Introduction

Mammalian aging is associated with a gradual decline in body condition that negatively impacts physical capacity, fertility, and quality of life. Malnutrition and weight loss can accelerate these deleterious effects, particularly in species with high physiological demands1,2 Equine aging is associated with decrements in skeletal muscle and reproductive function that are similar to those observed in humans, and older horses (> 15 years of age) are now estimated to comprise up to one-third of the global equine population3,4,5 Horses may also suffer from sub-optimal nutrition and under-recognized signs of age-related decline, particularly when compounded by poor dental health and gastrointestinal problems4 As in humans, dietary supplementation with vegetable oils is commonly recommended for older horses to prevent weight loss, as they are calorie dense and easy to digest6,7 The majority of commercially available dietary oils are derived from corn, soybeans, or flax seeds containing predominantly polyunsaturated fatty acids (PUFAs), which are generally recommended over saturated fats for optimal health benefits based largely on human data8 However, the biological impacts of dietary PUFA supplementation in horses are less clear, particularly in the setting of advanced age.

Dietary PUFAs are further classified as omega-3 (n-3) or omega-6 (n-6) fatty acids based on the position of their first double bond counting from the terminal methyl group9 Both classes are metabolized by the same desaturase, elongase and oxygenase enzymes to generate distinct long-chain PUFA derivatives with diverse and often opposing biological functions. In the equine industry, dietary oils containing predominantly n-6 linoleic acid (LA) extracted from corn and soybeans are most frequently used because they are easily accessible and affordable10 This parallels a much greater intake of n-6 compared to n-3 PUFA in the modern human diet, which is in contrast to the more balanced intake of n-6 LA and n-3 α-linolenic acid (ALA) in the ancestral diets of humans and equids11,12 A greater proportion of dietary n-6 PUFA intake has been hypothesized to favor greater inflammatory signaling and increased risk of age-related diseases in humans,13 prompting recommendations for dietary supplementation with n-3 PUFA given their putative anti-inflammatory effects and diverse health benefits14 However, the biological impacts of n-3 and n-6 PUFA are diverse and vary widely across species, physiological systems and disease risk profiles, leading to ongoing debate over their optimal ratio in the diet and its relevance to long-term health outcomes15,16.

Age-related decrements in physical performance and fertility have been associated with reduced mitochondrial capacity in skeletal muscle and follicular cells of horses,17,18,19 and humans20,21 Recently, we found that supplementing the diets of aged mares with n-3 PUFA in combination with a proprietary blend of antioxidants, L-carnitine and other supportive nutrients for 8 weeks improved oocyte metabolic capacity, granulosa cell mitochondrial function, and fertility outcomes22 However, the independent effects of dietary PUFA (in the absence of supportive supplements) were not investigated, nor whether the impacts of these dietary interventions extended to other tissues and cell types. Consequently, the aim of the present study was to broadly characterize the effects of dietary n-3 and n-6 PUFA supplementation on platelet, skeletal muscle and follicular mitochondrial function in aged mares in the presence and absence of supportive dietary nutrients. Impacts on the fatty acid composition of serum, follicular fluid, muscle and granulosa cell phospholipids was also investigated, along with markers of inflammation and skeletal muscle oxidative stress23.

Materials and methods

Animals and experimental design

All procedures and experimental protocols on live animals were approved by the Institutional Animal Care and Use Committee at Colorado State University under protocol number 1554 and are reported in accordance with ARRIVE guidelines. The study was done using mares of light-horse breeding (not draft or pony mares). All mares were leased or owned by CSU for research; therefore, private owner consent was not applicable. Cyclic, non-lactating light-horse mares (n = 12, 19–27 years of age) were housed in adjacent dry lots with sheds, salt blocks, and water provided ad libitum. All animals were fed a base diet of grass/alfalfa mix hay at approximately 2% body weight daily with 57 g of a commercial vitamin and mineral forage balancer (Forage Balancer, Purina® Free Balance® 12:12, Purina Animal Nutrition, Gray Summit, MO, USA). After at least 4 weeks on the base diet, baseline measurements and samples were collected as described below, and mares were divided into two age-matched groups randomly assigned to receive dietary PUFA supplementation as 60 mL of flaxseed oil (N3; 57% ALA, 17% LA; n = 6, 22.5 ± 1.2 year) or corn oil (N6; 59% LA, < 1% ALA; n = 6, 23.2 ± 1.1 year) mixed in 227 g of alfalfa pellets once daily in the morning for the duration of the study. Oils were fortified with antioxidants (d-alpha tocopherol acetate and ascorbyl palmitate) and stored at room temperature and away from direct heat or sunlight to prevent auto-oxidation and preserve shelf life24 Following 6 weeks of the N3 or N6 diets, samples were collected to evaluate the independent impacts of dietary PUFA enrichment on the outcomes described below, and all mares began receiving additional dietary supplementation (DS) designed to support equine metabolic and gastrointestinal health and formulated to support cellular metabolism, including vitamins, trace minerals, amino acids, L-carnitine, antioxidants, prebiotics, and probiotics (187 g daily; Platinum Performance Inc., Buellton, CA). A complete fatty acid analysis of the N6 and N3 oils, and full list of DS ingredients are provided in the Supplemental Information. Samples were collected again following 6 weeks on N3 or N6 diets with the additional DS for final measurement of the outcomes described below. Morphometric assessments were performed at 2-week intervals and included BCS (1–9)25 and cresty neck score (0–5)26.

