Oleic acid abomasal infusion limits lipolysis and improves insulin sensitivity in adipose tissue from periparturient dairy cows

Excessive adipose tissue (AT) lipolysis around parturition in dairy cows is associated with impaired AT insulin sensitivity and increased incidence of metabolic diseases. Supplementing cows with oleic acid (OA) reduces circulating biomarkers of lipolysis and improves energy balance. Nevertheless, it is unclear if OA alters lipid trafficking in AT. In the liver and skeletal muscle, OA improves mitochondrial function and promotes lipid droplet formation by activating perilipin 5 (PLIN5) and peroxisome proliferator-activated receptor α (PPARα). However, it is unknown if this mechanism occurs in AT. The objective of this study was to determine the effect of OA on AT lipolysis, systemic and AT insulin sensitivity, and AT mitochondrial function in periparturient dairy cows. Twelve rumen-cannulated Holstein cows were infused abomasally following parturition with ethanol (CON) or OA (60 g/d) for 14 d. Subcutaneous AT samples were collected at 11 ± 3.6 d before calving (−12 d), and 6 ± 1.0 d (7 d) and 13 ± 1.4 d (14 d) after parturition. An intravenous glucose tolerance test was performed on d 14. Adipocyte morphometry was performed on hematoxylin and eosin-stained AT sections. The antilipolytic effect of insulin (1 μg/L) was evaluated using an ex vivo explant culture following lipolysis stimulation. PLIN5 and PPARα transcription and translation were determined by real-time quantitative PCR and capillary electrophoresis, respectively. RNA sequencing was used to evaluate the transcriptomic profile of mitochondrial gene networks. In CON cows, postpartum lipolysis increased the percentage of smaller (<3,000 μm 2 ) adipocytes at 14 d compared with −12 d. However, OA limited adipocyte size reduction at 14 d. Likewise, OA decreased lipolysis plasma markers nonesterified free fatty acids and β-hydroxybutyrate at 5 and 7 d. Over the 14-d period, compared with CON, OA increased the concentration of plasma insulin and decreased plasma glucose. During the glucose tolerance test, OA decreased circulating glucose concentration (at 10, 20, 30, 40 min) and the glucose clearance rate. Moreover, OA increased insulin at 10 and 20 min and tended to increase it at 30 min. Following lipolysis stimulation, OA improved the antilipolytic effect of insulin in the AT at 14 d. PLIN5 and PPARA gene expression decreased postpartum regardless of treatment. However, OA increased PLIN5 protein expression at 14 d and increased PPARA at 7 and 14 d. Immunohis-tochemical analysis of AT and RNA sequencing data showed that OA increased the number of mitochondria and improved mitochondrial function. However, OA had no effect on production and digestibility. Our re-sults demonstrate that OA limits AT lipolysis, improves systemic and AT insulin sensitivity, and is associated with markers of mitochondrial function supporting a shift to lipogenesis in AT of periparturient dairy cows.


INTRODUCTION
During the periparturient period, which extends from 3 wk before to 3 wk after parturition, dairy cows have an increase in nutrient and energy requirements due to fetal growth, parturition, and the onset of lactogenesis.During this period, DMI is at its lowest, which puts the cow in a state of negative energy balance (NEB; Grummer et al., 2004).As a result, triacylglycerol (TAG) molecules stored in the adipose tissue (AT) are hydrolyzed through the process of lipolysis and nonesterified free fatty acids (NEFA) are released into circulation.Enhanced AT lipolysis coincides with a period of reduced insulin secretion (Dębski et al., 2017) and insulin sensitivity in adipocytes, which redirects energy to the mammary gland (De Koster and Opsomer, 2013).As lactation progresses, AT becomes more responsive to insulin and stores energy surplus as fatty acids (FA) incorporated into TAG in a process known as lipogenesis.
However, intense and protracted lipolysis rates, which are reflected by elevated concentrations of plasma NEFA and BHB, are strongly associated with reduced milk production, poor reproductive performance, and disease susceptibility.Similarly, extended periods of insulin resistance predispose cows to inflammatory and metabolic diseases (De Koster and Opsomer, 2013).
Targeted FA supplementation in periparturient cows increases the energy density of the diet, which may minimize NEB and its consequences such as lipolysis.We demonstrated that including oleic acid (OA) in lactation diets improves energy balance and decreases plasma NEFA (de Souza et al., 2021a).Increasing the percentage of OA in the diet of early-lactation cows improved DMI and reduced BW and BCS losses (de Souza et al., 2021b).Reductions in BCS reflect AT mass losses due to lipolysis.These findings suggest that OA induces changes in the activity of the metabolic pathways of AT that favor lipogenesis over lipolysis.However, the specific mechanisms by which OA modulates FA trafficking systemically and in AT of periparturient cows are not completely understood.
Rodent models demonstrate at least 2 possible mechanisms for OA lipogenic effect.First, OA stimulates pancreatic insulin secretion (Fujiwara et al., 2005).Second, in HepG2 and primary hepatocyte cells, OA promotes lipid droplet (LD) formation through the activation of perilipin 5 (PLIN5) and nuclear transcription factor peroxisome proliferator-activated receptor α (PPARα; Zhong et al., 2019;Najt et al., 2020).PPARα is a key regulator of lipid metabolism.PLIN5 is a LD protein that enhances the formation of new lipids, targets FA for mitochondrial oxidation, and inhibits lipolysis activity.The role of PLIN5 in regulating LD formation in highly oxidative tissues such as heart, skeletal muscle, and liver is well characterized (Kimmel and Sztalryd, 2014).However, whether OA can modulate PLIN5 and PPARα activity in bovine AT is unknown.
There are reports indicating that OA has anti-inflammatory properties in AT and enhances the capacity to oxidize FA in mitochondria.For instance, in rodent models, OA promotes anti-inflammatory polarization of AT macrophages (Camell and Smith, 2013).In myoblasts, OA improves mitochondrial function especially through the enhancement of FA oxidation (Lim et al., 2013).These effects may be beneficial to dairy cows because excessive lipolysis in AT induces an inflammatory response characterized by AT macrophages infiltration (Contreras et al., 2015).Adipose tissue macrophages trafficking is mediated in part by the release of osteopontin, which is a chemoattractant, that also promotes the polarization of these mononuclear immune cells toward a pro-inflammatory phenotype (Lee et al., 2013).Lipolysis also promotes the production of inflamma-tory cytokines, including IL6, that inhibits mitochondrial biogenesis and further dysregulates AT lipolysis (Vieira-Potter, 2014).However, whether OA exerts this effect in the AT of periparturient cows is unknown.Therefore, in the present study, we examined the role of OA on lipid metabolism and mitochondrial function in AT of periparturient dairy cows.Using abomasal infusion of OA, we provide evidence that OA reduces FA trafficking out of AT in periparturient dairy cows by limiting adipocyte lipolysis, improving insulin sensitivity and mitochondrial function in AT, and enhancing systemic insulin secretion and sensitivity.

