Subclinical ketosis leads to lipid metabolism disorder by down-regulating the expression of Acetyl-CoA acetyltransferase 2 (ACAT2) in dairy cows

Ketosis is a metabolic disease that often occurs in dairy cows postpartum and results from disordered lipid metabolism. Acetyl-coenzyme A (CoA) acetyltransferase 2 ( ACAT2 ) is important for balancing cholesterol and triglyceride ( TG ) metabolism; however, its role in subclinical ketotic dairy cows is unclear. This study aimed to explore the potential correlation between ACAT2 and lipid metabolism disorders in subclinical ketotic cows through in vitro and in vivo experiments. In the in vivo experiment, liver tissue and blood samples were collected from healthy cows ( CON , n = 6, β-hydroxybutyric acid [ BHBA ] concentration < 1.0 mM) and subclinical ketotic cows (subclinical ketosis [ SCK] , n = 6, BHBA concentration = 1.2–3.0 mM) to explore the effect of ACAT2 on lipid metabolism disorders in SCK cows. For the in vitro experiment, bovine hepatocytes ( BHEC ) were used as the model. The effects of BHBA on ACAT2 and lipid metabolism were investigated via BHBA concentration gradient experiments. Subsequently, the relation between ACAT2 and lipid metabolism disorder was explored by transfection with siRNA of ACAT2. Transcriptomics showed an upregulation of differentially expression genes ( DEGs ) during lipid metabolism and significantly lower ACAT2 mRNA levels in the SCK group. Compared with the CON group in vivo, the SCK group showed significantly higher expression levels of peroxisome proliferator–activated receptor γ ( PPARγ ) and sterol regulator element binding protein 1c ( SREBP1c ) and significantly lower expression levels of peroxisome proliferator-activated receptor α ( PPARα ), carnitine palmitoyl-transferase 1A ( CPT1A ), sterol regulatory element binding transcription factor 2 ( SREBP2 ), and 3-hydroxy-3-meth-ylglutaryl-CoA reductase ( HMGCR ). Moreover, the SCK group had a significantly higher liver TG content and significantly lower plasma total cholesterol ( TC ) and free cholesterol ( FC ) content. These results were indicative of TG and cholesterol metabolism disorders in the liver of dairy cows with SCK. Additionally, the SCK group showed an increased expression of Perili-pin-2 ( PLIN2 ), decreased expression of apolipoprotein B ( APOB ), and decreased plasma concentration of very low-density lipoproteins ( VLDL ) and low-density lipoproteins-cholesterol ( LDL-C ) by downregulating ACAT2, which indicated an accumulation of TG in liver. I n vitro experiments showed that BHBA induced an increase in the TG content of BHEC, decreased content TC, increased expression of PPARγ and SREBP1c, and decreased expression of PPARα, CPT1A, SREBP2, and HMGCR. Additionally, BHBA increased the expression of PLIN2 in BHEC, decreased the expression and fluorescence intensity of ACAT2, and decreased the VLDL and LDL-C contents. Furthermore, silencing ACAT2 expression increased the TG content; decreased the TC, VLDL, and LDL-C contents; decreased the expression of HMGCR and SREBP2; and increased the expression of SREBP1c; but had no effect on the expression of PLIN2. These results suggest that ACAT2 downregulation in BHEC promotes TG accumulation and inhibits cholesterol synthesis, leading to TG and cholesterol metabolic disorders. In conclusion, ACAT2 downregulation in the SCK group inhibited cholesterol synthesis, increased TG synthesis, and reduced the contents of VLDL and LDL-C, eventually leading to disordered TG and cholesterol metabolism.