Sample collections

Blood, ovarian follicle contents, and skeletal muscle biopsies were collected at baseline, following 6 weeks of the N3 or N6 diets, and following an additional 6 weeks on N3 or N6 diets with the DS as described above. All methods were carried out in accordance with relevant guidelines and regulations. Blood was collected by jugular venipuncture and centrifuged at 3500 x g for isolation of serum. Serum was frozen at -80 °C for future biochemical analyses. Isolation of platelets was conducted according to Abcam protocol (Abcam, Cambridge, UK)27 Briefly, blood was collected into 60mL syringes containing 10% citrate-phosphate-dextrose buffer. Samples were transferred into tubes and centrifuged at 200 x g for 20 min to obtain platelet-rich plasma (PRP). PRP was removed from the tubes, diluted 1:1 with HEPES-buffered solution containing 1µM prostaglandin I2 (Sigma-Aldrich, St. Louis, MO), and centrifuged at 100 x g for 20 min to remove blood cells. The supernatant was removed, centrifuged at 800 x g for 10 min, and the resulting platelets were resuspended in Tyrode’s buffer containing 5 mM glucose, 3 mg/mL bovine serum albumin, and 1 µM prostaglandin I2. Platelets were maintained on ice until preparation for respirometry experiments.

Histrelin (0.5 mg IM; Doc Lane, Lexington, KY) was administered to induce follicle and oocyte maturation in vivo when the dominant follicle exceeded 35 mm and endometrial edema was revealed by ultrasonography. Cumulus oocyte complexes, granulosa cells, and follicular fluid were collected by ultrasound-guided follicular aspirations of preovulatory follicles at 20 ±2 h after induction as previously described28 Briefly, the follicular wall was punctured, and follicular fluid (5–10 mL) was collected from the center of the follicular antrum, to avoid blood and cell contamination, into a tube. The follicle was then flushed with Flush Medium (Vigro Complete Flush Solution, Vetoquinol, Fort Worth, TX, USA) with 10 USP units/ml heparin (Sagent, Schaumburg, IL) at 38.5 °C, and the aspirate was collected in a sterile bottle. Follicular fluid samples were frozen at -80 °C for later biochemical analyses. Sheets of granulosa cells were manually collected from the aspirate, washed in Flush Medium, centrifuged at 800 x g for 5 min, treated with red blood cell lysis buffer (Roche, Basel, Switzerland) for 1 min, and then suspended in Flush Medium for later respirometry experiments or frozen for biochemical analyses. Cumulus oocyte complexes were identified using a stereomicroscope with a warming plate to maintain approximate body temperature. Sections of cumulus cells were cut using needle tips from the complex, prior to placing the remaining cumulus oocyte complex in hyaluronidase (80 U/ml) for a few minutes prior to pipetting to separate the remaining cumulus cells from the oocyte. Isolated cumulus cells were frozen at -80 °C for future biochemical analyses, and denuded oocytes were preserved in MOPS-buffered medium (G-MOPS™; Vitrolife, Gothenburg, Sweden) and maintained at 4 °C until metabolic flux analyses described below.

Punch biopsies of the trapezius muscle were immediately rinsed in BIOPS preservation buffer containing (in mM) 10 Ca-EGTA (0.1 M free calcium), 20 imidazole, 20 taurine, 50 K-MES, 0.5 DTT, 6.56 MgCl2, 5.77 ATP, and 15 phosphocreatine (pH 7.1 with KOH). Approximately 100 mg muscle samples were snap frozen in liquid nitrogen for later analyses, while the remaining sample was transferred to fresh BIOPS buffer on ice until preparation for high-resolution respirometry.

Serum cytokine immunoassay

Serum samples were assayed for various cytokines using a bead-based multiplex assay at an external laboratory (Cornell Animal Health and Diagnostic Center, Ithaca, NY). Samples containing equine cytokines were mixed with anti-cytokine beads and incubated, followed by detection using fluorochrome-conjugated antibodies. Cytokine values were calculated based on five standards included in the assay. Methodology is described in detail elsewhere29.

Analysis of phospholipid fatty acid composition

Total phospholipids were extracted from skeletal muscle biopsies, granulosa cells, follicular fluid, and serum for determination of fatty acid composition by gas chromatography essentially as previously described30 Briefly, 10 mg of muscle was homogenized in 800 µL of precooled methanol in a glass homogenizer, while 50 µL of freeze-thawed granulosa cell lysates, follicular fluid, or serum were added directly to 600 µL of precooled methanol in a glass tube. After vortexing samples for 30 s, tubes were centrifuged at 900 × g for 5 min to pellet the non-phospholipid fraction. The supernatant (containing phospholipids) was transferred to a fresh glass tube, to which 25 µL of methoxide solution was added to synthesize methyl esters from hydrolyzed phospholipids. The reaction was stopped after 3 min by adding 75 µL of methanolic HCl, and fatty acid methyl esters (FAMEs) were extracted by adding 700 µL of hexane, transferring the upper hexane layer to 2 mL chromatography vials. FAMEs were dried under nitrogen flow and resuspended in hexane for gas chromatography (GC) analysis using an Agilent Technologies DB-225 30 m x 0.250 mm x 0.25 μm column (model 122–2232, J&W Scientific) on an Agilent 6890 Series Gas Chromatograph with a flame ionization detector30 A fatty acid standard mixture containing 23 FAMEs ranging from 16:0 to 24:1 was injected into the GC at 1 mg/mL every 6–7 runs to recalibrate the retention times and order of fatty acid peaks on the chromatograph, enabling relative quantitation of all the long-chain fatty acids commonly found in mammalian phospholipids expressed as a percentage of total phospholipid fatty acids.

