| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
activation and long-chain fatty acids alter lipogenic gene networks in bovine mammary epithelial cells to various extents
,
,,1
* Department of Animal and Avian Sciences, University of Maryland, College Park 20742
Mammalian NutriPhysioGenomics, Department of Animal Sciences, and
Division of Nutritional Sciences, University of Illinois, Urbana 61801
1 Corresponding author: jloor{at}illinois.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
(PPARG), which, along with its lipogenic target genes, is upregulated in bovine mammary tissue during lactation. Thus, PPARG might represent an important control point of bovine milk fat synthesis. We tested lipogenic gene network expression via quantitative PCR of 19 genes in bovine mammary epithelial cells cultured with 16:0, 18:0, cis-9 18:1, trans-10 18:1, trans-10,cis-12 18:2 [t10c12 conjugated linoleic acid (CLA)], 20:5, ethanol (control), and the PPARG agonist rosiglitazone (ROSI). Triplicate cultures were maintained for 12 h with 50 µM ROSI or 100 µM LCFA. Responses common to 16:0 and 18:0 relative to the control included significantly greater expression of INSIG1 (+298%, +92%), AGPAT6 (+137%, +169%), FABP3 (+755%, +338%), and FABP4 (+171%, 157%). These were coupled with greater intracellular lipid droplet formation and mRNA of ACSS2, LPIN1, SCD, and SREBF2 in response to 16:0, and greater DGAT1 and THRSP with 18:0. Trans-10 18:1 and t10c12 CLA reduced expression of FASN (–60%, –31%), SCD (–100%, –357%), and SREBF1 (–49%, –189%). Furthermore, t10c12 CLA downregulated ACSS2, FABP3, INSIG1, SREBF2, and THRSP expression. Expression of SREBF1 was lower with cis-9 18:1 (–140%) and 20:5 (–125%) compared with the control. This latter LCFA also decreased SCD, SREBF2, and LPL expression. No effects of LCFA or ROSI on PPARG were observed, but ROSI upregulated (+39% to +269%) expression of ACACA, FASN, LPIN1, AGPAT6, DGAT1, SREBF1, SREBF2, and INSIG1. Thus, these genes are putative PPARG target genes in bovine mammary cells. This is the first report showing a direct effect of trans-10 18:1 on bovine mammary cell lipogenic gene expression. The coordinated upregulation of lipogenic gene networks in response to ROSI and saturated LCFA offers support for PPARG activation in regulating bovine milk fat synthesis.
Key Words: nuclear receptor • milk fat • genomics
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
In nonruminants, most LCFA, and specifically polyunsaturated fatty acids (PUFA), are natural ligands and bind to peroxisome proliferator-activated receptor- (PPARG), eliciting changes in gene expression and rates of lipogenesis (Bensinger and Tontonoz, 2008, Berger and Moller, 2002, Desvergne et al., 2006). Messenger RNA expression of PPARG and several potential target genes in mammary tissue was upregulated (Bionaz and Loor, 2008b) at the onset and throughout lactation in dairy cows. Therefore, it was proposed that milk fat synthesis regulation via exogenous or mammary-synthesized LCFA ligands (e.g., palmitic acid) could take place through PPARG-mediated changes in gene expression.
There is a high correlation between mRNA expression of target genes and recruitment of lipogenic transcription factors and their coregulatory proteins to promoter regions, suggesting that gene expression analysis is useful for inferring transcriptional activity (Bennett et al., 2008). In addition, an increase in metabolic flux requires overexpression of most of the enzymes in a biosynthetic pathway, challenging the idea of a "limiting enzyme" (Morandini et al., 2005). This is an important consideration when interpreting data from transcriptomics studies. Therefore, measurement of mRNA for multiple genes and their networks in a pathway is essential.