INTRODUCTION
Ketosis, which commonly affects postpartum dairy cows, results from the negative energy balance in highyielding dairy cows with low energy intakes that cannot meet milk production requirements (Herdt, 2000).To satisfy energy demands, large amounts of lipids are mobilized, resulting in elevated nonesterified fatty acid Subclinical ketosis leads to lipid metabolism disorder by down-regulating the expression of Acetyl-CoA acetyltransferase 2 (ACAT2) in dairy cows Shendong Zhou, 1 Mengru Chen, 1 Meijuan Meng, 1 Nana Ma, 1 Wan Xie, 1 Xiangzhen Shen, 1 Zhixin Li, 2 and Guangjun Chang12* (NEFA) and β-hydroxybutyric acid (BHBA) concentrations in the blood, which in turn causes ketosis (Baird, 1982).Ketosis in dairy cows can be categorized as clinical and subclinical ketosis (SCK), with the clinical diagnosis mainly based on the detected BHBA blood concentration.SCK is usually considered for BHBA concentrations of 1.2-3.0mmol/L, while clinical ketosis is considered for BHBA concentrations ≥ 3.0 mmol/L (Ospina et al., 2010, Suthar et al., 2013, Shin et al., 2015).Previous studies have shown that ketosis increases the risk of other diseases such as abomasum displacement, mastitis, and lameness.(Dohoo and Martin, 1984, Sheldon et al., 2008, Duffield et al., 2009, Cheong et al., 2011).SCK poses great risks since its clinical symptoms are not obvious and often ignored (Berge and Vertenten, 2014, Overton et al., 2017, Yan et al., 2020).Therefore, further studies on the pathogenesis of SCK in dairy cows are essential for its control.
Many studies have reported a disturbance in lipid metabolism in ketotic cows, including hepatic triglyceride (TG) deposition and a decrease in cholesterol levels (Graber et al., 2010, Zhang et al., 2019, Klein et al., 2020).Cholesterol is a type of lipid involved in animal cell membrane structures and certain signaling pathways; its homeostasis is important in ketotic cows (Loor et al., 2007, Gross et al., 2015).Previous studies have shown that cholesterol demands during lactation are met by continuously upregulating the expression of cholesterol metabolic pathway enzymes from 3 weeks before delivery to 1 week after delivery, with an equilibrium at 5 weeks postnatal (Schlegel et al., 2012b).However, compared with healthy cows, ketotic cows have lower expression levels of the enzymes involved such as 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR, related to cholesterol synthesis), ATP binding cassette subfamily A member 1 (ABCA1, related to cholesterol transport), and ATP binding cassette subfamily G member 8 (ABCG8, related to bile acid metabolism) (Schlegel et al., 2012a, Gross et al., 2019, Wang et al., 2022).Therefore, cholesterol homeostasis regulation is crucial for cows with ketosis.
The liver, being the key hub of metabolism in the body, is not only the main site of cholesterol synthesis and metabolism but also participates in regulating the homeostasis of cholesterol (Luo et al., 2020).Acetyl-CoA acetyltransferase (ACAT) regulates cholesterol balance by converting free cholesterol (FC) to cholesteryl esters (CE) that are stored in cells (Chang et al., 1997).Intracellular CE participates in the formation of lipid droplets (LDs) with perilipins and phospholipids (Chang et al., 2001, Horton et al., 2003) and in the construction of apolipoproteins such as very low-density lipoprotein (VLDL) and low-density lipoprotein-cholesterol (LDL-C).These apolipoproteins transport TG from the liver to the whole body and are closely related to atherosclerosis and fatty liver disease (Liu et al., 2014, Gross et al., 2015, Geng et al., 2017).The 2 types of ACAT are ACAT1 and ACAT2; ACAT1 is widely expressed, while ACAT2 is expressed only in the liver and intestine and is therefore important for maintaining TG metabolic balance in the liver (Romeo, 2022).
Previous studies have shown that in porcine intramuscular preadipocytes, ACAT2 overexpression inhibits preadipocyte differentiation (Zhang et al., 2014, Wang et al., 2016).Additionally, ACAT2 inhibition in mice improves high-fat-diet-induced glucose intolerance (Pramfalk et al., 2022), while ACAT2 overexpression directly leads to atherosclerosis (Melchior et al., 2013, Nordestgaard, 2016).However, few studies have examined the regulation of lipid metabolism by ACAT2 in cows with SCK.Therefore, our objective was to investigate the relationship between ACAT2 and lipid metabolism disorders, including triglyceride and cholesterol metabolism disorders in subclinical ketosis dairy cows, thus providing a theoretical basis for the control of SCK in dairy cows.

Ethical Approval
The study strictly adhered to the guidelines of and was approved by the Animal Ethics Committee of Nanjing Agricultural University, China (NJAU.PZ2020102).Sampling was performed as per the specifications of the Animal Research Institute Committee of Nanjing Agricultural University, China.

Animal, Diet and Experimental Design
Twelve dairy cows, comprising 6 in the healthy control group (CON) and 6 in the SCK group, were selected for a follow-up experiment.The dairy cows in this study were chosen from the Modern Bengbu Pasture in Anhui, China.Ketone blood concentrations < 1.0 mM were considered healthy and concentrations of 1.2-3.0mM were indicative of SCK.The blood ketone concentrations of dairy cows in the pasture were detected using a cow blood ketone meter (Cat # eB-K03, Haiyi Technology Co., LTD., Beijing, China) and tail venous blood samples.Cows in the CON and SCK groups were numbered and randomly sampled.
All cows were fed a total mixed and allowed to drink freely.The nutritional composition of the diet, and the specific information about the cows are provided in Supplementary Table S1, and Supplementary Table S2 respectively.Additionally, on the 15th day after delivery, 5 mL of whole blood was collected using a

Sample Collection
The collected blood samples were transported on ice to the laboratory.The plasma was isolated from the anticoagulant blood samples by centrifugation at 3,000 × g at 4°C for 10 min.Plasma was stored at −80°C for the following analysis.
After anesthetizing the cows using 1% lidocaine hydrochloride (Jilin Huamu Animal Health Products Co., LTD, China), the site of the liver biopsy was chosen by tracing a line between the right shoulder joint and the ischial tubercle and determining its intersection between the penultimate ribs; approximately 500 mg of liver tissue was obtained from.The liver tissue was placed on a sterile gauze and washed with 1 × PBS while monitoring the removal of fat and blood, frozen in liquid nitrogen, transported to laboratory, and stored at −80°C.

Cell culture
Primary bovine hepatocytes (BHEC) were acquired from Professor Juan J. Loor (Department of Animal Sciences and Division of Nutritional Sciences, University of Illinois, Urbana).These cells were previously verified by our group and have been used as an in vitro model in many studies (Xue et al., 2022, Xie et al., 2023).
Since high BHBA concentration in the blood of dairy cows can lead to ketosis, the BHEC were treated with BHBA (Aladdin Bio-Chem Technology Co., Shanghai, China) to establish an in vitro model of ketotic cows.The cells were cultured in a complete medium containing 10% fetal bovine serum and 1% penicillin or streptomycin and incubated at 37°C in a 5% CO 2 incubator.This study used 4-8 generations of cells.