High-resolution fluo-respirometry

Rates of cellular respiration and H2O2 release were measured in platelets, granulosa cells and skeletal muscle fibers by high-resolution fluo-respirometry using an Oroboros Oxygraph system with an O2k-Fluo LED2-Module (Oroboros Instruments, Innsbruck, Austria) essentially as previously described31 Platelets and granulosa cells were freshly isolated as described above and added directly to the oxygraph chamber containing 2 mL of mitochondrial respiration medium (MiR05) composed of (in mM) 0.5 EGTA, 3 MgCl2 hexahydrate, 60 lactobionic acid, 20 taurine, 10 KH2PO4, 20 HEPES, 110 sucrose, and 0.1% fatty acid-free bovine serum albumin (pH 7.1 with KOH) at 37 °C with room air-saturated oxygen (~ 160 µM). Following standardized instrumental and chemical background calibrations, the oxygen consumption rate (OCR) of intact platelets and granulosa cells was measured to determine basal rates of cellular oxidative metabolism supported by endogenous substrates. Cells were then permeabilized with 5 µg/mL digitonin to enable free diffusion of exogenous metabolic substrates to mitochondria without compromising inner membrane integrity32 Muscle biopsies were gently teased in BIOPS on ice and permeabilized with 50 µg/mL saponin before being added to the oxygraph chamber as described above, but with ~ 400 µM oxygen to overcome diffusion limitations of respirometry in permeabilized muscle fibers31.

Following permeabilization, 1 mM malate and 0.05 mM palmitoylcarnitine were added to samples in the oxygraph chamber to assess the capacity for non-coupled “LEAK” respiration supported by fatty acid oxidation (facilitated by proton leak across the inner mitochondrial membrane), followed by the addition of 2.5 mM ADP to generate the oxidative phosphorylation (OXPHOS)-linked OCR. The extent of respiratory control by ADP was expressed as the OXPHOS coupling control factor [1-(LEAK/OXPHOS)] for fatty acid oxidation, where a maximum value of 1.0 represents fully coupled mitochondria (100% control of respiration by ADP), and 0 represents fully uncoupled mitochondria (0% respiratory control by ADP). Finally, the maximal OXPHOS-linked OCR was determined by titrating 5 mM pyruvate and 10 mM succinate to the oxygraph chamber to maximize rates of electron supply from carbohydrate and fatty acid oxidation and a fully reconstituted citric acid cycle33.

The net rate of mitochondrial H2O2 release (JH2O2) was measured during basal and maximal OXPHOS-linked OCR states by monitoring the accumulation of chamber resorufin (Ex/Em 571/585 nm), the stable fluorescent product of 1:1 oxidation of 5 µM Amplex UltraRed by H2O2 in the presence of 1 U/mL horseradish peroxidase31,34 Sample OCR and JH2O2 are expressed relative to mg wet weight of muscle fiber bundles and total protein content of platelet or granulosa cells added to the oxygraph chamber determined by bicinchoninic acid (BCA) assay (Thermo Scientific, #23225).

Oocyte metabolic flux analysis

Basal (non-stimulated) rates of single oocyte oxygen consumption (OCR; reflecting mitochondrial respiration) and extracellular acidification (ECAR; representing glycolytic flux) were measured using a custom-built metabolic micro-sensor as previously described35,36 Briefly, single denuded oocytes were gradually warmed to 38.5 °C before being transferred to the sensor micro-chamber containing 180 µL of G-MOPS maintained at the same temperature within hours of follicular aspiration. Oocytes were positioned directly over the working electrodes of a pre-calibrated amperometric oxygen sensor and potentiometric pH sensor to obtain stable measurements of OCR and ECAR, respectively.

Protein immunoblotting

Frozen muscle samples (~ 50 mg) were homogenized in lysis buffer containing (in mM) 150 NaCl, 1 EDTA, 1 EGTA, 5 sodium pyrophosphate, 1 sodium orthovanadate, 20 sodium fluoride added to Mammalian Protein Extraction Reagent (Pierce Cat# 78501) with supplemental Protease Inhibitor Cocktail (Sigma Aldrich Cat# P8340). Protein concentrations of 10,000 × g homogenate supernatants were detected by BCA assay. Extracted proteins (30 µg) were electrophoresed on 4–12% Bis-Tris gels and transferred to PVDF membranes, then blocked for 1 h at room temperature with 5% non-fat milk before incubating with the following primary antibodies (1:1000 dilution) overnight at 4 C°: Cu/ZN SOD (SOD101, StressGen, Victoria, BC), Mn SOD (SOD111, StressGen, Victoria, BC), NOX4 (PA5-72816, Thermo Scientific, Waltham, MA), 4-Hydroxynonenal (MA527570, Thermo Scientific, Waltham, MA), Malondialdehyde (MA527560, Thermo Scientific, Waltham, MA) Secondary antibodies including Goat anti-Rabbit IgG (AB6721, Abcam, Waltham, MA) and Goat anti-Mouse IgG (AB6789, Abcam, Waltham, MA) were used at a 1:3000 dilution for 1 h at room temperature. Blotted proteins were imaged using SuperSignal West Dura Extended Duration Substrate (Lot#VL314742) and a UVP ChemStudio blot imager (Analytik Jena, Germany), normalizing band densities to total protein staining of a 80–100 kDa range of bands near the target protein in each lane using AmidoBlack (Sigma A8181), quantified using ImageJ software (NIH; Version 1.51; https://imagej.net/ij/index.html).