Our central hypothesis was that activation of PPARG in bovine mammary cells would upregulate mRNA expression of lipogenic target genes identified primarily in rodent studies. Furthermore, we hypothesized that different LCFA regulate expression of genes controlling networks of lipid metabolism to various extents, and potentially through PPARG. The specific objectives were to determine the effects of PPARG activation as well as exogenous saturated and unsaturated LCFA on mRNA expression of genes that compose an interactive network controlling in vivo bovine mammary cell lipid metabolism (Bionaz and Loor, 2008a,b).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Approximately 48 h prior to the last subculture before initializing the experiment, cells were allowed to grow in a basal medium similar to that of Peterson et al. (2004), with modifications. The basal medium was composed of Minimum Essential Medium/Earle’s Balanced Salts HyQ (MEM/EBSS, HyClone) with insulin (5 mg/L, I6634, Sigma, St. Louis, MO), hydrocortisone (1 mg/L, H0888, Sigma), transferrin (5 mg/L, T1428, Sigma), ascorbic acid (5 µM/L, A4544, Sigma), sodium acetate (5 mM/L, S5636, Sigma), and penicillin/streptomycin (10 mL/L, sv30010, HyClone). The basal medium was supplemented with fetal bovine serum (10%, Gibco, Invitrogen, Carlsbad, CA) and growth-promoting hormones (1 mg/L of progesterone, P8783, Sigma; 0.05% lactalbumin, L5385, Sigma; 0.05% 
-lactose, 47287-U, Sigma). At approximately 24 h before applying treatments (approximately 90% confluence), cells were cultured in a lactogenic medium as reported by Peterson et al. (2004), with modifications. The lactogenic medium was prepared as the basal medium, except that high-glucose Dulbecco’s modified Eagle’s medium (HG-DMEM, HyClone) was used. The lactogenic medium was devoid of fetal bovine serum and was supplemented with BSA (1 g/L) and prolactin (2.5 mg/L).
Treatments were the known PPARG ligands (Walkey and Spiegelman, 2008) rosiglitazone (ROSI; Cayman Chemical, Ann Harbor, MI), stearate (18:0; N-18-A, Nu-Chek Prep Inc., Elysian, MN), and oleate (cis-9 18:1, 1022, Matreya, Pleasant Gap, PA) as well as eicosapentanoate (20:5, N-20-A, Nu-Chek Prep Inc.), palmitate (16:0; N-16-A, Nu-Chek Prep Inc.), trans-10 18:1 (kindly provided by M. Pete Yurawecz, US Food and Drug Administration, Washington, DC), and t10c12 CLA (no. 1249, Matreya).
Preparation of LCFA and ROSI
The stock solution of 30 mM LCFA was prepared in 13 x 100 mm Pyrex glass tubes with screw caps by saponification at 42°C with an equimolar solution of NaOH in water. After saponification, the soap was suspended with absolute ethanol to obtain a final solution of 95% ethanol. After preparation, the LCFA stock solution was screw-capped using paraffin and stored at –20°C until use. A 95% ethanol + 30 mM NaOH solution served as the control. Most of these LCFA are PPARG ligands in rodents as well as in human adipocyte cultures (Sauma et al., 2006), and exogenous t10c12 CLA and 20:5 have been shown to elicit different responses in bovine mammary lipid synthesis (Loor et al., 2005b; Harvatine and Bauman, 2006).
Preliminary Time-Course Study.
The MacT cells from several 75-cm2 vented flasks were pooled after trypsinization in a 50-mL tube and mixed thoroughly before transfer to 6-well plates for addition of treatments. As reported above, cultures were maintained for 24 h in a lactogenic medium before treatments were applied (approximately 70% confluence). For RNA extraction, cells were cultured in triplicate with 100 µM palmitate (16:0), trans-10,cis-12 18:2 (t10,c12 CLA), and 50 µM ROSI plus the control for 0, 3, 6, 12, and 24 h in a lactogenic medium. Cells were harvested in 1 mL of TRIzol reagent (Invitrogen) and stored at –80°C until RNA extraction.
The preliminary study was conducted to determine a suitable incubation time based on peak mRNA expression of the lipogenic genes acetyl-coenzyme A carboxylase (ACACA), fatty acid synthetase (FASN), stearoyl-CoA desaturase (SCD), SREBF1, lipin 1 (LPIN1), and the fatty acid translocase CD36 molecule [thrombospondin receptor] (CD36; Figure 1). The PCR data were calculated using a 6-point standard curve and normalized by the geometric mean of MRPL39, UXT, and RPS9. Final data were obtained as the percentage change relative to time 0 for all treatments and were corrected by the control data. Statistical analysis of the qPCR and Oil Red O staining results was performed using the MIXED procedure (version 9.1, SAS Institute, Cary, NC) with repeated measures to evaluate the effects of treatment, time, and treatment x time on normalized mRNA abundance of genes. The model included the fixed effect of time (0, 3, 6, 12, and 24 h) and the random effect of culture plates. For intracytoplasmic lipid droplets, measurement cells were subjected to the same conditions as reported for RNA extraction in duplicate for 24 h, and lipid droplets were visualized with Oil Red O staining (see below). The model included 0 h as a covariate with treatments as fixed effects and culture plates as the random effect. Results indicated that a 12-h incubation was suitable to obtain maximal responses in mRNA expression. Clearly, this represented a compromise because it is unlikely that all genes would have maximal expression at the same point in time.
|
90% confluence, and the treatments were applied. Cells were cultured with 100 µM each LCFA or 50 µM ROSI (Walkey and Spiegelman, 2008). Each treatment was performed in triplicate (i.e., 3 separate flasks per treatment) and cells were harvested at 12 h of incubation.