Cell Experiment Design
BHBA was diluted to 0, 1, 2, 4, 6, and 8 mM in PBS.Concentration and time gradients were used to optimize the BHBA treatment time and concentration.Next, BHEC were treated with 0-, 2-, and 4-mM BHBA for 12 h to investigate how BHBA affects ACAT2 expression and lipid metabolism.Finally, the relationship between ACAT2 and lipid metabolism was explored by transfecting a small interfering RNA (siRNA) of ACAT2 into BHEC.The TG, TC, VLDL, and LDL-C levels were determined.The expression levels of the genes and proteins related to TG and cholesterol metabolism were analyzed by real-time quantitative polymerase chain reaction (RT-qPCR), Western blotting, and immunofluorescence.

Cell Viability
Briefly, 10 4 BHEC were inoculated into each well of 96-well microplates and cultured for 24 h.The cells were then cultured with different concentrations of BHBA (0, 1, 2, 4, 6, and 8 mM) for 12 h or with BHBA (0, 2, and 4 mM) for 0, 3, 6, 12, 16, and 24 h.Each treatment was repeated 6 times.Next, 10 μL of cell counting kit-8 (CCK8) reagent was added to each well and incubated at 37°C in 5% CO 2 for 80 min.The optical density (OD) of each sample at 450 nm was measured using an enzyme labeling instrument (Thermo Fisher Scientific, United States).

Cell Transfection
The ACAT2 siRNA was designed and synthesized by Shanghai Sangong Biotechnology Co., Ltd.BHEC were cultured in 24-well microplates.When the cell density reached 70%, cells transfected with the ACAT2 siRNA were used as the treatment group (si-ACAT2), while those transfected with siRNA nonhomologous to BHEC comprised the negative control group (NC).After transfection with the siRNAs using Liposome 3000 according to the manufacturer's instructions, the BHEC were cultured for 48 h.

Immunofluorescence
For this analysis, a 14-mm diameter circular coverslip was placed into each well in a 24-well microplate for BHEC to adhere to.Each well was inoculated with 5 × 10 4 BHEC and cultured for 24 h, then cocultured with BHBA (0, 2, and 4 mM) for 12 h.The cells were fixed by adding 500 μL of 4% paraformaldehyde (P6148, Sigma-Aldrich) to each well for 20 min and washed 3 times (5 min each) with PBS.The cell permeability was improved by adding 500 μL of 0.3% Triton X-100 (T8200, Solarbio, Beijing, China) to each well, incubating at room temperature for 15 min, and washing with PBS (3 washes, 5 min each).Next, 500 μL of 5% bovine serum albumin was added to each well and incubated at 37°C for 1 h.A primary ACAT2 antibody (1:200; Cohension, Shanghai, China) was then added to each well, incubated at 4°C for 12 h, and washed with PBS (3 washes, 5 min each).Following this, a Cy3-labeled secondary antibody (1:500, A0562; Beyotime, Shanghai, China) was added and incubated at 37°C for 1 h in the dark.The cells were washed with

ELISA Determination of NEFA, INS, GC and VLDL in plasma or cell supernatant
We analyzed the concentration of NEFA, insulin (INS), glucagon (GC) and VLDL in plasma or cell supernatant by enzyme-linked immunosorbent assay (ELISA).The above indicators were tested in accordance with the instructions of kits (Yihe Biotechnology Co., Ltd., Shanghai, China).The detection of bovine NEFA by ELISA is described below as an example; the remaining assays were all performed in a similar manner.First, all the reagents were placed at room temperature (25°C) for 20 min and shaken evenly to ensure that they were fully dissolved.The calibration standards (50 μL each) and the samples (50 μL each, diluted to 1:4 with the sample diluent) were added to a 96-well plate.A NEFA detection antibody labeled with horseradish peroxidase (HRP) was then added to each calibration standard and sample well, covered with a sealed film, and left for 60 min at 37°C.The supernatants were then discarded and the wells washed 5 times with the washing liquid.After the last wash, the reaction termination solution was added, which changed the color from blue to yellow.The color intensity of the solution at 450 nm was measured by enzyme labeling (Thermo Fisher Scientific, United States).The concentration of the standard and its corresponding OD was used to draw a standard linear regression curve.The NEFA concentration of each sample was subsequently calculated using the regression curve.
The plasma concentration of INS, GC, and VLDL were each analyzed in a similar manner and the specific details of the ELISA kits are provided in Supplementary Table S3.

Determination of BHBA, FC, TC, TG, and LDL-C Contents in Plasma or Cell Supernatants
The plasma concentration of BHBA was detected by the water-soluble tetrazolium 1 method as per the instructions of a β-hydroxybutyric acid (BHBA) content assay kit (Cat # BC5085, Sloarbio, Beijing, China).Free cholesterol (FC) plasma concentrations determined using an automatic biochemical analyzer (MNCHIP, Celercare V5, Tianjing, China).The TG contents in the liver tissues and cell supernatants were measured as per the instructions of a TG content assay kit (AKFA003M, Beijing Boxbio Science & Technology Co., Ltd., China).The total cholesterol (TC) and LDL-C contents in the plasma and cell supernatants were measured according to the corresponding TC assay kit (Cat # A111-1-1, Jiancheng Bioengineering Institute, Nanjing, China) and LDL-C assay kit (Cat # A113-2-1, Jiancheng Bioengineering Institute, Nanjing, China) instructions.