Statistical analysis

Data from all experiments are presented as group means ± SEM. Following confirmation of normal data distributions using the Shapiro-Wilk test, group comparisons were first performed by one-way ANOVA with repeated measures for each PUFA diet group at baseline, 6 weeks post-diet, and following an additional 6 weeks post-diet + DS. Following a significant ANOVA, individual group comparisons were evaluated by Tukey’s HSD test for pairwise group comparisons of fatty acid composition and immunolotting data, and by Fisher’s LSD test for pairwise comparisons of fluo-respirometry data for greater statistical power given greater within- and across-group variability. Impacts of the PUFA diets were also assessed by paired t-tests on combined diet and + DS data compared to baseline from each individual, and by independent samples t-test for comparisons between diets at the same timepoint. A P value of < 0.05 was considered significant for all statistical analyses. All statistical analyses were performed using GraphPad Prism software (version 10; https://www.graphpad.com/).

Results

Animal morphometrics and serum cytokines

Impacts of the dietary interventions on animal condition and serum cytokines are summarized in Table S1. There were no significant effects of PUFA supplementation or addition of the dietary supplement (+ DS) on animal body weights or body condition scores, the latter remaining within the normal range of 4–6 for healthy mares across all groups and time points25 Similarly, cresty Neck Scores26 averaged within the normal range of 1.5–1.7 across groups and time points, with the exception of a significant decrease following 6 weeks on the N3 diet that increased following the addition of the DS (P < 0.05 by Tukey’s HSD test). Serum cytokines levels were highly variable across individuals and groups throughout the study, with significant differences being detected only in tumor necrosis factor-alpha (TNF-α), which increased relative to baseline levels following 6 weeks on the N3 diet (P = 0.043) and following 6 weeks of the N6 diet (P = 0.007) by Tukey’s HSD test.

PUFA diets impact serum, muscle, and follicular phospholipid fatty acid composition

The fatty acid composition of phospholipids extracted from serum (Table S2), follicular fluid (Table S3), skeletal muscle (Table S4), and granulosa cells (Table S5) was investigated at baseline (BL), following 6 weeks of PUFA supplementation (6 weeks), and following 6 additional weeks with the supportive dietary supplements (DS). Overall, both PUFA diets tended to increase the total phospholipid PUFA content and decrease monounsaturated fatty acids (MUFA) across all sample types, with variable and comparatively minor effects on total saturated fatty acids. Impacts of the N6 diet on phospholipid PUFA composition were generally more robust than the N3 diet across samples, and effects of both diets tended to be augmented or unaffected by six additional weeks of feeding with the DS. Notable exceptions were a suppression of ALA levels in follicular fluid in N3-fed mares following the DS (Table S3), and a marked suppression of n-6 arachidonic acid (ARA) with reciprocal increases in long-chain n-3 PUFA in granulosa cells of N3-fed mares following addition of the DS (Table S5). When combining the 6 week and + DS data to improve statistical power for detecting main effects of the PUFA diets, the N6 diet significantly increased proportions of LA (P = 0.0008–0.04 by paired t-test) and total n-6 PUFA (P = 0.0007–0.019 by paired t-test) in all four sampling sites, with trends for reciprocal decreases in ALA and/or total n-3 PUFA that only reached statistical significance in serum (P = 0.015 and 0.002 by paired t-test, respectively). Conversely, the N3 diet tended to increase proportions of ALA and/or total n-3 PUFA across samples, but this only reached statistical significance for total n-3 PUFA versus baseline in granulosa cells (P = 0.0005 by paired t-test). Collectively, these trends led to a significantly greater n-6/n-3 PUFA ratio in the N6 versus N3 diet groups (P < 0.05 by independent samples t-tests) across all sample types (Fig. 1). Taken together, these results demonstrate effects of the N6 and N3 diets on mare phospholipid composition that are generally consistent with their fatty acid composition across multiple biological compartments.

Fig. 1
figure 1

Polyunsaturated fatty acid (PUFA) composition in various tissues following dietary oil feeding. The ratio of omega-3 to omega-6 PUFA was determined in serum (A), follicular fluid (B), skeletal (trapezius) muscle tissue (C), and granulosa cells (D) prior to (BL) and following 6 weeks of corn (N6) or flaxseed (N3) diets, and following 6 additional weeks with the added dietary supplement (+ DS). Follicular fluid and granulosa cells were collected from preovulatory dominant follicles 24 ± 2 h after maturation induction and skeletal muscle biopsies were obtained from the trapezius muscle. Graphs represent mean ± SEM of N = 6 mares per group. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Tukey’s HSD test. Noted P values indicate significant differences in the combined 6 wk and + DS time points between N6 and N3 cohorts by independent sample t-tests at P < 0.05.