Design and Treatments
All experimental procedures were approved by the Institutional Animal Care and Use Committee (IA-CUC #11/16-188-00) at Michigan State University (East Lansing).The experiment began on November 22, 2018, and ended on March 25, 2019.Twelve rumencannulated multiparous Holstein cows at the Michigan State University Dairy Cattle Teaching and Research Center were used for this study.Cows were assigned into 6 blocks based on BCS, previous lactation 305d mature-equivalent yield, and parity.The averages (mean ± SD) were 3.53 ± 0.22, 32,182 ± 3,752 kg, and 2.67 ± 0.65 for BCS, mature-equivalent yield, and parity, respectively.All animals received a common diet during the close-up (30 d pre-calving to parturition) and fresh (1-15 DIM) periods (Table 1) to meet the requirements of the animals as determined by NRC (2001).Cows were housed in individual tiestalls throughout the experiment and milked twice daily (0400 and 1500 h).Access to feed was blocked once a day to allow for collection of orts and to offer new feed.Feed intake was recorded, and cows were offered 115% of expected intake at 1000 h daily.Water was available ad libitum and the tiestalls were bedded with sawdust and cleaned twice daily.Standard reproduction and health herd checks and breeding practices were maintained during this study.
Treatments were 0 (CON; n = 6) and 60 g/d of abomasally infused OA (n = 6; O1008-1G, Sigma-Aldrich).Because the OA comes as a viscous liquid, 60 ± 1.0 g was suspended in ethanol and the volume was brought up to 200 mL.The dose was selected based on results from a dose response study (Prom et al., 2021).Abomasal infusion devices were inserted into the abomasum 5 d before the beginning of the study.Infusion lines (0.5 cm i.d.polyvinyl chloride tubing) passed through the rumen fistula and sulcus omasi into the abomasum (Lock et al., 2007).Lines were checked Abou-Rjeileh et al.: OLEIC ACID AND ADIPOSE LIPID MOBILIZATION daily throughout the study to ensure proper placement.The 200-mL suspension of the daily dose of OA (≥99% cis-9 18:1 as a free FA) and 0 g/d treatment were mixed in individual glass jars daily.Hence, the OA infusion included ~130 mL of ethanol and the CON included 200 mL.The infusate solution was divided into 4 equal infusions per day occurring every 6 h beginning at noon the day after calving for 14 d.Infusate solutions were delivered into infusion lines using 60-mL plastic syringes.The following is the same infusion protocol that followed for both groups of cows to ensure consistency: (1) lines were flushed 3× with 50 mL of water to ensure the line is not blocked, (2) 50 mL of the appropriate treatment was infused, and (3) lines were flushed with 30 mL of ethanol once and 50 mL of water 4×.Although CON cows received ~70 mL/d more ethanol, we do not expect any effect of the ethanol infusions on either group because postpartum dairy cows have suffi-cient metabolic capacity to cope with high dietary concentration of primary alcohols (Raun and Kristensen, 2011).However, future studies providing equal amounts of ethanol to both treatments and minimizing the use of ethanol are important.