RNA Extraction and Transcriptome Sequencing
Trizol (Cat # 9108, TaKaRa, Shiga, Japan) was used to extract the total RNA from the liver samples.A spectrophotometer (NanoDrop 2000, Thermo Fisher Science, United States) was used to measure the RNA concentration and purity, and ratios of absorbance at 260 and 280 nm of 1.9-2.1 were used for subsequent experiments.The RNA integrity was evaluated using an RNA Nano 6000 analysis kit (Agilent Technologies, CA, United States).Following the quality inspection, all RNA samples were frozen in liquid nitrogen and shipped to Beijing Baimaike Biotechnology Co., Ltd., to build complementary DNA (cDNA) libraries, the effective concentrations of which were quantified by RT-qPCR (effective library concentration > 2 nM).Library sequencing was performed by BioMarker Technologies (Beijing, China) using a HiSeq X system (Illumina, Unit States).The reads containing joints and those of low quality (reads with an N ratio > 10% and those with > 50% of the total reads with a mass value Q ≤ 10) were removed to acquire clean, high-quality data.
The differentially expressed genes (DEGs) between the CON and SCK groups (Fold change >2 and false discovery rate <0.01) were screened and analyzed using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment.

RT-qPCR Analysis
An Evo M-MLV Mix Kit with gDNA Clean for qPCR (Cat # AG11728, Accurate Biotechnology Co., Ltd., Changsha, China) was used to reverse transcribe 1 μg of the total RNA into cDNA.Primer sequences of the following genes were designed online using the Primer BLAST tool from the NCBI and validated for usability (Supplementary Table S4): genes associated with lipid synthesis, namely peroxisome proliferator-activated receptor γ (PPARγ) and sterol regulatory element binding factor 1c (SREBP1c); genes associated with lipolysis, namely proliferator-activated receptor α (PPARα) and carnitine palmitoyltransferase 1A (CPT1A); and genes associated with cholesterol metabolism, namely apolipoprotein B (APOB), ACAT2, Zhou et al.: SUBCLINICAL KETOSIS AND LIVER HEALTH OF COWS SREBP2, HMGCR, and perilipin 2 (PLIN2).An qPCR SYBR Green Master Mix (Cat # 11201ES03, Yeasen Biotechnology Co., Ltd., Shanghai, China) and ABI 7300 fast RT-PCR system (Applied Biosystems, United States) were used to perform RT-qPCR, data were analyzed using the 2 −ΔΔCt method, and β actin (ACTB) was used as the reference gene for relative mRNA abundance calculations.

Western Blotting Analysis
Liver tissue samples (100 mg) were pulverized with liquid nitrogen and subsequently mixed with 1 mL of a radioimmunoprecipitation assay lysis buffer (Beyotime Biotechnology Co. Ltd., Shanghai, China) and 1 mm of phenylmethylsulfonyl fluoride and incubated on ice for 20 min to extract the total protein.The sample was then centrifuged at 12,000 rpm at 4°C for 5 min and the supernatant removed to measure the total protein concentration using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific, United States).The protein sample was subsequently diluted to 7 μg/mL.A sodium dodecyl sulfate (SDS, 5 × ) protein loading buffer was added to the total protein and denatured in a 99°C water bath for 5 min, following which all proteins were separated by 10% acrylamide SDS-PAGE (Category number [Cat#] PG112, Epizyme, Shanghai, China) and subsequently transferred to a 0.2-μm methanol-activated polyvinylidene fluoride membrane (Bio-Rad, United States).The membrane was blocked at room temperature for 2 h with 7% skim milk and incubated with diluted primary antibodies overnight at 4°C.The primary antibodies used, which were ensured to be homologous to bovine by > 90%, were as follows: PPARγ (1:500, Cat# WL01800) and PPARα (1:1000, Cat# WL00978) purchased from Wanleibio (Shenyang, China); SREBP1c (1:1000, Cat# ab28481) and APOB (1:1000, Cat #AF6213) purchased from Beyotime Institute of Biotechnology (Shanghai, China); CPT1A (1:1000, Cat# DF12004) purchased from Affinity Biosciences (United States); and ACAT2 (1:1000, Cat# CQA1677), SREBP2 (1:1000, Cat# CQA6043), HMGCR (1:1000, Cat# CQA1064), and PLIN2 (1:1000, Cat# CQA2156) purchased from Cohension Bio (Shanghai, China).The membranes were subsequently washed with 1 × tris-buffered saline (TBST), incubated with the secondary antibodies corresponding to the primary antibodies for 2 h at room temperature, and washed with 1 × TBST.The membranes were immersed in an ECL Plus Western blotting substrate (Vazyme, Nanjing, China) and the strips were imaged using a ChemiDoc MP system (Bio-Rad, United States).The protein band intensities were analyzed using Image Lab (version 5.2.1.62311,Bio-Rad, United States).ACTB (1:1000, Cat # AF7018, Affinity, United States) was chosen as the reference protein to normalize the target protein, which was performed using the ratio of the intensity of the target protein to that of ACTB.
Data from the animal experiments and BHEC transfection experiments were compared using an independent sample t-test and data from the cell culture assays were analyzed using one-way ANOVA with Dunnett's post hoc test.All results are expressed as the mean ± standard deviation.The residuals of each variable were tested for normality distribution using Shapiro-Wilk and joint hypotheses tests using Levene's test.Statistical significance was set at P < 0.05, with P < 0.01 representing extremely significant differences.Data were visualized using Graph Pad Prism (version 8.0, Graph Pad Software, La Jolla, United States).

Plasma Metabolism Indices
As shown in Figure 1, the blood metabolic indicators signaling the development of ketosis were dramatically altered in the SCK group compared with those of the CON group (Figure 1A).The plasma concentrations of BHBA (P < 0.01) and GC (P < 0.05) were significantly higher in the SCK group than in the CON group.The INS plasma concentration of the CON group was significantly lower than that of the SCK group (P < 0.01) (Figures 1C and D).

