25 mL) were plated on Seahorse culture plates and centrifuged at 2000 × g at 4° C for 20 min. Mitochondria were visualized by bright field microscopy to verify density. The EC or EF assay buffer (155 mL) was added, and the mitochondria were incubated in a 37° C non-carbon dioxide incubator for 10 min before the assay was performed. An EC buffer without BSA was used to prepare optimized concentrations of ADP (port A), oligomycin (port B), FCCP (carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone) (port C), and antimycin A and rotenone (port D) for the EC assay. An EF buffer without BSA was used to prepare optimized concentrations of rotenone (port A), succinate (port B), antimycin A (port C), and TMPD (N, N, N′, N′-tetramethyl-p-phenylenediamine) (port D) for the EF assay. EC and EF assays were performed using the XFe96 Extracellular Flux Analyzer (Seahorse Bioscience).

Mitochondrial DNA (mtDNA) was measured after mitochondrial respiration measurements. The mtDNA isolation protocol was adapted from Baskin et al. [30]. In brief, the Seahorse assay buffer was aspirated from wells and replaced with Trizol (200 uL) to extract DNA from mitochondria samples (mtDNA). Chloroform (50 uL) was added to the mitochondria. The samples were vigorously mixed, and centrifuged for 15 min at 12,000 × g, 4° C. The aqueous layer was discarded. The extraction buffer (4 M guanidine thiocyanate, 50 mM sodium citrate, and 1 M Tris base) was added to the mtDNA and organic fractions. The samples were inverted several times. After a 10-min incubation at room temperature, samples were centrifuged for 30 min at 300 × g, 4° C. The upper phase containing mtDNA was collected and polyacryl carrier was added (Molecular Research Center PC152) and samples were inverted several times and incubated at room temperature for 5 min. mtDNA was pelleted by centrifuging at 12,000 × g for 5 min at 4° C. The pellet was washed 3 times with 70% ethanol, allowed pellets to air dry, and resuspended in a minimal volume of 8 mM NaOH and then vortexed. A mix of 1 M HEPES and 100 mM EDTA was added to each isolated mtDNA sample. mtDNA samples were incubated at room temperature overnight, and mtDNA concentrations were measured using a Nanodrop and mitochondrial gene expression was quantified by qPCR.

qPCR

qPCR was performed on all samples to quantify mtDNA according to vendor protocol with iTaq Universal Sybr green reagent (Bio-Rad). A total of 5 ng of mtDNA was used per sample for NADH dehydrogenase subunit 1 (mtND1; forward primer 5′-CCC ATT CGC GTT ATT CTT-3′ and reverse primer 5′-AAG TTG ATC GTA ACG GAA GC-3′), and for mitochondrial ribosomal RNA (mt16S; forward primer 5′-CCG CAA GGG AAA GAT GAA AGA C-3′ and reverse primer 5′-TCG TTT GGT TTC GGG GTT TC-3′). Ct values were compared between samples to confirm mitochondrial plating densities between wells after respiration measurements.

Mitochondrial protein analysis

Protein concentration of previously isolated mitochondrial extracts was determined by the bicinchoninic acid reaction using the BCA Protein Assay Kit (Pierce Chemical Co.). Total OXPHOS protein was measured in hepatic mitochondria samples. A total of 5 μg of protein from each sample was separated using SDS-PAGE with a 15% SDS-PAGE gel. Protein was transferred to a polyvinylidene fluoride membrane in transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol; Bio-Rad Laboratories). After drying the membrane for 1 h at room temperature, membranes were incubated in Revert 700 total protein stain (LI-COR Biosciences) and total protein (loading control) was detected using the LI-COR Odyssey Imager (LI-COR Biosciences). After washing, membranes were then blocked with 5% nonfat dry milk in Tris-buffered saline for 1 h and incubated on a shaker overnight at 4°C with primary antibody (1:500; Total OXPHOS; Abcam). After washing, membranes were incubated for 1 h with fluorophore-conjugated secondary antibody, and bands were detected using the LI-COR Odyssey Imager (LI-COR Biosciences). Densitometric analysis was carried out using the LI-COR Imager Software (LI-COR Biosciences).