Sample and Data Collection
Subcutaneous AT (SCAT) samples were obtained from the right flank at 11 ± 3.6 d before expected calving date (PreP), and 6 ± 1.0 d (PP1) and 13 ± 1.4 d (PP2) after parturition.In short, after local anesthesia (15 mL of 2% lidocaine hydrochloride, VetTek) and aseptic preparation of the surgical area using iodine scrub, iodine prep, and alcohol, a vertical skin incision of 5 cm was made.The site of the incision was moved ventrally 3 to 4 cm at each collection time point.This anatomic location offers a wider surgical area for repeated AT sampling, it is less likely to have fecal contamination, and it is seldom used as an injection site.Around 5 g of SCAT was collected.Part of the SCAT was used in the AT lipolysis assay as described below, another part was snap-frozen in liquid nitrogen and stored at −80°C for RNA and protein extraction, and the remaining was fixed in 4% paraformaldehyde for 72 h for histology analysis.The skin was closed using a continuous interlocking suture with Braunamid (USP1, Aesculap).Sutures were removed 14 d after each procedure.
Blood samples were collected on d −14, 3, 5, 7, 10, 12, and 14 relative to calving.Blood was drawn via coccygeal venipuncture using coated collection tubes (K 2 EDTA) before morning feeding and stored on ice.Samples were then centrifuged at 2,000 × g for 15 min at 4°C for plasma fraction collection, and then stored at −20°C until further analysis.
Samples of all diet ingredients and orts from each cow were collected daily and stored in plastic bags at −20°C until processed.Milk samples were collected at each milking on d 2, 4, 7, 9, 11, and 14 postpartum and stored in sealed tubes with preservatives (Bronopol tablet; D&F Control Systems) at 4°C for component analysis.Milk yield and feed offered and refused were recorded daily throughout the 14 d of the experiment.Body weight was taken on d 4, 6, 8, 11, 13, and 15 postpartum and body condition was scored by 3 trained investigators on d 4, 8, 11, and 15 postpartum on a 5-point scale (Chirivi et al., 2022).On d 10 postpartum, fecal samples (~400 g) were collected every 6 h, totaling 4 samples per cow, to estimate nutrient digestibility.The 6-h interval over 24 h accounts for diurnal variation.Feces were stored in a sealed plastic cup at −20°C.

Sample Analyses
Plasma concentrations of insulin, glucose, NEFA, and BHB were determined using an Olympus AU640e chemistry analyzer (Olympus America) at the Michigan State University Veterinary Diagnostic Laboratory (East Lansing).Concentrations of triglyceride, total protein, and albumin were determined using the CataChemWell-T analyzer (Catachem Inc.).
Diet ingredients, orts, and fecal samples were dried at 55°C in a forced-air oven for 72 h.Dried fecal samples for each cow were then composited per cow.Dried samples were ground with a Wiley mill (1-mm screen; Arthur H. Thomas).Feed ingredients were analyzed for absolute DM, ash, CP, and starch, and orts and feces for absolute DM and ash by Cumberland Valley Analytical Services (Waynesboro, PA) as described by Boerman et al. (2017).Absolute DM was determined by drying samples in an oven at 105°C using the National Forage Testing Association reference method (Shreve et al., 2006).Indigestible NDF was estimated as NDF after 240-h in vitro fermentation (Goering and Van Soest, 1970) and was used as an internal marker to estimate fecal output to determine apparent total-tract digestibility of nutrients (Cochran et al., 1986).The amount of FA in the infusate was considered for the intake, digestibility, and absorption of FA.Feed, orts, and fecal FA concentrations were direct methylated with a 5% HCl (Boerman et al., 2017).
Milk samples were analyzed for fat, true protein, and lactose concentrations by mid-infrared spectroscopy using the Bentley Instruments Combi-System (Central Star Cooperative).Yields of 3.5% FCM, ECM, milk energy, and milk components were calculated using milk yield and component concentrations from each milking and summed for a daily total.

Glucose Tolerance Test
An intravenous glucose tolerance test (GTT) was performed on d 14 postpartum at 0730 h around 1 h after abomasal infusions while cows were blocked from feed.We calculated 50% dextrose injection volumes for each cow according to the following formula: injection volume = 1.67 mmol glucose/kg of BW.Blood samples were taken at −10 min before dextrose injection and then every 10 min through 120 min, and at 150 min and 180 min relative to injection start time.Jugular catheters were flushed with 4 mL of 4.2% sodium citrate saline before and after each blood collection.Samples were stored on ice and centrifuged within an hour of collection at 2,000 × g for 15 min at 4°C.Plasma was collected and stored at −20°C until further analysis.Parameters included delta maximum (μIU/mL), clear-ance rate (%/min), and time to reach half-maximal concentration (min).Parameters were calculated based on the equations by Oliveira et al. (2016).