Full size image

Cell type-specific effects of dietary interventions on mitochondrial bioenergetics

Impacts of the dietary interventions on cellular mitochondrial function varied according to the cell type being studied and outcome examined. In platelets, the N6 and N3 diets both tended to decrease basal OCR and fatty acid-supported LEAK and OXPHOS states (Figs. 2A-C), and favored an increase in OXPHOS coupling control in N6 (P = 0.009 by Fisher’s LSD test) and N3 after the DS (P = 0.009 by Fisher’s LSD test (Fig. 2D). Both diets also tended to decrease in maximal OXPHOS-linked OCR, which only reached statistical significance following addition of the DS (P < 0.05 by Fisher’s LSD test; Fig. 2E). Trends across timepoint and respiratory states were nearly identical between N6 and N3 diet groups, with no significant differences between diets.

Fig. 2
figure 2

Platelet oxidative metabolism. High resolution respirometry was performed on platelets isolated from mares at baseline (BL), following 6 weeks of feeding the flaxseed (N3) or corn oil (N6) supplements, and following an additional 6 weeks with dietary supplement (+ DS). Rates of oxygen consumption (JO2) were measured in intact cells under basal condition supported by endogenous substrates (A), followed by permeabilization with digitonin and the addition of malate and palmitoylcarnitine to generate the LEAK JO2 (B). ADP was then added to generate the OXPHOS supported by fatty acid oxidation (FAOx; C) and enable calculation of FAOx-linked coupling control (D), followed by the addition of pyruvate, glutamate and succinate to support maximal OXPHOS-linked JO2 (E). Graphs represent mean ± SEM of N = 6 mares per group. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Fisher’s LSD test following a significant ANOVA.

Full size image

In skeletal muscle fiber bundles, both diets similarly decreased fatty acid-supported LEAK (Fig. 3A; P ≤ 0.001) and OXPHOS states (Fig. 3B; P < 0.01) by Fisher’s LSD test. However, the DS significantly augmented this decline in the LEAK state (P ≤ 0.001 by Fisher’s LSD test), leading to marked increases in OXPHOS coupling control in the N6 (P = 0.006 by Fisher’s LSD test) and N3 (P = 0.002 by Fisher’s LSD test) groups (Fig. 3C). Maximal OXPHOS-linked OCR was lower following 6 weeks of the N6 diet (P = 0.033 by Fisher’s LSD test) and after DS (P = 0.021 by Fisher’s LSD test) compared to baseline, but was unaffected in the N3 diet group (Fig. 3D).

Fig. 3
figure 3

Skeletal muscle oxidative metabolism. High resolution respirometry was performed on skeletal muscle fiber bundles (3–4 mg) isolated from trapezius biopsies obtained at baseline (BL), following 6 weeks of feeding the flaxseed (N3) or corn oil (N6) supplements, and following an additional 6 weeks with dietary supplement (+ DS). Rates of oxygen consumption (JO2) were measured following by permeabilization with saponin and the addition of malate and palmitoylcarnitine to generate the LEAK JO2 (A). ADP was then added to generate the OXPHOS supported by fatty acid oxidation (FAOx; B) and enable calculation of FAOx-linked coupling control (C), followed by the addition of pyruvate, glutamate and succinate to support maximal OXPHOS-linked JO2 (D). Graphs represent mean ± SEM of N = 6 mares per group. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Fisher’s LSD test following a significant ANOVA.

Full size image

In granulosa cells, the N6 and N3 diets had no significant effects on OCR or OXPHOS coupling control under any substrate condition after 6 weeks of feeding (Fig. 4A-E). However, the additional 6 weeks with DS increased fatty acid-supported OCR in N6 (P = 0.033 by Tukey’s HSD test) and N3 (P = 0.027 by Fisher’s LSD test) groups (Figs. 4C), as well as maximal OXPHOS-linked OCR in N6 (P = 0.039 by Fisher’s LSD test) and N3 (P = 0.007 by Fisher’s LSD test) groups (Figs. 4E). Across all cell types examined, there were no significant differences between the N3 and N6 diets for any of the assayed time points.

Fig. 4
figure 4

Granulosa cell oxidative metabolism. High resolution respirometry was performed on granulosa cells obtained during aspiration of preovulatory follicles at baseline (BL), following 6 weeks of feeding the flaxseed (N3) or corn oil (N6) supplements, and following an additional 6 weeks with the dietary supplement (+ DS). Rates of oxygen consumption (JO2) were measured in intact cells under basal condition supported by endogenous substrates (A), followed by permeabilization with digitonin and the addition of malate and palmitoylcarnitine to generate the LEAK JO2 (B). ADP was then added to generate the OXPHOS supported by fatty acid oxidation (FAOx; C) and enable calculation of FAOx-linked coupling control (D), followed by the addition of pyruvate, glutamate and succinate to support maximal OXPHOS-linked JO2 (E). Graphs represent mean ± SEM of N = 6 mares per group. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Fisher’s LSD test following a significant ANOVA.

Full size image

PUFA diets with dietary supplement increase oocyte oxygen consumption rate

Oocyte OCR was unaffected by 6 weeks of feeding the N6 or N3 diets, but increased significantly following an additional 6 weeks with the dietary supplement in N6 (P = 0.012 by Fisher’s LSD test) and N3 (P = 0.037 by Fisher’s LSD test) groups (Fig. 5A). Oocyte ECAR was not significantly impacted by either PUFA diet or DS (Fig. 5B), suggesting an independent effect of the DS to stimulate basal rates of oocyte oxidative metabolism.