RNA Extraction, RNA Quality Assessment, and Real-Time PCR
Total RNA was extracted from the cells using ice-cold TRIzol reagent (Loor et al., 2005a). Concentrations of RNA were quantified with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Genomic DNA was removed with DNase using RNeasy Mini Kit columns (Qiagen, Valencia, CA). A portion of the RNA was diluted to 100 mg/L using DNase/RNase–free water before quantitative real-time PCR (qPCR). Sufficient cDNA was prepared to run all selected genes. Each cDNA was synthesized by qPCR using 100 ng of RNA, 1 µL of dT18 (Operon Biotechnologies, Huntsville, AL), 1 µL of 10 mM deoxynucleotide 5'-triphosphate mix (Invitrogen), 1 µL of Random Primers (Invitrogen), and 7 µL of DNase/RNase–free water. The mixture was incubated at 65°C for 5 min in an Eppendorf Mastercycler Gradient and kept on ice for 3 min. A total of 9 µL of Master Mix, composed of 4.5 µL of 5x First-Strand Buffer, 1 µL of 0.1 M dithiothreitol, 0.25 µL (100 U) of SuperScript III RT (Invitrogen), 0.25 µL of RNase Inhibitor (Promega, Madison, WI), and 3 µL of DNase/RNase–free water was added. The reaction was performed using the following temperature program: 25°C for 5 min, 50°C for 60 min, and 70°C for 15 min. Complementary DNA was then diluted 1:3 with DNase/RNase–free water.
For qPCR, 4 µL of diluted cDNA was combined with 6 µL of a mixture composed of 5 µL of 1x SYBR Green Master Mix (Applied Biosystems, Foster City, CA), 0.4 µL each of 10 µM forward and reverse primers, and 0.2 µL of DNase/RNase–free water in a MicroAmp Optical 384-well reaction plate (Applied Biosystems). Each sample was run in triplicate and a 6-point relative standard curve (4-fold dilution) plus the nontemplate control were used. The reactions were performed in an ABI Prism 7900 HT SDS instrument (Applied Biosystems) under the following conditions: 2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C, and 1 min at 60°C. The presence of a single PCR product was verified by the dissociation protocol using incremental temperatures to 95°C for 15 s plus 65°C for 15 s. Data were analyzed with the 7900 HT Sequence Detection Systems Software (version 2.2.3, Applied Biosystems).
Primer Design and Testing
Primer Express 3.0 software (Applied Biosystems), optimized for use with Applied Biosystems qPCR systems, was used for primer design, using the default features except for the amplicon length, which was fixed at a minimum of 100 bp. Primers were designed to span exon-exon boundaries when possible to avoid amplification of genomic DNA. The exon junctions were uncovered by blasting the sequence against bovine genome (Genome Browser Gateway, 2008). Primers were aligned against publicly available sequences in NCBI (National Center of Biotechnology Information, 2008) and UCSC Genome Browser (Genome Browser Gateway, 2008). Before qPCR, primers were tested using the same protocol as for qPCR but without the dissociation step in a 20-µL reaction. Five microliters of the PCR product was run in a 2% agarose gel stained with ethidium bromide to assess the presence of the product to a expected size and the presence of primer-dimer; the rest was purified using a Qiaquick PCR purification kit (Qiagen) and sent to be sequenced at the Core DNA Sequencing Facility of the Roy J. Carver Biotechnology Center at the University of Illinois, Urbana. Only primers with high specificity evaluated by a single band on agarose gel, the absence of primer-dimer, amplification of the right cDNA as verified by sequencing, and a unique peak in the dissociation curve after the qPCR reaction were used. Sequencing results for all genes are reported in a previous publication (Bionaz and Loor, 2008b).
Additional details on selected genes and real-time PCR performance for each gene analyzed are reported in Supplemental Tables 1 and 2 (http://jds.fass.org/content/vol92/issue9).