AT Lipolysis Assay
After collection, SCAT was immediately placed in 30 mL of Krebs Ringer Bicarbonate HEPES Buffer (Teknova; pH 7.4) and transported to the laboratory at 37°C for assessing AT lipolysis as described in Chirivi et al. (2022).Briefly, AT was cut into approximately 100-mg tissue fragments using sterile surgical scissors and placed into 24-well plates containing 1 mL of Krebs Ringer Bicarbonate HEPES supplemented with 3% FA-free BSA (Millipore-Sigma).Culture plates with AT explants were pre-incubated at 37°C for 20 min on a shaker.Basal lipolysis was determined without the addition of any reagent.Stimulated lipolysis was determined after a β-adrenergic agonist challenge with isoproterenol hydrochloride (ISO, I6504, Millipore-Sigma; 1 μM).Insulin response was assessed using 1 μg/L of insulin solution from bovine pancreas (INS, I0516, Sigma-Aldrich).Adipose tissues were incubated with insulin for 1 h before ISO stimulation.Three hours after ISO stimulation, culture media were collected, snapfrozen in liquid nitrogen, and stored at −80°C until further analysis.All treatments were prepared fresh immediately before the lipolysis assay.To assess lipolysis, culture media were analyzed for glycerol quantification using free glycerol reagent (Millipore-Sigma, F6428).Briefly, 67 μL of media sample, blank (Krebs Ringer Bicarbonate HEPES + 3% BSA), and standards (Glycerol standard solution, G7793, Sigma-Aldrich) were loaded into a 96-well plate in duplicates.We added 200 μL of the free glycerol reagent.The plate was incubated for 5 min at 37°C.Absorbance was measured at 540 nm.The intra-and interassay coefficients of variation were 4.24% and 6.90%, respectively.Glycerol content was measured using the following formula: where A Sample = absorbance of media sample, A Blank = absorbance of blank well, and A Standard = absorbance of standard well.Glycerol content was normalized by AT weight (nmol/mg).

RNA Extraction and Gene Expression Analysis
The RNA from the SCAT samples was extracted using a TRI Reagent (R2050-1 Zymo Research) and the

Abou-Rjeileh et al.: OLEIC ACID AND ADIPOSE LIPID MOBILIZATION
Quick-RNA Miniprep plus kit (R1058 Zymo Research) as described in Chirivi et al. (2022).The purity, concentration, and integrity of total RNA were evaluated using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific) and Agilent Bioanalyzer 2100 (Agilent Technologies).All samples had a 260:280 nm ratio between 1.85 and 2.06 and an RNA integrity number >7.Reverse transcription was performed with 300 ng of RNA using 4 μL of the qScript cDNA SuperMix (95048 Quantabio) for 5 min at 25°C, 30 min at 42°C, and 5 min at 85°C.cDNA was stored at −20°C until further analysis.

RNA-Sequencing Analysis
The RNA-sequencing (RNA-seq) data sets from AT samples collected from these cows are publicly available from GEO (Gene Expression Omnibus, GSE159224) and the analyses methods are described in Salcedo-Tacuma et al. (2020).Changes in gene expression between PreP and PP1-PP2 data sets were quantified and analyzed against the treatments with OA using fragments per kilobase of transcript per million (FPKM; Supplemental File; https: / / doi .org/ 10 .6084/m9 .figshare.21265968;Abou-Rjeileh et al., 2023c).

Protein Extraction and Analysis
Proteins were extracted from ~100 mg of snap-frozen SCAT samples using radioimmunoprecipitation assay buffer (R3792, Teknova) supplemented with protease cOmplete mini EDTA-free protease inhibitor (Roche) and phosphatase inhibitors (phosphatase inhibitor cocktail II J61022, Thermo Fisher Scientific).The concentration of protein was determined using the Pierce BCA Protein Assay Kit (Thermo Scientific, 23225;Chirivi et al., 2022).Protein samples were separated by capillary electrophoresis using the 12-230 kDa Wes Separation Module capillary cartridges in the Simple Protein Wes system (SM-W004, ProteinSimple) following the manufacturer's protocol.The optimal protein concentration for the antibodies used in these experiments was 0.5 mg/mL for PPARα, vinculin, and osteopontin, and 0.4 mg/mL for PLIN5.Antibodies and dilutions used were as follows: mouse monoclonal PPAR α/NR1C1 antibody (1:50, NB300-537 Novus Biologicals), mouse monoclonal anti-LSDP5 (E3; 1:10, sc-514296, Santa Cruz), mouse monoclonal anti-vinculin (7F9; 1:50, sc-73614, Santa Cruz), and mouse monoclonal osteopontin/OPN antibody (1:50, NB110-89062 Novus Biologicals).The anti-mouse module for the Wes (DM-002, ProteinSimple) kit that includes luminol-S, peroxide, antibody diluent 2, streptavidin-HRP, and anti-mouse secondary antibody was used for detection.The chemiluminescence of secondary antibodies generates signal peaks that were used for analysis.These signal peaks were transformed into digital images depicting bands as observed in a western blot analysis.Using Compass software (ProteinSimple), the peak areas of PPARα and PLIN5 were estimated and normalized against the peak area of vinculin.The peak area for osteopontin was estimated and normalized against the total protein detected with the total protein detection module kit (DM-TP01, ProteinSimple).