Fig. 5
figure 5

Oocyte metabolism. Oocytes obtained from preovulatory follicles of mares prior to (BL), following 6 weeks of N6 or N3 supplementation (6 weeks), and after an additional 6 weeks of diets with the dietary supplement (+ DS). A custom-built metabolic micro-sensor was used to measure (A) basal rates of oxygen consumption (OCR) as a measure of oxidative metabolism and (B) extracellular acidification rate (ECAR) as a measure of anaerobic glycolysis. Graphs represent mean ± SEM. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Fisher’s LSD test following a significant ANOVA.

Full size image

PUFA supplements increase ROS release from granulosa cells, but not platelets

In platelets, both N6 and N3 diets tended to decrease H2O2 release across the three respiratory states examined, but this only reached statistical significance in the N6 + DS group under basal conditions (P = 0.007 by Fisher’s LSD test) due to high variability across samples (Figs. 6A-C). Conversely, both N6 and N3 diets tended to increase H2O2 release after 6 weeks of supplementation, with trends for attenuation following addition of the DS that did not reach statistical significance (Figs. 7A-C).

Fig. 6
figure 6

Platelet ROS release. Rates of cellular reactive oxygen species (ROS) release rates were measured in platelets isolated from mares at baseline (BL), following 6 weeks of feeding the flaxseed (N3) or corn oil (N6) supplements, and following an additional 6 weeks with dietary supplement (+ DS). Rates of hydrogen peroxide release (JH2O2) were measured as the accumulation of resorufin fluorescence in respirometry chambers containing intact cells under basal conditions supported by endogenous substrates (A), following permeabilization with digitonin and the addition of malate, palmitoylcarnitine and ADP to generate fatty acid-supported OXPHOS (FAOx; B), and following the addition of pyruvate, glutamate and succinate to support maximal OXPHOS-linked JO2 (C). Graphs represent mean ± SEM of N = 6 mares per group. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Fisher’s LSD test following a significant ANOVA.

Full size image
Fig. 7
figure 7

Granulosa cell ROS release. Rates of cellular reactive oxygen species (ROS) release rates were measured in granulosa cells obtained during aspiration of preovulatory follicles at baseline (BL), following 6 weeks of feeding the flaxseed (N3) or corn oil (N6) supplements, and following an additional 6 weeks with dietary supplement (+ DS). Rates of hydrogen peroxide release (JH2O2) were measured as the accumulation of resorufin fluorescence in respirometry chambers containing intact cells under basal conditions supported by endogenous substrates (A), following permeabilization with digitonin and the addition of malate, palmitoylcarnitine and ADP to generate fatty acid-supported OXPHOS (FAOx; B), and following the addition of pyruvate, glutamate and succinate to support maximal OXPHOS (C). Graphs represent mean ± SEM of N = 6 mares per group. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Fisher’s LSD test following a significant ANOVA.

Full size image

PUFAs without supportive DS favors greater ROS release from muscle mitochondria

In permeabilized skeletal muscle fiber bundles, 6 weeks of feeding the N6 or N3 supplements alone significantly increased H2O2 release during fatty acid-supported OXPHOS (Fig. 8A; P < 0.001) and maximal OXPHOS (Fig. 8B; P < 0.05) by Fisher’s LSD test compared to baseline, which significantly decreased to near baseline levels following the addition of the DS to both diets (P < 0.0001). Expression of NADPH oxidase (NOX4), a major physiological source of mitochondrial ROS in skeletal muscle,37 tended to increase following N6 or N3 feeding and DS periods, reaching statistical significance when diet groups were combined to increase statistical power (Fig. 8C; P < 0.05 by paired t-test). Conversely, expression of the cytosolic (CuZn) and mitochondrial (Mn) isoforms of superoxide dismutase tended to be lower following the N6 and N3 diets and/or DS periods (Fig. 8D-E). Levels of lipid peroxidation-derived aldehyde protein adducts malondialdehyde (MDA; Fig. 8F) and 4-hydroxynonenal (4-HNE; Fig. 8G) were largely unaffected by the dietary interventions, with the exception of lower 4-HNE adducts following 6 weeks of N6 feeding compared to baseline (P = 0.021 by Tukey’s HSD test) that was reversed by addition of the DS. Taken together, these results suggest that dietary PUFA supplementation without the supportive DS increases muscle mitochondrial ROS release capacity and reduces antioxidant enzyme expression, but does not increase tissue oxidative stress.

Fig. 8
figure 8

Skeletal muscle ROS release and related parameters. Skeletal muscle (trapezius) biopsies were processed for assessment of reactive oxygen species (ROS) release or protein immunoblotting at baseline (BL), following 6 weeks of feeding the flaxseed (N3) or corn oil (N6) supplements, and following an additional 6 weeks with dietary supplement (+ DS). Rates of hydrogen peroxide release (JH2O2) were measured as the accumulation of resorufin fluorescence in respirometry chambers containing saponin-permeabilized fiber bundles following the addition of malate, palmitoylcarnitine and ADP to generate fatty acid-supported OXPHOS (FAOx; A), and following the addition of pyruvate, glutamate, and succinate to support maximal OXPHOS (B). Muscle protein homogenates were immunoblotted for NADPH oxidase 4 (NOX4; C), mitochondrial superoxide dismutase (MnSOD; D), cytosolic superoxide dismutase (CuZnSOD, E), and protein adducts of the lipid peroxidation aldehydes malondialdehyde (MDA; F) and 4-hydroxynonenal (4-HNE, G)-protein adducts. Graphs represent mean ± SEM of N = 6 mares per group. Superscripts differ at P < 0.05 between weeks within mare’s supplementation group by Tukey’s HSD test following a significant ANOVA.