Oil Red O Staining
The intracytoplasmic lipid droplets in cells were visualized with Oil Red O staining, which was applied to cells fixed with 10% formalin after a wash with PBS (Carson, 1990). Fixation was performed for 30 to 60 min. Fixed cells were washed gently with water, and 2 mL of 60% isopropanol was added to cover the bottom of each well and kept for 5 min. Isopropanol was removed and 1 mL of a filtered working solution of Oil Red O stain (300 mg/100 mL of 99% isopropanol) was added. Cells were incubated for 5 min, Oil Red O staining was removed, and cells were rinsed with tap water. Then, 1 mL of hematoxylin (HHS-16, Sigma) was added for 1 min to stain the nucleus of the cells, and cells were rinsed carefully with tap water. Pictures were taken using an inverted microscope at 400x magnification (Supplemental Figures 1 and 2; http://jds.fass.org/content/vol92/issue9/). Quantification of intracytoplasmic triglycerides was determined by extracting Oil Red O staining as described previously (Ramirez-Zacarias et al., 1992). The quantification of the extract was performed using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies) at 510 nm. A 6-point standard curve with 2-fold dilution of Oil Red O solution was used to obtain the final values (Figure 2).
Data Transformation, Relative mRNA Abundance, and Statistical Analysis
Raw qPCR data were transformed based on a 6-point standard curve and normalized by the geometric mean of 3 previously tested internal control genes (MRPL39, RPS9, and UXT; Bionaz and Loor, 2007). Normalized data were further transformed to obtain a perfect mean of 0.0% for the control, and differences in mRNA expression of treatments were transformed to percentage of mRNA relative to the control mean (Bionaz and Loor, 2008a,b). The median of the relative percentages of mRNA abundance among genes was calculated as reported previously (Bionaz and Loor, 2008b); that is, 1/E(
Ct), where E is the efficiency of PCR amplification and Ct is calculated as [Ct gene – geometrical mean of Ct internal controls] considering all treatments (see Supplemental Table 2 for PCR performance and Supplemental Figure 3 for relative percentage mRNA abundance among measured genes; http://jds.fass.org/content/vol92/issue9/). The GLM procedure (version 9.1, SAS Institute Inc., Cary, NC) was used to evaluate the treatment effects on percentage of mRNA abundance relative to the control and lipid droplet formation. Fixed effects in the model were treatments, whereas the random effects were replicates (n = 3 cultures/treatment). Treatment means were separated using Fisher’s least significant difference pair-wise comparisons (Peterson et al., 2004). Significance was declared at P 
0.05.
Expression Clustering and Gene Network Development
Hierarchical clustering (Figure 3) was performed using fold changes in mRNA expression for each treatment relative to the control by using Genesis software (Sturn et al., 2002). Summary networks among the genes most affected by treatments were developed using Ingenuity Pathway Analysis (IPA; Ingenuity Systems, Redwood City, CA; www.ingenuity.com; Figure 4; Supplemental Figures 4 to 6; http://jds.fass.org/content/vol92/issue9). The gene expression data set comprising gene identifiers, corresponding expression values, and pair-wise comparison P-values were uploaded into IPA. Connections among genes were developed based on results from this study as well as on known relationships available in the IPA knowledge base. This is a proprietary manually curated database containing relationships from the published literature on rodents and humans.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
|
|
Among genes associated with de novo fatty acid synthesis, expression of ACACA and FASN was lower with 18:0 (–56 and –69%) and t10c12 CLA (–97 and –31%) relative to the control (Table 1). Trans-10 18:1 decreased expression of FASN by 61%. The mRNA expression of SCD was upregulated with 16:0 (95%) but was downregulated by the presence of LCFA containing 1 or more double bonds, namely, cis-9 18:1 (–428%), trans-10 18:1(–100%), t10c12 CLA (–357%), and 20:5 (–205%; Table 1). Surprisingly, 18:0 did not affect SCD mRNA expression.
Although none of the treatments affected expression of glycerol-3-phosphate acyltransferase, mitochondrial (GPAM), expression of diacylglycerol O-acyltransferase homolog 1 (mouse; DGAT1) was greater with 18:0 (89%). Expression of 1-acylglycerol-3-phosphate O-acyltransferase 6 (AGPAT6) was greater in response to all LCFA treatments except trans-10 18:1. Expression of lipin 1 (LPIN1) was greater with 16:0 (141%) but was lower with cis-9 18:1 (–63%).