Immunohistochemistry
Subcutaneous AT were fixed in 4% paraformaldehyde, blocked in paraffin, and then sectioned into 4-μm slices by the Michigan State University Investigative Histopathology Laboratory.Following pretreatment protocol, standard micro-polymer staining was performed at room temperature on the Biocare intelliPA-TH automated stainer.All staining steps were followed by rinses in TBS Autowash Buffer (Biocare).Nonspecific proteins were blocked using Background Punisher (Biocare) for 5 min.Primary antibodies were diluted in normal antibody diluent (Scytek), rabbit polyclonal antibody anti-PLIN5 (1:50; PAB12542, Abnova), and mouse monoclonal anti-SDHA (succinate dehydrogenase complex flavoprotein subunit A) antibody (1:500; ab14715, Abcam) and incubated for 1 h, followed by HRP Polymer detection (Biocare), rabbit on Farma for PLIN5, and mouse on Farma HRP for SDHA for 30 min.The reaction was developed with Romulin AEC (Biocare) for 5 min.The sample was counterstained with CATHE hematoxylin diluted 1:10 (Biocare) for 2 min, dehydrated, cleared, and mounted with synthetic mounting media.Brightfield images were collected using either a 10× Plan Fluor (NA 0.30) or 40× Plan Fluor (NA 0.75) objective on a Nikon Eclipse Ni upright microscope (Nikon Instrument Inc.), configured with a Nikon DS-Fi2 color camera.Images were collected and analyzed using the Nikon NIS Elements software (version 5.21.03).Adiposoft plugin (v.1.15) for ImageJ Fiji (version 2.0.0) was used for adipocyte size analyses (Ferland et al., 2020).Adipocyte areas were divided into 7 bins ranging from smaller (<3,000 μm 2 ) to larger (>9,001 μm 2 ) adipocytes.The SDHA intensity was also measured using ImageJ as described in Crowe and Yue (2019).All image capture and analysis were carried out with identifier blind to treatments.

Statistical Analysis
Production responses, digestibility, GTT, and plasma data were analyzed using the GLIMMIX procedure of SAS (version 9.4, SAS Institute Inc.).Variables evaluated over time were analyzed with repeated measures according to the following model: where Y ijklm = the dependent variable, μ = the overall mean, F i = the fixed effect of treatment, T j = the fixed effect of time, B k = the random effect of block, C l (B k F i ) = the random effect of cow within block and treatment, J m = random effect of Julian date of parturition, F i × T j = the fixed effect of the interaction between treatment and time, and e ijklm = the residual error.The Julian date was kept in the model only when its estimate in the covariance parameter was greater than zero (de Souza et al., 2021a).Spatial power was used as a covariance structure for most variables because their repeated measurements were not equally spaced.First-order autoregressive was used for DMI and milk yield because their repeated measurements were equally spaced (Littell, 2006).Variables determined once were analyzed without repeated measures, time, interactions, or nested effects in the model.Blood samples taken on d −14 relative to calving were tested as covariates.The pre-calving samples were analyzed as covariates only and not as part of the treatment or treatment and time interaction effects.Covariates were nonsignificant and removed from the model except for glucose and insulin.Removing the covariate had no change on the interpretation of the data.Normality of the residuals was checked with normal probability and box plots and homogeneity of variances with plots of residuals versus predicted values.Denominator degrees of freedom were adjusted by the containment method, the SAS default method for the analysis performed.Significance of main effects was declared at P ≤ 0.05, and tendencies were declared at 0.05 < P ≤ 0.15.Sliced comparisons of the effect of treatment at a specific time point occurred when treatment by time was significant (P ≤ 0.15).
Protein, RNA, histology, and RNA-seq data were analyzed using a mixed-effect model in JMP (JMP, Version 15.SAS Institute Inc.) including the random effect of cow and the fixed effect of treatment, time, and their interactions.The PreP sample was included in the model as a covariate.In the figures, PreP values are used as a visual representation for each group pre-infusion.In some protein expression analysis, the postpartum data (PP1 and PP2) was normalized by the prepartum data (PreP) and is reported as "relative to PreP" in the figures.Time relative to parturition was included as a repeated measure.Compound symmetry was used as a covariance structure.The RNA data were natural log-transformed.Results are presented as mean ± standard error of the mean unless stated otherwise.
Significance was declared at P ≤ 0.05, and tendencies were declared at 0.05 < P ≤ 0.15.Results presented at a specific time point (i.e., PP1, PP2, minute, or specific day) in the text and figures reflect treatment by time interaction.Sliced comparisons occurred when treatment × time at least tended to be significant (P ≤ 0.15).Significance for the main effects was declared at P ≤ 0.05.For treatment by time interactions, it was declared at P ≤ 0.15 due to the difficulty to get a higher sample size and the variation in molecular biology procedures.When interaction was significant, modified protected least significant difference was used to evaluate all possible pairwise comparisons.
Measurement of plasma metabolites throughout the study reflected similar results.Compared with CON, OA decreased plasma glucose (60.60 vs. 55.60 ± 1.73 mg/dL; P = 0.01) and increased the concentration of plasma insulin (2.04 vs. 2.12 ± 0.46 μIU/mL; P < 0.01, Figure 2C).However, we did not observe interactions between treatment and time for glucose (P = 0.49) or insulin (P = 0.61), indicating that the treatment effects were consistent over the 14-d postpartum.
Adipose tissue insulin sensitivity was assessed using an ex vivo approach, which challenges AT with a β-adrenergic agonist ISO and evaluates the anti-lipolytic effect of insulin.We observed a treatment by time interaction where, compared with CON, OA improved the antilipolytic effect of insulin in SCAT at PP2 by ~86% (Figure 2D; P < 0.01).