Liver Transcriptomics Data
As shown in Supplementary Table S5, Q30 (A higher Q30 represents a good detection quality of the sample) ≥ 93.27% in the CON group and Q30 ≥ 93.91% in the SCK group, which showed that these samples had low sequencing error rates and were of high quality.The average of the total reads, mapped reads, GC content, and mapped rate were 38396450, 37760348, 48.69%, and 96.54% in the CON group; and 41499852, 39958374, 47.67%, and 96.28% in the SCK group, respectively.A heatmap was generated after the hierarchical clustering analysis of the DEG expression profiles in the SCK and CON groups (Supplementary Fig. S1A).The colors indicate different DEG expression levels, with blue and red indicating gene expression downregulation and upregulation, respectively.The resulting heatmap also showed the same DEG expression cluster in each group, indicating a good sample biological repetition.The volcano plot showed a total of 55 DEGs (30 upregulated and 25 downregulated) in the SCK group compared with the CON group (Supplementary Figs.S1B and  C).

GO and KEGG Pathway Enrichment Analysis of DEGs
GO enrichment analysis classified the DEGs using 3 GO categories: biological process (BP), cellular component (CC), and molecular function (MF).The DEGs were enriched into 17 BP terms, 13 CC terms, and 6 MF terms (Supplementary Figure 2A) by comparing the CON and SCK groups.Compared with the CON group, the SCK group exhibited 4 significantly enriched BP terms involved in lipid metabolism: positive regulation of fatty acid biosynthetic process, VLDL particle assembly, cholesterol esterification, and cholesterol metabolism (Supplementary Figure 2B).Compared with the CON group, the SCK group exhibited one significantly enriched CC term involved in lipid metabolism, namely mitochondrial small ribosomal subunit (Supplementary Figure 2C), and 3 significantly enriched MF terms involved in cholesterol metabolism, namely cholesterol monooxygenase (side-chain cleaving), cholesterol O-acyltransferase, and sphingosine-1-phosphate receptor activities (Supplementary Figure 2D).The top 20 significant KEGG pathways following the KEGG enrichment analysis of the DEGs showed 3 terms involved in lipid metabolism: cholesterol metabolism, PPAR signaling pathway, and glucagon signaling pathway (Supplementary Figure 3).
These results show that the DEGs between the CON and SCK groups were closely associated with lipid metabolism, including the TG metabolism and cholesterol metabolic pathways.

Disturbed Cholesterol and TG Metabolism in SCK Livers
Transcriptomics show that the DEGs between the CON and SCK groups were mainly enriched in the lipid metabolism pathways, including TG and cholesterol metabolism.Therefore, we used RT-qPCR and Western blotting to analyze the lipid metabolism pathways in dairy cows.
Figure 3 shows that, compared with the CON group, the SCK group had significantly higher liver TG contents (Figure 3A), relative PPARG and SREBP1c mRNA expression levels (P < 0.01), and relative protein abundance of PPARγ and SREBP1c (P < 0.01), while the relative PPARα and CPT1A mRNA expression levels (P < 0.01; Figure 3B) and relative protein abundance of PPARα (P < 0.01) and CPT1A (P < 0.05) were significantly lower (Figures 3D and E).
These results showed a decrease in cholesterol synthesis, increase in TG synthesis, and inhibition of fatty acid β-oxidation in the SCK cows compared with the CON cows, indicating the occurrence of cholesterol and TG metabolic disorders, which was consistent with the transcriptomics results.

ACAT2 Expression and Its Downstream Factors in SCK
Since ACAT2 was one of the more interesting DEGs obtained from the transcriptome analysis, RT-qPCR and Western blot analyses were used to confirm the relationship between ACAT2 and lipid metabolism in the SCK group.The results showed that the relative mRNA and protein expression levels of ACAT2 were significantly lower in the SCK group compared with Comparisons between 2 groups were calculated using independent sample t-test.For all bar plots shown, blue represents CON group and red represents SCK group, and the data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).
those of the CON group (P < 0.01; Figures 4A-C).These results were consistent with the transcriptomics results.
To explore the role of ACAT2 in regulating lipid metabolism, we examined the downstream of CE mediated by ACAT2 in the body, including the mRNA and protein expression levels of APOB and PLIN2, as well as the plasma concentrations of VLDL and LDL-C.The results showed that compared with the CON group, the relative mRNA and protein levels of PLIN2 in the livers of the SCK group were significantly increased (P < 0.01), while the relative mRNA and protein levels of APOB were significantly decreased (P < 0.01; Fig- ures 4A-C).Additionally, the plasma concentrations of VLDL-C and LDL-C were significant lower in the SCK group compared with CON group (P < 0.01; Figure 4D).
These results indicated that ACAT2 downregulation in the SCK group promoted the formation of LDs and inhibited the transport of TG from the liver by reducing the synthesis of VLDL and LDL-C, thereby aggravating the disordered TG metabolism.

BHBA-Induced Lipid Metabolism Disturbance in BHEC
First, we used concentration and time gradient experiments to optimize the BHBA treatment of dairy cow hepatocytes.CCK-8 assay results showed that treating BHEC with 6 and 8 mM of BHBA for 12 h significantly reduced the cell viability (P < 0.01; Figure 5A).The cell viability was also significantly reduced in BHEC treated with 2 mM of BHBA for 24 h (P < 0.05; Figure 5B) and 4 mM of BHBA for 18 h (P < 0.01; Figure 5C).Based on these results and the clinical diagnostic criteria for blood ketone concentrations in cows with clinical and SCK, we selected BHEC treated with BHBA at 0, 2, and 4 mM for 12 h for the subsequent experiments.
As shown in Figure 6, compared with 0 mM, both 2-and 4-mM BHBA treatments caused a significant increase in TG concentration (P < 0.01) and a very significant decrease in TC concentration in BHEC (P < 0.01; Figure 6A).Compared with 0 mM, both 2and 4-mM BHBA treatments significantly decreased the relative mRNA and protein expression levels of SREBP2, HMGCR, PPARα, and CPT1A (P < 0.01, Figures 6B-E) and significantly increased the relative mRNA and protein levels of PPARγ and SREBP1c in BHEC (P < 0.01; Figures 6C-E).
These results demonstrated that BHBA leads to disordered TG and cholesterol metabolism in BHEC, which can be aggravated with increasing BHBA concentrations.