Full size image

Discussion

The present study is the first to evaluate the biological impacts of dietary essential n-6 and n-3 PUFA supplementation on serum, blood cells, skeletal muscle and ovarian follicles of older mares. Our results demonstrate that adding PUFA-rich vegetable oils to a standard equine diet increases phospholipid PUFA content across multiple tissues and elicits modest effects the distribution of n-6 and n-3 PUFAs that are generally consistent with their fatty acid composition. Impacts of the N6 diet on phospholipid composition were generally more robust than the N3 diet, while both PUFA diets elicited similar effects on serum cytokines and cellular mitochondrial function. The addition of the dietary supplement tended to improve mitochondrial OXPHOS coupling control and reduce ROS release across all cell types tested, suggesting that these supportive nutrients may be advantageous for obtaining optimal responses to dietary PUFAs in aged mares. Taken together, these results provide novel insight to the biological effects of essential fatty acid supplementation in horses and add to a growing body of evidence that at least some of these effects are largely independent of variations in the n-6/n-3 PUFA ratio.

Similar to the ancestral human diet, the natural equine diet of grass is rich in n-3 ALA compared to n-6 LA,11,12 while modern industrialization has produced diets for both species containing much higher quantities of LA10 This has prompted widespread recommendations for dietary supplementation with n-3 PUFA to better balance the endogenous distribution of n-3 and n-6 PUFA and avoid the putative pro-inflammatory effects of excessive dietary n-6 PUFA intake13 However, few studies have reported health benefits or serum fatty acid composition changes following dietary n-3 PUFA supplementation in horses. This may be due in part to a much great abundance of n-6 over n-3 PUFA in membrane phospholipids. Indeed, we found 3 to 8-fold greater n-6 than n-3 PUFA levels across the four biological compartments investigated prior to dietary PUFA supplementation. This was augmented further by the N6 diet, and only modestly attenuated by 2 months on the N3 diet. While the N3 diet tended to increase ALA in all samples, it also increased LA (perhaps due to its presence in flaxseed oil at 17%), which is consistent with previous findings in plasma of young horses38 However, the more bioactive longer chain n-3 PUFA derivatives of ALA, docosahexaenoic acid (DHA), docosapentaenoic acid (DPA) and eicosapentaenoic acid (EPA), were only readily detected in granulosa cells and muscle tissue, likely reflecting their preferential incorporation to membrane phospholipids. This is in contrast to arachidonic acid (ARA), the long-chain n-6 PUFA derivative of LA, which was readily found in all biological samples including serum and follicular fluid. While the efficiency of long-chain PUFA synthesis from ALA and LA in vivo has been debated,39 the greater levels of ARA may result from the much higher levels of LA than ALA present across tissues. Indeed, N3 supplementation tended to decrease ARA and increase long-chain n-3 PUFA in granulosa cells and follicular fluid, but not in serum or muscle. More robust displacement of n-6 PUFA with reciprocal increases in long-chain n-3 PUFA have been reported in multiple biological fluids of horses following long-chain n-3 PUFA supplementation,40,41 but the physiological benefits of these changes are unclear and merit further investigation.

Despite the significantly higher levels of n-6 PUFA following the N6 diet in the present study, 5 of 6 cytokines detected in serum were not significantly impacted by either diet. Moreover, TNF-α increased following both N6 and N3 diets, which resolved after addition of the DS in the N6, but not N3 diet. The mechanisms for these effects are unclear, but they highlight a potential pro-inflammatory effect of dietary oil supplements in mares regardless of PUFA composition. Dietary n-6 and n-3 PUFA have been shown to exacerbate gut inflammation and serum cytokine elevations in patients with Crohn’s Disease and murine models,42 indicating that the enteric environment can influence systemic effects of dietary PUFA intake. While care was taken to avoid rancidity of oils through proper storage and daily rationing, we cannot rule out the possibility of naturally oxidized oil intake in triggering gut inflammation as described in other species43 Resolution of serum TNF-α elevations following addition of the DS to the N6 diet supports the need for supportive nutrients to prevent these potential effects of dietary PUFA supplementation, though this appeared to be insufficient following the N3 diet. Taken together, these results emphasize the importance of considering biological effects of dietary oils beyond those putatively linked to the n-6/n-3 PUFA ratio and inflammation.

PUFAs have diverse effects on mitochondrial function that vary depending on the cell type, method and duration of administration, and biological context being investigated44 This is the first study to evaluate impacts of dietary PUFA supplementation on mitochondrial function in equids, and our results demonstrate cell type-specific effects on respiratory capacity, OXPHOS efficiency and ROS release. In platelets and muscle, both N6 and N3 diets potently decreased mitochondrial respiratory capacity irrespective of substrates used, which tended to decrease further after addition of the DS. These decreases were comparatively greater in the LEAK state, leading to increases in OXPHOS coupling control in both cell types. Lower OCR and OXPHOS efficiency rates have been reported platelet and muscle mitochondria from aged humans,45,46 suggesting the effects of dietary interventions herein may be advantageous for improving bioenergetic efficiency in aged mares. Indeed, previous studies have demonstrated beneficial effects of dietary PUFA supplementation on indices of OXPHOS efficiency in rat cardiac muscle47 and human skeletal muscle48 Similar effects of the diets on platelet and muscle mitochondrial bioenergetics herein have also been reported in response to aging in humans,46 suggesting that platelets may serve as a less invasive means of assessing the impacts of aging and nutritional interventions on muscle mitochondria. However, we caution against making broad generalizations of mitochondrial effects observed in one cell type to another given vastly different metabolic demands and phenotypes.