Fatty acid composition of hepatic phospholipids classes

Total lipids were extracted from liver samples with 2:1 (vol:vol) chloroform: methanol [31]. For the analysis of total lipids and triglycerides (TG), the chloroform phase containing the lipids was transferred to a weighed test tube and dried under nitrogen gas at room temperature. The weight of the empty test tube weight was subtracted from the weight of the dried lipids and test tube to calculate the extracted total lipid. The extraction of TG was obtained using solid-phase extraction [27] and analyzed using an enzymatic colorimetric assay (Millipore Sigma).

The separation of phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine, phosphatidylglycerol (PG), and CL was accomplished using SPE cartridges from Agilent Technologies, Inc. according to Suzuki et al. [27]. The fatty acids of each phospholipid class were methylated at 76° overnight using 5% hydrochloric acid in methanol [28]. The resulting fatty acid methyl esters were extracted with hexane and analyzed via gas chromatography equipped with a 30-m Omegawax 320 fused silica capillary column (Supelco) helium carrier gas. The oven temperature started at 175°C and increased at a rate of 3°C/min until reaching 220°C. Retention times of samples were compared with standards for fatty acid methyl esters (Matreya, LLC and Nu-Check Prep Inc.). Fatty acids are reported as a percent of total identified [32].

Statistical analysis

Data are presented as the mean ± the SEM except fatty acid profile of the diet is presented as SD. Differences between LD and SO diet groups were analyzed using a 2-sample t-tests. Non-normal datasets were analyzed with the Mann–Whitney test. Data was tested for Gaussian distribution with the Shapiro–Wilk test and QQ plots. To test for any statistical outliers, we compared individual values to a cutoff range calculated as mean ± 3(SD). All statistical tests were 2 tailed and performed using STATA (StataCorp LLC) at the 5% significance level and data were graphed with GraphPad Prism (GraphPad Software).

Results

Body weight, food intake, body composition, and hepatic lipids

Cumulatively, neither body weight (Figure 1A) nor food intake (Figure 1B) differed between diet groups. In contrast, there were transient differences between food intake and body weight between Days 19 and 69. Body composition data showed that the LD diet group had a higher percent change in body weight and adipose mass on Day 65, but not at Day 100 (Figure 1C and D). Percent change in lean mass was not different between diet groups at any time points (Figure 1E). Neither the liver weight nor the hepatic concentrations of total lipids or TG were different between the 2 diets groups (Figure 1F–H).

FIGURE 1.

Mouse physiological parameters. (A) Body weight and (B) food intake were measured every 2 days, except for a brief period during the COVID-19 university shutdown (Days 14–35) where body weight was measured every 3–4 days. Body composition was measured using Echo MRI at Days 0, 65, and 100. (C) Percent change in body weight (BW), (D) Adipose mass and (E) Lean mass from Day 0. (F) Mean liver as percent of total BW. Hepatic total lipids (G) and triglycerides (H). Data presented as mean ± SEM. A–E, G and H were analyzed by 2-sample t-test. F was analyzed by Mann-Whitney test. N = 12 per diet group, for A–E, G and H. N = 7 per diet group for F. P < 0.05 (∗).

Fasting blood glucose or glucose disappearance during the glucose tolerance test levels as measured on Days 1, 85, and 91 were different between diet groups (Supplemental Figure 1I–M).

Mitochondrial CL species

CL species were calculated relative to the total CL detected from isolated mitochondria. The percentages of total CL of the 72 carbon CL species are listed in Figure 2, and all detected hepatic CL species are listed in Supplemental Table 3. Mice fed SO diet had significantly greater L₄CL compared with the LD diet group (Figure 2A and F; P 0.004). The LD diet group had significantly higher OA-containing CL species (e.g., L₃O₁CL, L₂O₂CL) compared with the SO diet group (Figure 2B–D P < 0.001). There were no differences between diet groups for the O₄CL species (Figure 2E). The SO diet group had higher relative quantities of the LA-rich 72 carbon CL species compared with the LD diet group that had higher relative quantities OA-rich 72 carbon CL species (Figure 2F).