Effect on the Expression of ACAT2 and Its Downstream Factors in BHEC Induced by BHBA
As shown in Figure 7, compared with 0 mM, both 2-and 4-mM BHBA treatments significantly decreased the relative mRNA and protein expression levels of ACAT2 in BHEC (P < 0.01; Figures 7A and B) and significantly increased the relative expression of PLIN2.Additionally, the 2-mM BHBA treatment significantly increased the relative protein abundance of PLIN2 in BHEC (P < 0.05) and 4 mM of BHBA significantly increased the relative protein abundance of PLIN2 in BHEC (P < 0.01; Figure 7B).Furthermore, compared with the 0-mM treatment, both 2-and 4-mM of BHBA decreased the fluorescence intensity of ACAT2 (Figure 7C) and significantly decreased the LDL-C and VLDL contents in BHEC (P < 0.01).

Effect of ACAT2 Silencing on Lipid Metabolism and ACAT2 Downstream Factor Expression in BHEC
To verify the efficiency of si-ACAT2, we examined the expression of ACAT2 in BHEC by RT-qPCR and immunofluorescence.Compared with the NC group, the si-ACAT2 group showed significantly lower relative ACAT2 expression levels (P < 0.01; Figure 8A As shown in Figure 9, compared with the NC group, the si-ACAT2 group had significantly higher TG contents (P < 0.01) and significantly lower TC contents (P < 0.05; Figure 9A).The relative expression levels of HMGCR and SREBP2 were significantly lower (P < 0.01) and the relative expression of SREBP1c was significantly higher (P < 0.01) in the si-ACAT2 group compared with those of the NC group.Additionally, the VLDL and LDL-C contents in the si-ACAT2 group were significantly lower than those of the NC group (P < 0.01).
These results suggest that ACAT2 silencing promotes TG synthesis in BHEC and inhibits the outward transport of TG in cells by reducing the synthesis of VLDL and LDL-C, which eventually leads disordered TG metabolism.Moreover, ACAT2 silencing leads to disordered cholesterol metabolism by reducing cholesterol synthesis.In conclusion, ACAT2 silencing causes lipid metabolism disorders in BHEC.