The PPARG agonist ROSI did not affect expression of genes associated with LCFA uptake or intracellular activation and transport, including LPL (Table 1). However, it resulted in greater expression of ACACA (85%), AGPAT6 (269%), DGAT1 (76%), FASN (39%), and LPIN1 (116%), suggesting that these genes are putative PPARG target genes in MacT cells and likely in bovine mammary tissue.
mRNA Expression of Transcriptional Regulators
Relative to the control, SREBF1 expression was lower with cis-9 18:1 (–140%), trans-10 18:1 (–49%), t10c12 CLA (–189%), and 20:5 (–125%; Table 1). Similarly, sterol regulatory element-binding transcription factor 2 (SREBF2) expression was lower with cis-9 18:1 (–83%), t10c12 CLA (–71%), and 20:5 (–35%) but was greater with 16:0 (53%). Among the genes known to regulate SREBP activity, mRNA expression of SREBF chaperone (SCAP) was lower (–76%) with 20:5. Expression of insulin-induced gene 1 (INSIG1) was greater with 16:0 (298%), 18:0 (92%), and 20:5 (84%) but was lower with cis-9 18:1 (–93%) and t10c12 CLA (–95%). Expression of the transcription factor THRSP was greater with 18:0 (97%) and lower with t10c12 CLA (–73%) relative to the control. None of the treatments affected PPARG expression. Rosiglitazone led to greater expression of SREBF1 (100%), SREBF2 (35%), and INSIG1 (142%), also suggesting that these genes are putative PPARG target genes in bovine.
Gene Expression Clusters
The hierarchical algorithm implemented by Genesis attempts to find successive clusters using previously established clusters; that is, it uses a "top-down" (or divisive) approach, beginning with the whole set, and proceeds to divide it into successively smaller clusters (Sturn et al., 2002). This clustering analysis revealed that the mRNA expression profiles obtained essentially grouped into 3 patterns of expression (Figure 3). Treatments 16:0, 18:0, and ROSI were grouped in a cluster characterized primarily by upregulation (Table 1) of several genes, including FABP4, ACSS2, FABP3, LPIN1, INSIG1, THRSP, and DGAT1. Within this group, the most similar expression pattern across the 19 genes examined was that between 16:0 and 18:0. A second cluster was composed of expression patterns obtained with treatments trans-10 18:1, t10c12 CLA, and cis-9 18:1. Within this cluster, expressions patterns with t10c12 CLA and cis-9 18:1 were further grouped together based on their more similar expression relative to trans-10 18:1. The third major cluster of expression was obtained with treatment 20:5.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
is one of the most studied nuclear receptors because of its central role in regulating metabolism in relation to the composition and concentration of LCFA in the cellular environment (Desvergne et al., 2006). The fact that LCFA are potent agonists of PPARG in nonruminants led us to infer that it could be a potential player in mammary lipid metabolism. Among the genes studied, CD36, LPL, FABP4, and ACSL1 are known PPARG target genes in nonruminants (Berger and Moller, 2002). In addition, genes related to de novo fatty acid synthesis (ACACA, FASN), desaturation (SCD), and TAG synthesis (AGPAT6) are upregulated in adipose tissue of rats (Way et al., 2001) or humans (Kolak et al., 2007) by PPARG agonist treatment (GW1929, ROSI). As summarized in Figure 4, treatment with ROSI increased the expression of genes related to de novo fatty acid synthesis (ACACA, FASN), TAG synthesis (AGPAT6, DGAT1, LPIN1), transcriptional regulation of lipid synthesis (SREBF1, SREBF2), and the SREBF1-repressor INSIG1. These observations strongly indicate a role for PPARG in de novo fatty acid synthesis regulation, partly because the estimated control flux coefficient through the pathway was estimated to be 63% for ACACA and for 37% for FASN in bovine mammary homogenates (Wright et al., 2006). Together, the results support a previous suggestion (Bionaz and Loor, 2008b) that PPARG activation could be a regulatory point of milk fat synthesis.
Our observations with ROSI are striking because they suggest that a single nuclear receptor with LCFA as its natural ligands could potentially regulate not only target genes (ACACA, FASN, LPIN1), as in nonruminants, but also a subset of lipogenic transcription factors (i.e., SREBF1 and SREBF2). In addition to potentially binding up to 2 fatty acids simultaneously, it has been proposed (Itoh et al., 2008) that the large pocket in PPARG could allow for a graded response to varying compositions of the cellular pool of LCFA ligands (e.g., saturated and unsaturated LCFA). A complex regulatory network might be at work in mammary cells such that several proteins serve as putative checkpoints to regulate milk fat synthesis. As an example, increased SREBF1 expression upon PPARG activation could lead to upregulation of downstream lipogenic genes such as ACACA, FASN, and AGPAT6 (Figure 4). Under a different scenario, if the observed increase in INSIG1 mRNA expression attributable to ROSI extended to the protein level, more SREBP1 might be retained in the endoplasmic reticulum via SCAP-INSIG1 binding; that is, SREBP1 would be rendered inactive (Horton et al., 2003; Kast-Woelbern et al., 2004), as suggested previously (Bionaz and Loor, 2008b).