FIGURE 2.

Dietary fat alters CL species in mouse liver. (A) L₄CL; (B) L₃OCL; (C) L₂O₂CL; (D) L₃OCL; (E) O₄CL; (F) % of 72C species. CL species of 72 carbons are presented as a percent of total 72 carbon species labeled by mass to charge ratio. The five 72 carbon species of CL are composed of LA and OA fatty acids. Data are presented as mean ± SEM. Data were analyzed by 2-Sample t-test. N = 5 mice per diet group. P < 0.05 (∗).

Mitochondrial function

To analyze any differences in mitochondrial respiratory activity, oxygen consumption rates (OCR) were measured using the electron coupling (EC) assay that measures respiration in coupled mitochondria through complexes II–V and the EF that measures sequential respiration through complexes I–V in uncoupled mitochondria (Figure 3A and B). Both diet groups had similar OCR levels throughout the EC assay (Figure 3A and C). There was an increase in basal OCR and another increase in OCR upon the addition of succinate in the mitochondria of SO diet-fed mice compared with the LD diet mitochondria in the EF assay (P values 0.03 and 0.01). No other significant differences in OCR were seen through the course of the EF assay (Figure 3B and D). OCR measurements were independent of a change of mitochondrial content as measured by mtDNA (Supplemental Figure 2). These findings suggest that mitochondrial function, not mitochondrial quantity, is increased in the livers of the SO diet group in the EF assay.

FIGURE 3.

Dietary fat impacts mitochondrial activity but not OXPHOS protein concentrations in mouse liver. (A & C) Electron coupled assay (EC); (B & D) Electron flow assay (EF); (E) Western blot image of complexes I–V. (F) Quantitation of densitometrey from OXPHOS protein expression in Western blots. Sample wells were normalized to total protein, as described in the methods. Data are presented as mean ± SEM. Panels C–D were analyzed by Mann-Whitney test. Panel F was analyzed by 2-sample t-test. N = 4–5 per diet group. P < 0.05 (∗).

To determine whether protein concentrations of the mitochondrial ETC complexes were altered between diets, a western blot was performed to measure OXPHOS protein concentrations in the hepatic mitochondria samples. No differences in protein concentrations of the ETC protein complexes were detected between the diet groups as measured by western blots (Figure 3E and F).

Fatty acid composition of hepatic phospholipid classes

Mice fed the SO diet group, compared with the LD diet group, had increased incorporation of LA into all phospholipids including CL and PG, a precursor phospholipid that can be converted into CL (Table 1; P < 0.05). In all phospholipids but PG, mice fed the SO diet had lower amount of OA and palmitic acid (PA) compared with the mice fed the LD diet. Interestingly, DHA was increased in all isolated hepatic phospholipid classes in the LD diet-fed mice compared with the SO diet-fed mice despite DHA being nondetectable amounts in both diets (Table 1; P < 0.05).

TABLE 1.