DISCUSSION
The peripartum period is a demanding phase, with studies showing that peripartum cows with a negative energy balance are at risk of developing ketosis (Es- Comparisons between 2 groups were calculated using independent sample t-test.For all bar plots shown, blue represents CON group and red represents SCK group, and the data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).posito et al., 2014posito et al., , Shin et al., 2015)), especially in highyield cows, which have a high incidence of > 50% (Du et al., 2018, Dong et al., 2022).This period typically includes low glucose and INS contents and high GC contents in the blood (Han van der Kolk et al., 2017).To meet the glucose requirements during lactation, the body reduces the glucose requirements for peripheral organs other than the mammary gland by decreasing the blood INS content, and also promotes adipose tissue decomposition through the upregulation of blood GC (Baird, 1982, Zhang et al., 2019), thus increasing the blood concentration of NEFA (Oetzel, 2004, Gross et al., 2013).NEFA accumulation leads to an excess of acetyl-CoA, which is used to synthesize ketone bodies such as BHBA, and eventually leads to ketosis in dairy cows (Esposito et al., 2014, Han van der Kolk et al., 2017).Above-average NEFA and BHBA concentrations can cause impaired immune function and an increased susceptibility to other diseases such as abomasum displacement, metritis, and mastitis (Scalia et al., 2006, Stengarde et al., 2008, Moyes et al., 2009).Additionally, NEFA and BHBA concentrations beyond the normal range also reduce ovarian sensitivity to luteinizing hormone and follicle-stimulating hormone, leading to oocyte toxicity and impaired fertility in dairy cows (Dohoo and Martin, 1984, Leroy et al., 2011, Miqueo et al., 2019).Therefore, the prevention and control of ketosis in cows is crucial, especially in those with SCK.
Numerous studies have clearly demonstrated the occurrence of disorders of lipid metabolism in ketotic cows and the risk of higher fat accumulation in the liver (Herdt, 2000, Suthar et al., 2013, Esposito et al., 2014).PPARγ and SREBP1c, 2 important upstream factors that participate in lipid metabolism regulation (Nafikov et al., 2013), promote the expression of en-zymes associated with TG synthesis (Chen et al., 2019, Yang et al., 2022).PPARα catabolizes lipids by regulating lipid transport in the liver and β-oxidation (Xie et al., 2020), with PPARα and CPT1A inhibition in the livers of ketotic cows (Han van der Kolk et al., 2017, Alharthi et al., 2018).We showed a significant upregulation in the mRNA and protein expressions of PPARγ and SREBP1c and a significant downregulation in the mRNA and protein expressions of PPARα and CPT1A in the SCK group, indicating increased fat synthesis and decreased fat oxidation in the liver, leading to a TG disorder in the liver.Furthermore, we showed that BHBA induces an increase in the mRNA and protein expressions of PPARγ and SREBP1c, and a decrease in the mRNA and protein expressions of PPARα and CPT1A in BHEC, leading to a disorder of TG metabolism in BHEC.In summary, both in vivo and in vitro assays confirmed the transcriptomics results.
Previous studies have shown that lactation in mammals leads to an increase in cholesterol demand (Smith et al., 1998, Athippozhy et al., 2011).The expression levels of proteins related to cholesterol metabolism in cows gradually increase and stabilize during lactation, thus maintaining cholesterol homeostasis in the body (Espenshade and Hughes, 2007, Graber et al., 2010, Schlegel et al., 2012b).However, we found that compared with the CON group, the mRNA and protein expressions of SREBP2 and HMGCR in the SCK group were significantly decreased, indicating the occurrence of severe cholesterol metabolic disorders or a longer duration to restore cholesterol homeostasis in the SCK group.We also showed that BHBA induced a reduction in TC content and downregulation of HMGCR and SREBP2 expression in BHEC, suggesting that BHBA can cause cholesterol metabolic disorders in BHEC.Previous studies have shown reduced TC levels and cholesterol synthesis in cows with ketosis, which are consistent with the results of this study (Loor et al., 2007, Schlegel et al., 2012b, Wang et al., 2022).
ACAT2 is located in the endoplasmic reticulum and is highly sensitive to cholesterol concentrations.Its expression level is important for lipid metabolism regulation (Buhman et al., 2000).Relevant studies have shown that ACAT2 overexpression leads to fatty livers in mice (Nordestgaard, 2016) and that silencing ACAT2 effectively alleviates fatty livers induced by high-fat diets (Buhman et al., 2000, Willner et al., 2003, Romeo, 2022).Transcriptome sequencing reveal that the DEGs between the CON and SCK groups were significantly enriched during cholesterol metabolism and that ACAT2 expression in the SCK group was significantly downregulated compared with that of the CON group.Additionally, the livers of ketotic cows usually displayed physiological abnormalities.Interestingly, ACAT2 is specifically expressed in the liver, which suggests that it is important to study ACAT2 in the livers of ketotic cows (TY et al., 1997).Moreover, we determined the mRNA and protein expression levels of ACAT2 in the livers of the SCK group, which were significantly downregulated compared with those of the CON group.BHBA also reduced the mRNA and protein levels of ACAT2 and the fluorescence intensity of ACAT2 in BHEC.Similar results have been reported previously (Schlegel et al., 2012a, Wang et al., 2022).
Studies have shown that ACAT2 is degraded by the ubiquitination of its Cys277 site, which can be prevented by the oxidation of Cys277 by FC and free fatty acids (Wang et al., 2017).Therefore, the low ACAT2 expression levels in this study may be related to the decreased FC levels.Additionally, ACAT2 participates in lipid homeostasis regulation mainly through the esterification of FC to CE (Luo et al., 2020, Li et al., 2021).CE synthesized in the liver forms LDs with phospholipids and PLIN2 (Chang et al., 2001, Horton et al., 2003, Sastre et al., 2014) and participates in VLDL and LDL synthesis in the liver (Gross et al., 2019).TG is primarily transported from the liver to the circulating blood by VLDL and LDL-C, thus relieving hepatic fat deposits; LDL concentrations can also be determined by LDL-C measurements (Stone et al., 1987, Luo et al., 2020).However, in this experiment, the ACAT2 downregulation in the SCK group was accompanied by an increased PLIN2 expression and fat deposition in the liver and significant reductions in APOB expression (the assembled precursors of VLDL and LDL) and VLDL and LDL-C plasma concentrations compared with the CON group, leading to a blockage in TG transport in the liver.BHBA also upregulated PLIN2 expression and downregulated VLDL and LDL-C in BHEC.These results indicate that BHBA increases LD synthesis and inhibits VLDL and LDL-C syntheses by downregulating ACAT2 expression in BHEC, which eventually aggravates the disorder of TG metabolism.(E) The concentration of VLDL and LDL-C in supernatants of BHEC.Comparisons between 2 groups were calculated using independent sample t-test.The data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).
To further explore the role of ACAT2 in lipid metabolism regulation in subclinical ketotic cows, we silenced ACAT2 expression through siRNA transfection and assayed the related TG and cholesterol metabolism indicators.Silencing ACAT2 expression increased the TG content and decreased the TC content in BHEC and caused cholesterol metabolism disorders by downregulating HMGCR and SREBP2 expression.Additionally, silencing ACAT2 expression also upregulated SREBP1c expression in BHEC, did not affect the expression of PLIN2, and decreased the contents of VLDL and LDL-C.These results suggest that TG metabolic disorders in cows with SCK are caused by ACAT2 through TG synthesis upregulation and VLDL and LDL-C inhibition.This was consistent with a previous study on ACAT2 knockout mice that also observed the occurrence of lipid metabolism disorder in the liver (Pramfalk et al., 2022).
SREBPs are important for the activation of cholesterol and fatty acid synthesis, where SREBP2 and SREBP1c activate the transcription of genes involved in cholesterol and fatty acid synthesis, respectively (DeBose-Boyd and Ye, 2018, Luo et al., 2020, Wang et al., 2020).We observed that the decreased ACAT2 expression was accompanied by an increased SREBP1c expression and decreased SREBP2 expression both in vivo and in vitro.We hypothesized that ACAT2 downregulation in SCK affects TG and cholesterol metabolism by regulating SREBP2 and SREBP1c expression.Previous studies have shown that the contents of VLDL and LDL-C are mainly regulated by SREBPs (Strong et al., 2012, Al-Eitan et al., 2020, Zhang et al., 2020).A future study will therefore test the above hypothesis by regulating the expressions of ACAT2 and SREBPs.