Impacts of the PUFA supplementation on platelet and muscle mitochondrial ROS release were more variable. The N3 diet tended to decrease platelet ROS release with no additive effect of the DS, which is consistent with previous evidence for an attenuation of ROS production in platelets and neutrophils following 6 weeks of dietary n-3 PUFA supplementation in healthy humans49 However, both the N3 and N6 diets increased ROS release from muscle mitochondria of mares in the present study, which is consistent with effects of PUFA on multiple cell types,50,51 but in contrast to effects of n-3 PUFA supplementation on skeletal muscle of aged humans52 and mice fed a high-fat diet53 Greater ROS production down-regulates fatty acid oxidation in C2C12 skeletal muscle cells,54 corroborating the observed link between these two responses herein. This effect was associated with trends for higher NOX4 expression herein, which is known to be upregulated in multiple cell types by inflammatory cytokines,55 including TNF-α,56 suggesting potential links to the observed impacts on serum TNF-α herein. Both diets also tended to decrease SOD isozymes in muscle, which may further promote oxidative stress under states of elevated metabolic demand. However, PUFA supplementation tended to either decrease (in N6) or have no effect on indices of lipid peroxidation (in N3), consistent with previous findings in yearling horses57 This suggests that endogenous antioxidants or those added to the PUFA-rich oils to preserve shelf life must be sufficient to prevent overt oxidative stress despite a higher ROS release capacity in response to these diets. Indeed, addition of the DS fully reversed these elevations in muscle ROS release capacity, further emphasizing the efficacy and importance of including supportive antioxidants to PUFA-enriched diets to avoid adverse biological responses24.

We employed two different oxygen-sensing technologies to evaluate metabolic responses of granulosa cells and oocytes to the dietary interventions. While these assessments do not represent fertility directly, they provide important insights to the oocyte’s environment as it matures within the ovarian follicle. Granulosa cells line the inside of ovarian follicles and are involved in steroidogenesis and provide metabolic support to the oocyte. We found that while short-term PUFA supplementation had minimal effects on granulosa cell and oocyte metabolism, both N6 and N3 diets increased granulosa cell ROS release, which has been linked with reduced oocyte quality in humans58. Addition of the DS attenuated PUFA-induced elevations in granulosa cell ROS release and increased OXPHOS capacity, suggesting a more favorable environment for oocyte development. Indeed, oocyte OCR was markedly increased by addition of the DS, with no significant effects of the PUFA supplements alone. Oocyte mitochondrial function is critical for supporting maturation and early embryo viability59, and a greater OCR in equine oocytes is associated with improved development into transferable embryos after ICSI18,19. Equine oocytes store and utilize fatty acids as a primary source of energy, and oocytes from old mares contain fewer amounts of lipids compared to oocytes from younger mares19. Therefore, we postulate that the addition of supportive nutrients present in the DS, specifically L-carnitine and antioxidants, are required to reap the benefits of dietary fatty acids on the follicular environment and support optimal oocyte development in older mares and perhaps other species, as previously demonstrated by our group22. In addition, multiple trace minerals have been previously reported to exert beneficial effects on cellular antioxidant status and mitochondrial function in horses60,61.

Although the use of dietary oils is an effective method to increase caloric intake in horses, the present study suggests that oil supplementation, regardless of n-3 or n-6 PUFA content, may induce undesirable effects on mitochondrial oxidative capacity, reactive oxygen species production, and systemic inflammation. However, the addition of dietary support nutrients in combination with either oil type may mitigate these effects, providing a feasible method to optimize equine health following oil feeding. n-3 ALA supplementation with flaxseed oil did not provide significant benefits to cell function or metabolism compared to n-6 LA supplementation with corn oil, which may be due in part to a lack of its endogenous conversion to the more bioactive long-chain n-3 PUFA. This is consistent with evidence from many species and may indicate inherently low delta-6 desaturase expression or its inhibition by dietary PUFA supplements in horses, further indicated by the absence or very low levels of long-chain n-3 PUFA in equine samples herein. This may also explain the lack anti-inflammatory effects of n-3 ALA indicated by greater TNF-α concentrations observed following both N3 and N6 oil treatments. However, it is important to note that we cannot rule out potentially beneficial effects of the N3 and N6 oils on outcomes not assessed in our study. Nevertheless, our results indicate that oil supplementation alone may reduce mitochondrial capacity to oxidize acylcarnitines and increase ROS production in skeletal muscle, platelets, and granulosa cells. Therefore, potential long-term health risks in older horses associated with these effects should be considered to minimize adverse consequences. Our study provides evidence that additional supplementation of antioxidants, L-carnitine, and other supportive nutrients is sufficient to reduce these adverse effects, and are therefore recommended for maintain the systemic and reproductive health of older horses consuming dietary PUFA supplements.