Effect of LD and SO diets on hepatic phospholipid fatty acid composition¹

CL
	LD	SO	P value
PA (C16:0)	19.9 ± 0.50	17.7 ± 0.50	0.0056
SA (C18:0)	18.9 ± 0.70	18.0 ± 0.60	0.299
OA (C18:1)	9.24 ± 0.65	4.77 ± 0.50	<0.0001
LA (C18:2)	35.0 ± 0.70	43.2 ± 0.90	<0.0001
ARA (C20:4)	7.56 ± 0.50	8.62 ± 0.47	0.135
DHA (22:6)	2.77 ± 0.10	0.74 ± 0.04	<0.0001
PG
	LD	SO	Pvalue
PA (C16:0)	18.2 ± 0.70	17.7 ± 0.50	0.592
SA (C18:0)	16.6 ± 0.50	16.9 ± 1.20	0.860
OA (C18:1)	8.92 ± 0.27	7.13 ± 1.58	0.277
LA (C18:2)	34.4 ± 0.60	49.4 ± 0.92	0.0002
ARA (C20:4)	7.56 ± 0.50	8.62 ± 0.47	0.660
DHA (22:6)	3.67 ± 0.37	0.87 ± 0.10	<0.0001
PE
	LD	SO	Pvalue
PA (C16:0)	17.1 ± 0.40	14.6 ± 0.24	<0.0001
SA (C18:0)	19.6 ± 0.50	21.54 ± 0.50	0.0132
OA (C18:1)	9.90 ± 0.51	7.90 ± 0.33	0.0030
LA (C18:2)	5.65 ± 0.48	10.4 ± 0.70	<0.0001
ARA (C20:4)	27.8 ± 0.60	28.6 ± 0.70	0.391
DHA (22:6)	13.3 ± 0.40	4.19 ± 0.15	<0.0001
PC
	LD	SO	Pvalue
PA (C16:0)	24.7 ± 0.20	21.8 ± 0.10	<0.0001
SA (C18:0)	15.1 ± 0.20	16.5 ± 0.30	0.00190
OA (C18:1)	8.31 ± 0.17	4.76 ± 0.13	<0.0001
LA (C18:2)	11.4 ± 0.40	16.5 ± 0.50	<0.0001
ARA (C20:4)	23.2 ± 0.30	26.4 ± 0.30	<0.0001
DHA (22:6)	9.12 ± 0.20	2.72 ± 0.09	<0.0001

Fatty acid composition (%) of phospholipid fractions. (CL) cardiolipin; (PG) phosphatidylglycerol; (PE) phosphatidylethanolamine, (PC) phosphatidylcholine. LD, lard; SO, satflower oil. Data represented as mean ± SEM. Data were analyzed by 2-sample t-test. N = 12 per diet group. P < 0.05 (∗).

¹Fatty acids not included: myrsitic acid (C14:0), palmoleatic acid C16:1n7, vaccenic acid (C18:1n7), γ-linolenic acid (C18:3n6), icosanoic acid (C20:0), cetoleic acid (C20:1n9), eicosadienoic acid (C20:2n6), dihomo-gamma-linolenic acid (C20:3n6), behenic acid (C20:0), adrenic acid (C22:4n6), osbond acid (22:5n6), docosapentaenoic acid (C22:3n), lignoceric acid (C24:0), nervonic acid (C24:1n9).

Discussion

Dietary fat has been demonstrated to impact CL species and respiratory function in mitochondrial rich tissues, including cardiac muscle and liver, in experimental animal models of heart failure and obesity, which are associated with deficits of L₄CL [23,33,34]. Our finding that dietary LA increased L₄CL in liver of nonobese mice fed a moderate-level fat diet suggests that even in a nondiseased condition, dietary LA content impacts the mitochondrial lipidome. In a recent clinical study of healthy adults, we demonstrated that fortifying diets with <2 teaspoons of LA-rich oil per day increased L₄CL in peripheral blood mononuclear cells (PBMCs) [35]. Whether the increase of L₄CL of PBMCs reflects changes in mitochondrial rich tissues such as liver, muscle, or brown adipose tissue of humans is unknown. Although the use of dietary LA in these preclinical models of dysfunctional cardiac mitochondria has stressed the importance of L₄CL species, this yet remains to be seen in the hepatic tissues.

Although body weight and cumulative food intake were temporarily different between diet groups (between Days 19 and 69), there were no differences between these 2 measurements by Day 69. It is not clear if the temporary difference of energy balance between the 2 groups was due to infrequent diet feeding (during restricted access to the vivarium during the COVID-19 pandemic in March 2020) or due to other conditional factors (e.g., noise disruptions). There were no longer differences between the diet groups in body weight, body composition, glucose tolerance, or hepatic lipid content. The mice in this study were not obese nor did they develop NAFLD (hepatic fat remained <5% liver weight). It is common in individuals with NAFLD to have other comorbidities involved with metabolic dysfunction such as obesity, type 2 diabetes, and dyslipidemia [36,37]. Whether dietary fortification with LA can increase LA-rich CL species and mitochondrial respiration, however, in a mouse model for NAFLD is not known but is worth investigating in future studies.