Rosiglitazone did not affect the expression of LPL and ACSL1, both known PPARG target genes in nonruminant adipose tissue, suggesting that regulation of lipogenic gene expression through PPARG might differ between adipose tissue and mammary epithelial cells. Alternatively, there might be species differences in the genes regulated by PPARG as well as in the potency of their regulation. The LPL was one of the least abundant transcripts in MacT cells (Supplemental Figure 3). This might be indicative of an adaptation of immortalized mammary cells, which do not require LPL because, in vivo, this protein is exported to the endothelium for processing of lipoproteins. We did not observe an increase in lipid droplet formation with ROSI relative to the control despite the greater mRNA expression of lipogenic genes and ample availability of acetate in culture medium, which, coupled with greater ACACA and FASN mRNA expression, should have provided short-chain fatty acids and 16:0 for lipid droplet formation. We speculate, based on those data, that mammary cells have a requirement for LCFA for lipid droplet formation.
LCFA Regulation of Mammary Lipogenic Genes Through PPARG
Palmitic Acid.
As in other tissues, mammary LCFA uptake and metabolism are likely orchestrated by a large set of proteins (Bionaz and Loor, 2008b). Results with ROSI suggest that several of the examined lipogenic genes are PPARG targets; thus, their level of expression in response to LCFA provides an indication of whether they might act through this nuclear receptor. Clustering analysis (Figure 3) revealed that the similarity of expression patterns among tested genes relative to ROSI was 16:0 > 18:0 > 20:5 > t10c12 CLA = cis-9 18:1 > trans-10 18:1. Together, these data suggest that the effect of 16:0 and 18:0 on the transcription of lipogenic genes in MacT cells is partly driven by the activation of PPARG (Figure 4).
Genes upregulated by 16:0 and ROSI included AGPAT6, LPIN1, INSIG1, and SREBF2 (Table 1 and Figure 4). Furthermore, 16:0 was the only LCFA that concomitantly increased ACSS2 expression [a gene encoding a cytosolic protein with high specificity for acetate and, to some extent, propionate (Fujino et al., 2001)] as well as intracytoplasmic TAG content (Figure 2 and Supplemental Figures 1 and 2). A role for 16:0 in regulating de novo synthesis was observed previously in bovine and ovine dispersed mammary epithelial cells when exogenous 16:0 led to increased synthesis and incorporation of butyrate into TAG (Hansen and Knudsen, 1987). In addition, 16:0 is the preferred substrate for the initial acylation of L--glycerolphosphate via GPAM, leading to formation of sn-1-lysophosphatidic acid, which is believed to represent the rate-limiting step in TAG synthesis (Kinsella and Gross, 1973). Together, the data are suggestive of induction of de novo fatty acid synthesis and esterification by 16:0 through greater provision of substrate for the key enzymes ACACA and FASN, but also through greater acylation capacity. The latter effect might have involved binding and activation of PPARG.
In addition to the putative action of 16:0 through PPARG, increased TAG synthesis in mammary cells also could encompass upregulation in expression of the mammary-enriched LCFA-binding proteins FABP3 and FABP4. In nonruminants, these FABP play a pivotal role in the trafficking and provision of LCFA for enzymes of TAG synthesis as well as the LCFA activation of PPARG (Mashek and Coleman, 2006). Such an effect in bovine mammary tissue was proposed previously for FABP3 (Bionaz and Loor, 2008b). Upregulation of LPIN1, a coactivator of PPARG in mouse adipose (Reue and Zhang, 2008), provides additional evidence that 16:0 might increase activation of PPARG. In this regard, the fact that PPARG apparently is able to bind 2 substrates simultaneously (Itoh et al., 2008) is suggestive of a need for an LCFA for full activation of the transcriptional program via ROSI. This might explain the lower than expected response to ROSI.
Because 16:0 is the main product of de novo synthesis in mammary cells, we propose that this LCFA could serve as a feed-forward mechanism for copious milk fat synthesis during the normal course of lactation. The large accumulation of lipid droplets (Figure 2 and Supplemental Figures 1 and 2; http://jds.fass.org/content/vol92/issue9/) supports this hypothesis. Overall, data indicate that 16:0 elicits a positive effect on milk TAG synthesis, partly through PPARG and potentially through as yet unidentified transcription factors (Figure 4). Duodenal infusion of palmitic acid resulted in greater milk fat synthesis and milk 4:0 concentration and yield (Enjalbert et al., 2000).