CONCLUSION
Compared with healthy cows, subclinical ketotic cows demonstrated a reduction in cholesterol synthesis, increase in fat synthesis in the liver, and reduction in VLDL and LDL-C synthesis to inhibit TG outward transport through ACAT2 downregulation, which eventually led to TG and cholesterol metabolism disorders.
Zhou et al.: SUBCLINICAL KETOSIS AND LIVER HEALTH OF COWS vacuum blood collection system (Jiangsu Shoude Medical Instrument, Nanjing, China) with the anticoagulant heparin sodium and liver tissue samples collected by liver biopsies.
Zhou et al.: SUBCLINICAL KETOSIS AND LIVER HEALTH OF COWSPBS, incubated with a diamidino-2-phenylindole solution (D8417; Sigma-Aldrich, Shanghai, China) for 15 min at room temperature, and washed with PBS again.Finally, the cells were removed, fixed onto slides, and the fluorescence intensity of ACAT2 in the cells was observed by confocal laser scanning microscopy (LSM 710; Zeiss, Oberkochen, Germany).
Zhou et al.: SUBCLINICAL KETOSIS AND LIVER HEALTH OF COWS

Figure 1 .
Figure 1.The analysis of blood metabolic indicators in dairy cows.(A-D) The concentration of BHBA, NEFA, INS and GC in the plasma of dairy cows.Comparisons between 2 groups were calculated using independent sample t-test.For all bar plots shown, blue represents CON group and red represents SCK group, and the data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).
Figure 2. The analysis of cholesterol metabolism in dairy cows.(A) The concentration of FC and TC in the plasma of dairy cows; (B) The relative mRNA expression abundance of SREBP2 and HMGCR.(C-D) The relative protein expression abundance of SREBP2 and HMGCR.For all bar plots shown, blue represents CON group and red represents SCK group, and the data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).
Figure 3.The analysis of TG metabolism in dairy cows.(A) The concentration of TG in the plasma of dairy cows; (B) The relative mRNA abundance of PPARG and SREBP1c; (C) The relative mRNA abundance of PPARα and CPT1A; (D) The relative protein expression abundance of PPARG, SREBP1c, PPARα and CPT1A.Comparisons between 2 groups were calculated using independent sample t-test.For all bar plots shown, blue represents CON group and red represents SCK group, and the data are expressed as mean ± SEM, ** represents extremely significant difference between the 2 groups (P < 0.01).
) and lower fluorescence intensity of ACAT2 in BHEC (Fig-ure 8B), indicating the effective silencing of ACAT2 expression in BHEC.
Figure 4.The expression analysis of ACAT2 and its downstream factors in dairy cows.(A) The relative mRNA expression abundance of ACAT2, PLIN2 and APOB; (B-C) The relative protein expression abundance of ACAT2, PLIN2 and APOB; (D) The concentration of LDL-C and VLDL in the plasma of dairy cows.Comparisons between 2 groups were calculated using independent sample t-test.For all bar plots shown, blue represents CON group and red represents SCK group, and the data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).
Figure 5.Effect of BHBA on BHEC cell viability.(A) Effects of BHBA treatment with 0, 1, 2, 4, 6, and 8 mM on BHEC cell viability for 12 h; (B-C) Effect of treatment with 2 mM and 4 mM BHBA for 12 h on the viability of BHEC cells.Comparisons between groups were calculated using One-way ANOVA with Dunnett's post hoc test.The data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).

Figure 6 .
Figure 6.Effect on TG metabolism and cholesterol metabolism in BHEC induced by BHBA.(A) The concentration of TG and TC in supernatants of BHEC; (B) The relative mRNA expression abundance of SREBP2 and HMGCR; (B-C) The relative protein expression abundance of PPARG, SREBP1c, PPARα and CPT1A.Comparisons between groups were calculated using One-way ANOVA with Dunnett's post hoc test.The data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).

Figure 7 .
Figure 7. Effect on the expression of ACAT2 and its downstream factors induced by BHBA.(A-B) The relative mRNA and protein expression abundance of PLIN2 and APOB; (C) Fluorescence intensity of ACAT2 in BHEC, the scale bar was 10 μm.(D-E) The concentration of LDL-C and VLDL in supernatants of BHEC.Comparisons between groups were calculated using One-way ANOVA with Dunnett's post hoc test.The data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).
Figure 8. Verification of the silencing efficiency of ACAT2.(A) The relative mRNA expression abundance of ACAT2; (B) Fluorescence intensity of ACAT2 in BHEC, the scale bar was 10 μm.Comparisons between 2 groups were calculated using independent sample t-test.The data are expressed as mean ± SEM, ** represents extremely significant difference between the 2 groups (P < 0.01).

Figure 9 .
Figure 9.Effect of silencing ACAT2 expression on lipid metabolism and expression of ACAT2 downstream factors in BHEC.(A) The concentration of TG in supernatants of BHEC; (B) The relative mRNA expression abundance of HMGCR and SREBP2; (B) The relative mRNA expression abundance of PPARG, SREBP1c, PPARα and CPT1A; (C) The relative mRNA expression abundance of PLIN2.(E)The concentration of VLDL and LDL-C in supernatants of BHEC.Comparisons between 2 groups were calculated using independent sample t-test.The data are expressed as mean ± SEM, * represents significant difference between the 2 groups (P < 0.05), ** represents extremely significant difference between the 2 groups (P < 0.01).