The mice fed the SO diet had increased L₄CL species and decreased OA-rich CL species compared with the mice fed the LD diet. For the first time, we show that mice fed an LA-fortified diet (SO compared with LD diet) had higher respiration in mitochondria isolated from minced tissue ex vivo as measured by XFe Seahorse assay. In previous studies, we showed that the replacement of LA by OA has significant negative effects on both coupled respiration (ATP synthesis) and uncoupled respiration from fatty acids in isolated rat cardiac cells [20]. The 4 electron transfer complexes of the ETC (complexes I–IV) with the fifth complex (ATP synthase) work together in the IMM to form supercomplexes that generate ATP by providing efficient respiration [19,38,39]. LA-rich CL may provide optimal symmetry and scaffolding for coordinating the structural support, formation, stability, and activity of these supercomplexes [11,38]. CL has also been found to help structure cristae, and lack of CL can result in dysfunctional cristae membranes [39]. The L₄CL species is typically the dominant species in mitochondrial dense tissues and may be vital for both proper cristae formation and supercomplex formation resulting in optimal respiration.

The mitochondria isolated from livers of the mice fed the SO diet group showed no differences in OCR compared with the LD diet group in the EC assay. We reasoned that there were no differences in OCR between the 2 diet groups because neither group was metabolically impaired. We anticipate that enrichment of LA content in CL of the SO diet group in mice that are metabolically impaired, or with fatty liver, may help maintain or increase hepatic OCR (compared with OCR in liver mitochondria of LD diet group). In rats fed a choline deficient diet to induce fatty liver (compared with control diet mice without fatty liver), complex I activity was impaired [3]. In a different study where mice were fed a high-fat diet (60% kcal from fat), complex I–III activity (measured as NADH oxidation or succinate oxidation) was impaired compared with the lean controls [34]. In LDL receptor knockout mice with diet-induced NAFLD from a western diet (45% kcal from fat and 35% kcal from sucrose), there was a reduction in complex V protein concentrations and a 50% decrease in hepatic ATP content [40]. Because mitochondrial dysfunction may be an early event in the development of NAFLD, it is not surprising that there were no differences in OCR within the EC assay between the 2 diet groups in our study.

Mitochondria isolated from the livers of mice fed the SO diet exhibited elevated basal mitochondrial respiration in the EF assay compared with respiration in the LD diet group. We reported similar results in brown adipose tissue where uncoupled respiration was elevated at baseline and after succinate treatment under the same dietary conditions [24]. Uncoupled mitochondria perform inefficient ATP synthesis by allowing proton leakage across the IMM, requiring the ETC to work requiring more energy to synthesize the same amount of ATP [41]. Although this mechanism is well studied in brown adipose tissue, uncoupled mitochondria may provide relief from lipid toxicity seen in individuals with NAFLD by increasing beta oxidation of lipids to produce adequate ATP. In 2013, a group of researchers found that rats with diet-induced hepatic steatosis (fed a 60% kcal from fat, SO diet supplemented with high sucrose drinking water) treated with mitochondria uncoupler, 2,4-dinitrophenol methyl derivative, reduced liver triglyceride content by ∼40% independent of changes in body weight [42]. Rats also had lower fasting glucose, triglyceride, and insulin concentrations compared with vehicle-treated animals [42]. Although our mice did not have hepatic steatosis as per our assessment, our data may suggest that the SO diet uncoupled mitochondria are able to process excess fatty acids than the LD diet-fed mice.