OA Increases PLIN5 and PPARα Protein Expression in SCAT
To understand the mechanism behind OA effect on lipid metabolism in AT, we evaluated the gene and protein expression of PLIN5 and PPARα, which are known effectors of lipogenic pathways.We did not observe a treatment effect (P ≥ 0.11) or a treatment by time interaction (P ≥ 0.65) for PLIN5 and PPARA gene expression (Figures 3A and B).However, we observed a treatment by time interaction for PLIN5 protein expression (P = 0.15), where OA increased PLIN5 protein expression at PP2 compared with CON (Figure 3C; P = 0.02).We observed a treatment effect where OA increased PPARA protein expression compared with CON (Figure 3D; P = 0.01).Immunohistochemical analysis of AT using an antibody against PLIN5 reflected similar results as the translational analysis.It also demonstrated that PLIN5 is located around the nucleus and LD of adipocytes (Figure 3E).However, we did not perform colocalization analysis.

OA and Inflammatory Responses in SCAT
Because lipolysis intensity is associated with SCAT inflammation, we quantified the content of AT osteopontin, a strong chemoattractant of macrophages, and the gene expression of IL6 and IL10 cytokines.Compared with CON, OA decreased osteopontin protein expression at PP1 (P = 0.001) and PP2 (P < 0.0001; Figure 4A).We found no effect of treatment or treatment by time interaction on IL6 FPKM (Figure 4B; P ≥ 0.52).However, we observed a treatment by time interaction (P = 0.14) for IL10, where OA tended to increase IL10 FPKM at PP2 compared with CON (Fig- ure 4B; P = 0.13).

OA Promotes SCAT Mitochondrial Biogenesis and Alters Mitochondria-Related Gene Transcription
Immunohistochemical analysis of AT using an antibody against the mitochondrial-specific marker SDHA demonstrated a treatment by time interaction where OA increased SDHA mean intensity at PP2 compared with CON, reflecting enhanced mitochondrial biogenesis (Figures 5A and B; P = 0.04).The RNA-seq analysis indicates that OA infusion was associated with the expression of markers of mitochondrial function.Compared with CON, OA decreased sirtuin 3 (SIRT3) FPKM (P = 0.02).Oleic acid had no effect  PP2 compared with PP1 for both treatments (Figure 5C; P = 0.02).We detected no treatment by time interactions for any of the RNA-seq data.

OA Has No Effect on Milk Production and Nutrient Digestibility
We found no effect of treatment or an interaction for treatment by time for production responses (P ≥ 0.16) or nutrient digestibility variables (P ≥ 0.10; Supplemental Table S2A and B; https: / / doi .org/ 10 .6084/m9 .figshare.21265932;Abou-Rjeileh et al., 2023b).