Stearic Acid.
Although there was some overlap in responses between 18:0 and ROSI, current and previous data are suggestive of 18:0 as a less potent or "selective" ligand of PPARG. Stearate increased INSIG1 and AGPAT6 as well as DGAT1. Contrary to 16:0, however, 18:0 decreased de novo synthesis but increased mammary TAG synthesis from preformed fatty acids in mammary epithelial cell cultures (Hansen and Knudsen, 1987). We observed lower ACACA and FASN expression with 18:0 and other LCFA considered to be strictly of exogenous origin (e.g., trans-10 18:1, t10c12 CLA, 20:5; Table 1 and Supplemental Figure 4). Biologically, a positive effect of 18:0 on mammary cell TAG synthesis (Hansen and Knudsen, 1987; Enjalbert et al., 2000) can be explained by the fact that this LCFA is primarily taken up from blood and not synthesized endogenously; that is, mammary cells reduce de novo synthesis in favor of esterification. In fact, greater AGPAT6 mRNA with 18:0, but also with unsaturated LCFA (i.e., cis-9 18:1, t10c12 CLA, and 20:5; Table 1), which are of exogenous origin, could be explained by the need to load additional LCFA into a growing intracellular pool of lysophosphatidic acid.
A previous study with MacT cells showed that exogenous 18:0 (25, 50, and 100 µM) could have a positive effect on ACACA and FASN activity; that is, despite lower mRNA abundance, the efficiency of these reactions might somehow be enhanced by 18:0 (Jayan and Herbein, 2000). It also is possible that endogenous synthesis of cis-9 18:1 from absorbed 18:0 through SCD plays a role in driving greater TAG synthesis, as speculated previously (Loor and Herbein, 2003; Bionaz and Loor, 2008b). This idea is partly supported by the significant formation of lipid droplets in cells cultured with 18:0 relative to the control (Figure 2 and Supplemental Figures 1 and 2), and also by the lack of change in SCD mRNA expression (Table 1) or SCD activity with incremental 18:0 (Jayan and Herbein, 2000). Exogenous 18:0 led to greater intracellular cis-9 18:1 in MacT cells (Jayan, 1998). The SCD was the most highly expressed transcript among those measured in MacT cells (Supplemental Figure 3). Overall, data suggest that 18:0 could manipulate de novo milk fatty acid synthesis through effects on gene expression.
Overall LCFA Regulation of Mammary Lipogenic Genes
Trans-10 Biohydrogenation Intermediates.
Effects of t10c12 CLA in our experiment were similar to those in previous reports (Harvatine and Bauman, 2006; Bauman et al., 2008). However, genes such as GPAM, LPL, AGPAT6, and ACSL1 were unaffected by this CLA. This can be explained by several factors, including cell culture conditions, measurement of different isoforms across studies (e.g., AGPAT6 and ACSL1), use of a different approach for normalization of qPCR data, or their combination. Novel information in this regard is the negative effect of t10c12 CLA on expression of additional lipogenic genes such as FABP3 and ACSS2. Overall, our data confirmed that t10c12 CLA has potent negative effects on lipogenic gene networks in mammary cells (Table 1 and Supplemental Figure 6). However, in our study, t10c12 CLA induced greater intracellular TAG accumulation (Figure 2 and Supplemental Figures 1 and 2). We can infer a mechanism in which there was a high rate of uptake of CLA, coupled with its preferential distribution (and its metabolites) into the neutral lipid fraction (Banni et al., 2001). The former is supported by the large upregulation of CD36 (which was common for all LCFA), as observed previously in similar in vitro studies (Yonezawa et al., 2008). Data on cell counts (data not shown) suggest that there was no apparent increase in apoptosis with t10c12 CLA, in contrast with a previous report at a lower level of supplementation (Keating et al., 2008). A biological explanation for the large accumulation of lipid droplets with t10c12 CLA despite downregulation of classical lipogenic genes is not apparent.