Our data show from the EF assay that the SO diet group mice have increased OCR from the addition of succinate. Succinate is an intermediate in the citric acid cycle and is a substrate to help transfer electrons from FADH2 to ubiquinol, in both cases through succinate dehydrogenase (SDH), or complex II in the ETC [43]. Our findings support the role for dietary SO in enhancing SDH activity. The increase in plasma succinate has been associated with fatty liver conditions in both preclinical and clinical models [[44], [45], [46]]. In a diet-induced murine model of NAFLD mice eating a western diet (20% fat, glucose + sucrose-enriched water) a decrease in hepatic SDH activity and increase in hepatic succinate [44]. This same group theorized that the decrease in SDH activity may be to reduce ROS generation, as they also reported no differences in oxidative tissue damage in western diet-fed mice to controls [44]. As SDH participates in two very important metabolic pathways, this reduction in metabolic flexibility in the mitochondria could lead to problems with ATP generation. As our mice in each diet group are considered metabolically normal, perhaps our LA-fed mice could have lower amounts of hepatic or plasma succinate. Conducting this in dietary-induced NAFLD mice could be an interesting experiment to determine if feeding dietary LA could decrease hepatic succinate while maintaining the capacity for increased OCR in mitochondria. Although we saw no changes in liver lipid content between our diet groups, the SO diet-fed group still had increased baseline OCR and succinate-driven OCR compared with LD diet-fed group in our EF assay suggesting that even in healthy mice dietary fat quality does influence hepatic mitochondrial respiration.

The phospholipids, PG, PE, and PC are involved with de novo synthesis of CL as well as CL remodeling. Each phospholipid fraction exhibited increased LA in response to the SO diet-fed mice (compared with concentrations in phospholipids of livers from LD diet-fed mice). The fortification of PC and PE with LA is important because LA-rich CL can be generated by a set of remodeling enzymes such as tafazzin, monolysocardiolipin acyltransferase, or Alcat1. However, Alcat1 lacks the specificity for LA or other fatty acyls for CL remodeling [47]. Remodeling of CL by tafazzin is accomplished by the transfer of fatty acyl chains from both PC and PE to CL [11,14,17]. Several studies suggest that tafazzin and ALCAT1 protein concentrations may be influenced by diet-induced changes or metabolic status as well as diet quality. After 8 wk of high-fat diet, Alcat1 knockout mice had attenuated weight gain and lower blood glucose concentrations compared with wild-type mice [48]. In another study, obese mice had a higher hepatic tafazzin expression with no differences in the Alcat1 gene expression compared with control diet-fed mice [34]. In a study with db/db mice, mRNA expression and activity of Alcat1 was upregulated by the onset of type 2 diabetes [48]. Mice in this study were not obese nor metabolically impaired, but changes in the gene expression of either of these remodeling enzymes could be a possible explanation for differences infatty acid content found in CL. It would be useful to measure the impact of dietary LA in a model of NALFD and/or metabolic impairment, along with the expression of CL remodeling enzymes. Dietary fat and overall quality of diet may not only affect LA-rich CL species by reducing LA availability but may also alter the gene expression and the resultant protein concentrations of key CL remodeling enzymes. In contrast, OA and PA were increased in the phospholipid fractions of PE and PC, but not PG, in the LD diet-fed mice. As mentioned earlier, OA replacement with LA has been shown to alter fatty acid oxidation [20]. The increase of OA availability in PC and PE could lead to the alteration of CL species and a decrease in the respiration of uncoupled mitochondria in the LD diet group. It is notable that arachidonate, a desaturated and elongated metabolite of LA, was significantly increased only in PC but not in other phospholipids. The low rate of conversion of LA is consistent with many other studies showing that diets enriched with LA have very little predictive impact on concentrations of arachidonate [49].