DISCUSSION
Research in ruminants provides evidence for the beneficial effect of supplementing OA to lactating dairy cows on metabolic function and production.We have previously demonstrated that increasing supplementation of OA during the immediate postpartum improves energy balance and partitioning toward body reserves ( de Souza et al., 2021a).In the present study, we provide evidence for the lipogenic potential of OA that limits lipid mobilization during the first 2 wk after parturition.This finding is related to the enhancement of systemic and AT insulin sensitivity that in turn promotes lipogenesis in adipocytes.Moreover, OA lipogenic capacity coincides with an increased translation of PLIN5 and PPARα.Oleic acid abomasal infusion was also associated with markers of mitochondrial biogenesis and function, and reduced expression of osteopontin, a potent chemoattractant of immune cells in AT.
Hormonal changes during the postpartum period associated with parturition and the onset of lactogenesis favor lipolysis over lipogenesis in AT (Vernon et al., 1995).In our current study, postpartum lipolysis decreased triglyceride stores of fat cells, which was reflected by an increase in the percentage of smaller adipocytes (<3,000 μm 2 ).This coincided with an increase in the lipolysis plasma biomarkers NEFA and BHB.In cows, adipocyte size has long been used as a marker of adipose mass (Robelin, 1981) and has been shown to have a strong relationship with BCS and BW (Waltner et al., 1994).Our results coincide with similar reports of a decrease in adipocyte size as a reflection of lipolysis postpartum (Chilliard et al., 1984).In our study, OA limited the size reduction of the adipocytes, and decreased plasma NEFA and BHB, reflecting a lower lipolytic or a higher lipogenic rate.Although OA did not affect BW in the present experiments, these findings on adipocyte size distribution agree with the effect of OA decreasing BW and BCS losses observed by our group (de Souza et al., 2021b).In bovine adipocytes, OA has been shown to enhance adipogenesis and lipogenesis compared with SFA such as palmitic and stearic acid (Yanting et al., 2018).Hence, OA limits lipolysis in the AT and this could be due to a shift to fat storage.Further research is needed to validate OA lipogenic potential during NEB by tracking triglyceride biosynthesis in adipocytes and its effect on periparturient lipogenesis.Moreover, the direct effect of OA on the neutral lipases (i.e., HSL, ATGL) activity requires additional examination.
A possible mechanism by which OA limits FA trafficking out of adipocytes and reduces AT lipolysis is by increasing circulating insulin.The pancreatic peptide is the major antilipolytic hormone that potently stimulates lipogenesis in adipose cells (Kersten, 2001).In monogastrics, OA enhances insulin secretion in pancreatic β-cells (Fujiwara et al., 2005).In mid-lactation dairy cows, OA abomasal infusion (Prom et al., 2021) and dietary supplementation (de Souza et al., 2018(de Souza et al., , 2019) ) resulted in higher plasma insulin compared with control diets and other FA supplements.Our results agree with these studies as we observed a long-term effect of OA, wherein it increased insulin by 3.92% and decreased glucose by 8.25% over the 14-d period.However, our data highlight the need to further characterize insulin response following OA supplementation in dairy cows during the periparturient and early-lactation periods.
In the present study, the response to OA supplementation went beyond increasing plasma insulin.We examined whole-body insulin responsiveness to glucose during a glucose challenge.As expected, we observed a positive association between insulin and glucose reflecting a state of glucose homeostasis.This is also observed in dairy cows at different stages of lactation when a glucose challenge is used to determine insulin sensitivity (Zachut et al., 2013;Rico et al., 2017).During GTT, following glucose infusion, the glucose peak reflects insulin responsiveness, whereas clearance rate reflects insulin sensitivity (Sano et al., 1991).Our current results demonstrate for the first time in periparturient dairy cows that abomasal infusion of 60 g/d of OA reduces both the peak and the clearance rate of glucose.Similar results have been observed in rats where OA restored insulin sensitivity caused by a high-fat diet (Chen et al., 2018), and in humans where decreasing saturated FA and increasing monounsaturated FA in a diet improved insulin sensitivity but had no effect on insulin secretion (Vessby et al., 2001).Moreover, it is important to emphasize that with our current study, we cannot distinguish between the long-term effect of OA versus short-term on GTT results because both could be affecting the response observed.To focus on the long-term effect of OA, infusions right before the GTT should be avoided.Additionally, to further assess and quantify insulin sensitivity during OA supplementation, euglycemic clam studies are warranted.
In the present study, we determined that OA also causes changes in insulin sensitivity in the AT.Our results show that during ex vivo β-adrenergic stimulation of AT, OA-treated cows had higher anti-lipolytic responses to insulin reflecting enhanced AT insulin sensitivity.Hence, the decrease in glucose during the treatment period (14 d) and the GTT observed responses could also be caused, in part, by an increase in insulin sensitivity in the AT.Moreover, the lower glucose concentration in this study could reflect an insulin-stimulated glucose transport into AT as observed in humans (Ryan et al., 2000).The mechanisms regulating insulin sensitivity and response in dairy cows remain largely unknown.In contrast to OA responses observed in this study, palmitic acid supplementation has no effect on systemic insulin sensitivity in dairy cows (Piantoni et al., 2013;Mathews et al., 2016).To our knowledge, this is the first study that shows the possibility of enhancing insulin sensitivity systemically and in the AT in ruminants following OA supplementation.A possible mechanism for OA to improve insulin sensitivity is through enhancing the IR/IRS1/Akt insulin pathway (Pardo et al., 2015) and stimulating glucose transporter 4 translocation (Alkhateeb and Qnais, 2017).However, in our study, we did not evaluate the insulin pathway at the molecular level in adipocytes and this is a gap in knowledge that requires future research.
In the present study, OA reduced lipolysis and enhanced insulin sensitivity, and this observation coincided with higher translation of PLIN5 and PPARα in AT.To our knowledge, the present report is the first to characterize the dynamics of PLIN5 and PPARα expression in the AT of periparturient dairy cows.Long-chain FA have been shown to increase PLIN5 expression and lipid accumulation in primary rat hepatocytes (Grasselli et al., 2010).Similarly, in dairy cow hepatocytes, NEFA activate PLIN5 and promote TAG accumulation (Jia et al., 2019).For instance, within the adipocyte, OA may be re-esterified into TAG, especially when insulin sensitivity is increased as observed in this study.Like other perilipins, PLIN5 binds to the LD and protects it from lipolytic degradation.In different cell lines, it has been reported that PLIN5 binds to adipose triglyceride lipase (ATGL) on the LD and inhibits lipolysis under basal conditions (Wang et al., 2011a).In cardiac myocytes, OA promotes the interaction between PLIN5 and ABHD5 (α/β hydrolase domain-containing protein 5), mainly on the LD, and enhances lipogenesis by preventing access of lipolytic protein, such as ATGL, to stored TAG (Granneman et al., 2009).Further research is needed to confirm OAinduced PLIN5 expression and activation in AT of dairy cows, the mechanism behind its regulation of FA trafficking, and its interactions with the insulin signaling pathway.As for PPARα, currently, the literature shows conflicting results regarding the mechanism behind the effect of OA on this nuclear receptor.In HepG2 cells, PPARα appears to act upstream of PLIN5 to promote OA-induced lipid accumulation (Zhong et al., 2019).In contrast, in AML12 cells (Li et al., 2012) and primary mice hepatocytes (Najt et al., 2020), PLIN5 regulates PPARα expression.The inconsistent results observed could be related to the cell line used or the presence of a feedback loop between the 2 proteins.However, the mechanism of activation of the nuclear receptor PPARα in the presence of OA in AT is still unknown and will require gain and loss of function studies targeting PLIN5 and PPARα in bovine adipocytes.
The effect of OA on lipolysis was associated with a state of reduced inflammation in the AT as reflected by a decreased osteopontin protein expression.We have previously demonstrated that the expression of osteopontin is upregulated postpartum especially in cows with high rates of lipolysis (Contreras et al., 2017).Although we saw no difference between the 2 treatments with other cytokines such as IL6 and IL10, diets rich in OA have been shown to have beneficial effects on inflammatory-related diseases (extensively reviewed in Carrillo et al., 2012).For example, rodents fed diets containing high amounts of olive oil (>70% OA) exhibit an anti-inflammatory response (Yaqoob et al., 1994;Carrillo et al., 2012).The anti-inflammatory effect of OA in the AT of dairy cows requires further research as this study had a limited sample size and did not measure specific inflammatory biomarkers in blood, AT macrophages infiltration intensity, and the phenotype of immune cells trafficking into the AT.
There is evidence for mitochondrial dysfunction during the PP especially during NEB and inflammation (Swartz et al., 2021).Here we report that the OA treatment enhanced mitochondrial number and increase specific markers of mitochondrial function.The upregulation of SDHA as observed with the OA treatment in our study may indicate a positive effect on mitochondrial biogenesis.SDHA also plays an important role in mitochondrial function as it is a major catalytic subunit of succinate-ubiquinone oxidoreductase (Aldera and Govender, 2018).Oleic acid has been shown to upregulate genes related to FA oxidation, thus improving mitochondrial function and preventing insulin resistance (Rehman et al., 2020).In this study, this was reflected by enhanced expression of SOD2.Similarly, SIRT and PARP are 2 key regulators of mitochondrial metabolism.In white AT, PARP1 and 2 promote lipogenesis whereas SIRT1 inhibits it (Jokinen et al., 2017).In this study, OA decreased SIRT3, whereas Abou-Rjeileh et al.: OLEIC ACID AND ADIPOSE LIPID MOBILIZATION it increased PARP3 expression reflecting a state of enhanced lipogenesis.A possible mechanism for improved mitochondrial function is through the increased expression of PLIN5.In oxidative tissues, PLIN5 induces physical contact between LD and mitochondria (Wang et al., 2011b;Pribasnig et al., 2018).More recently, it has been shown that PLIN5 LD-mitochondrial coupling improves mitochondrial respiratory capacity and metabolic flexibility in human cells (Kien et al., 2022).Moreover, PLIN5 protects the mitochondria from a local surge in FA by promoting FA storage in LD and regulating LD hydrolysis during physiological stress (Wang et al., 2011b).Further research is required to validate OA involvement in mitochondrial function in AT of dairy cows through the activation of PLIN5.Also, it is important to emphasize that it is unknown if the effects mentioned above are exclusive to infusing OA or if infusing other unsaturated FA would result in similar outcomes.Therefore, future studies should focus on comparing OA against other FA, such as C18:2 and C18:3.On the other hand, OA can be produced through the desaturation of stearic acid by stearoyl-CoA desaturase.We focus our results and discussion on the effect of the exogenous source of OA.Whether endogenous OA has similar effects on AT metabolism requires further research.
We did not observe treatment effects on production response or nutrient digestibility variables.However, our study was primarily designed to investigate blood and molecular parameters rather than effects on production.In addition, we did not use FA supplements in the diets, which may have reduced the effect of OA as an emulsifier, resulting in a lack of effect on FA digestibility and, consequently, production responses.The tendency observed for increased total FA absorption with OA seems to be mainly related to the amount of FA from the infusate itself, and it was not sufficient to increase milk fat yield.On the other hand, a recent study by Prom et al. (2021) demonstrated that abomasally infusing OA up to 60 g/d linearly increased FA digestibility, 3.5% FCM, and ECM in cows receiving a diet containing a saturated FA supplement included at 1.8% DM.A meta-analysis that included calcium salts of palm FA (~46% C16:0 and ~38% OA) at ≤3% of diet DM tended to increase FA digestibility and improved production responses of dairy cows (dos Santos Neto et al., 2021).Other feeding studies using OA from FA blends also increased FA digestibility and milk fat yield of dairy cows while decreasing BW and BCS losses (de Souza et al., 2018(de Souza et al., , 2021a;;Prom and Lock, 2021).Further studies are needed to evaluate the effect of OA as an emulsifier in diets with different levels of FA supplementation.

CONCLUSIONS
In the present study, we provide evidence for 3 key mechanisms that explain the positive effect of OA on periparturient cow metabolic and lactation performance as this FA (1) limits FA trafficking out of AT, (2) improves systemic and AT insulin sensitivity, and (3) enhances mitochondrial biogenesis and may improve its function (Figure 6).These results demonstrate that OA may limit lipolysis by promoting lipogenesis during NEB periods such as the postpartum period in dairy cows.

Figure 6 .
Figure 6.Postpartum oleic acid (OA) infusions alter adipose tissue metabolism.Oleic acid limits fatty acid trafficking out of adipose tissue by limiting lipolysis, enhancing insulin sensitivity, and promoting lipogenesis, possibly through perilipin 5 (PLIN5) and peroxisome proliferatoractivated receptor α (PPARα) signaling.Oleic acid may also enhance mitochondrial biogenesis and function in adipose tissue.Created with BioRender.com.

Table 1 .
Ingredient and nutrient composition of diets