Although with different potency relative to t10c12 CLA, trans-10 18:1 decreased expression of FASN and SCD, both of which are crucial for de novo fatty acid synthesis and generation of cis-9 18:1 for TAG formation (Loor and Herbein, 2003). Recently, a lack of effect of short-term (4-d) intestinally infused (approximately 43 g/d) trans-10 18:1 on milk fat synthesis in lactating cows was reported (Lock et al., 2007). However, milk trans-10 18:1 reached only 1.1% of total fatty acids during infusions, well below the concentration calculated for maximal milk fat depression of 2.5% (Loor et al., 2005c; Kadegowda et al., 2008). Thus, higher amounts of trans-10 18:1 reaching the mammary gland could potentially induce milk fat depression by decreasing lipogenic gene expression. Trans-10 18:1 did not elicit effects similar to ROSI; thus, it does not appear to be a ligand of PPARG.
Oleic Acid.
Oleic acid resulted in similar effects on expression of lipogenic genes as t10c12 CLA (Table 1 and Supplemental Figure 4). This LCFA was the most potent inhibitor of SCD expression, which is suggestive of a feedback-inhibition mechanism. A previous study with MacT cells reported lower ACACA, FASN, and SCD activity with incremental exogenous cis-9 18:1 (25, 50, and 100 µM; Jayan and Herbein, 2000). Exogenous oleic acid also inhibited acetate incorporation into TAG by dispersed ruminant mammary cells (Hansen and Knudsen, 1987). Our in vitro data suggest that cis-9 18:1 can potentially have a negative effect on lipoprotein utilization by mammary tissue (i.e., LPL), trafficking (i.e., FABP3), and desaturation (i.e., SCD). Compared with 16:0 and 18:0, oleic acid affected expression of only AGPAT6 in a similar fashion as ROSI; thus, we can infer that this LCFA is a weak effector of PPARG.
Eicosapentaenoic Acid.
Eicosapentaenoic acid reduced expression of genes involved in the nonruminant SREBF1 network (Horton et al., 2003; Table 1 and Supplemental Figure 5). In addition to inducing lower mRNA of SREBF1 and SCAP, expression of INSIG1 increased. This latter effect might further decrease the activity of SREBF1 (Kast-Woelbern et al., 2004). The decrease in SCD mRNA by 20:5 was lower relative to t10c12 CLA. In this regard, it is noteworthy that activation of PPAR
(via GW0742) increased INSIG1 (mRNA and protein), inhibited proteolytic-processing of mature SREBF1, and reduced FASN, ACACA, and SCD, all of which led to reduced intracellular lipid accumulation in murine liver (Qin et al., 2008). Bovine mammary tissue expresses PPAR (M. Bionaz, S. L. Rodriguez-Zas, R. E. Everts, H. A. Lewin, and J. J. Loor, University of Illinois, Urbana; unpublished results). Together, these data suggest that 20:5 could regulate mammary lipid synthesis through both PPARG and PPAR. Duodenal infusion of fish oil with high 20:5 content resulted in lower milk fat concentration as well as lower yield of 16:0 and cis-9 18:1 in milk fat (Loor et al., 2005b), all of which agree with the observed inhibitory effect of 20:5 on SREBF1, SREBF2, and SCD. Among unsaturated LCFA, 20:5 increased expression of AGPAT6, a PPARG target gene.
Relevance of Gene Expression Data to In Vivo Studies.
Negative effects on milk fat synthesis in vivo have been convincingly demonstrated for t10c12 CLA (Baumgard et al., 2000; Loor and Herbein, 2003) and mixtures of trans-18:1 LCFA (Gaynor et al., 1994). Despite the lack of "strong" in vivo data (see Chilliard et al., 2001), the fact that all unsaturated LCFA decreased (significantly or numerically) mRNA expression of SREBF1, SREBF2, THRSP, SCD, ACACA, and LPL indicates that they have some ability to reduce in vivo mammary synthesis of fatty acids, TAG, or both (Enjalbert et al., 2000; Loor et al., 2005b; Kadegowda et al., 2008). One likely mechanism stems from an "indirect" inhibition of SREBF1 and SREBF2, that is, downregulation through an upstream transcription factor or nuclear receptor able to bind unsaturated LCFA (Qin et al., 2008). This suggestion is supported by significant downregulation (and tendencies) of the SREBF target genes ACACA, FABP3, LPIN1, and SCD (Table 1). Inhibition of SCD, as assessed by lower cis-9 18:1/18:0 milk fatty acid ratios or SCD mRNA (Baumgard et al., 2002), is a classical marker of milk fat depression induced by exogenous t10c12 CLA (Baumgard et al., 2000; Loor and Herbein, 2003).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication November 26, 2008. Accepted for publication March 22, 2009.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Related articles in JDS:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