PG is an important precursor for nascent CL synthesis via CL synthase [50]. Because PG is used to form nascent CL, which is then remodeled to form LA-rich CL species, higher CL-associated LA species in livers of the SO diet group was not unexpected. A study in mice with defective PG remodeling found that a decrease in LA-rich PG led to a significant decrease of L₄CL species in liver mitochondria [51]. Lysophosphatidylglycerol acyltransferase 1 is a key enzyme in catalyzing the acylation of lysophosphatidylglycerol to PG that allows for PG remodeling [51]. Because the LD diet had less LA in the PG, this could also explain why we see a decrease of LA in their CL concentrations as well. As this is another remodeling enzyme that can ultimately lead to a decrease of LA enriched CL. These data suggest that a depletion of LA in PG may be disruptive to proper L₄CL formation and mitochondrial function.

The higher DHA content of the liver phospholipids of mice fed the LD diet, compared with the SO diet, was unexpected because both diets had undetectable amounts of DHA. Pathologic CL remodeling has been linked with increased DHA content and is disruptive to normal IMM function [52,53]. Additionally, increased DHA content of CL in the mouse heart has been linked to reduced complex I, IV, V, and I + III activity [53]. Previous studies have shown that dietary LA rescued the DHA suppression of mitochondrial respiration [22,53]. These data seem to indicate that the L₄CL species is vital for optimal mitochondrial respiration, and the lack of dietary LA can have consequences on mitochondrial respiration. In a study focusing on the influence of dietary oils following 1 month of calorie restriction in mice, the calorie restriction group fed a soybean oil-enriched diet, containing ∼55% LA, had higher concentrations of CL-associated LA, and displayed higher complex I and II activity in the liver (compared with the calorie restriction group fed a fish oil-enriched diet) [54]. Importantly, the calorie-restriction group-fed fish oil diet had ∼3 times the percent of DHA in CL compared with all other groups [54]. Another group found that DHA in diet-induced obesity fed mice (60 kcal% from fat) had increased DHA incorporation into PC and PE fractions compared with their lean controls [23]. The increase of DHA into major phospholipids such as PC, PE, PG, and CL may be indicative of a phospholipid profile associated with metabolic impairment. In a study using mouse C2C12 myoblasts cell lines transfected with ALCAT1, it was found that these cell lines overexpressing ALCAT1 had reduced LA-rich CL species and higher DHA related species in CL [52]. Our data corroborate a link between higher DHA content in phospholipids, especially CL, with unfavorable mitochondrial function.

In summary, mice fed a diet fortified with an LA-rich oil had higher hepatic L₄CL (compared with mice fed an LD diet) and had higher baseline and succinate-driven mitochondrial respiration in the EF assay. Because the mice in this study did not have obesity, dysregulated glucose tolerance or fatty livers, we were unable to assess whether a diet fortified with LA-rich oil alters the progression of NAFLD. Because an early adaptive response of hepatic mitochondria may be to improve beta oxidation due to dietary lipid overload, the stage of severity of fatty livers may alter the effects of LA on mitochondrial respiration that we report here. Associative studies show that higher dietary or body concentrations of LA are related with reduced hepatic fat content. Thus, it will be worthwhile to examine whether dietary LA can reduce the progression of NAFLD in a randomized controlled trial.

Author contributions

The authors’ responsibilities were as follows – MAB: project conception; CAM, DBS, RMC, AA, GS, MAB: research plan; CAM, MAB: study oversight; CAM, DBS, RMC, AA, GS, KKB, DAB: data collection; CAM, DBS, RMC, AA, GS, KKB, AN: data analysis; CAM, DBS, MAB: writing; CAM, DBS, RMC, AA, GS, KKB, MAB: editing; CAM, DBS, RMC, AA, KKB, GS, MAB: final content; and all authors: read and approved the final manuscript.

Conflict of interest

MAB has a research grant funded by the United Soybean Board, and serves on the Board of Directors for the American Society for Nutrition. All other authors report no conflicts of interest.

Funding

This project was supported by the Ohio Agriculture Research and Development Center (Wooster, OH).

Data availability

Data described in the manuscript, code book, and analytic code will be made available upon request pending application, approval, and payment.

Appendix A. Supplementary data

The following is the Supplementary data to this article: