Modulatory effect of high saturated fat diet-induced metabolic disturbances on angiogenic response in hepatocyte RXRα knockout mice

Abstract

Metabolic syndrome and diabetes lead to pathological angiogenesis and angiopathy. Metabolic disturbances occur as an effect of genetic and environmental interaction. Hyperleptinemia accompanies obesity and leptin is a potent proangiogenic factor. The aim of the study was to investigate the effect of high fat diet-induced alterations in gene expression and angiogenic response in the hRXRα ko mice lacking of hyperglycemia.

hRXRα ko and control mice were fed either standard or high saturated fat (HF) diet for 7 weeks. Body weight and biochemical parameters (glucose, triglycerides, cholesterol), insulin and adipokines (leptin, adiponectin) were monitored. At sixth week of feeding, mice were subcutaneously injected for 6 days with matrigel containing bFGF. Then, matrigel plugs were used for immunohistochemical staining of cells with CD31 antibody and gene expression assessment (by microarray confirmed for some genes with quantitative real time PCR). For description of angiogenesis CD31 positive structures were counted in the matrigel sections. HF diet feeding of the hRXRα ko mice resulted in increased serum cholesterol and leptin level and in tendency to decrease angiogenesis (number of vessels with lumen). The microarray studies revealed that HF diet down-regulated genes related to angiogenesis (Nos3, Kdr) and up-regulated genes connected with apoptosis (activators of caspase 3, proapoptotic genes Bcl2) and proinflammatory pathway (NfκB pathway, Tnfα).

Summing up, HF diet feeding of hRXRα ko mice resulted in dyslipidemia and hyperleptinemia as well as impaired angiogenic response, and cell apoptosis. These results argue for independent participation of dyslipidemia and hyperleptinemia in pathology of angiogenic response associating metabolic syndrome.

Introduction

Obesity, metabolic syndrome and diabetes type 2 represent one of the dominating disorders of industrialized societies [17]. Metabolic disturbances occur as an effect of genetic background combined with the influence of environmental factors such as consumption of a high fat western diet. This results in a progressive development of obesity and insulin resistance (hyperglycemia, hyperinsulinemia), which are the main symptoms of metabolic syndrome (MS). One of the MS consequences is pathological angiogenesis that controls angiopathy such as retinopathy, nephropathy etc. [10, 22, 25].

Pathological angiogenesis is characterized by growth of new, leaky, immature blood vessels, which resemble primitive vascular network found in cancer [20] or atherosclerotic plaque [16]. Important role in the induction of this pathological angiogenesis play biochemical factors such as: dyslipidemia, modified lipoproteins (ox LDL), increased level of free fatty acids (FFA), insulin and insulin-like growth factor 1 (IGF-1), adipokines (leptin, adiponectin). Angiogenesis is also regulated by: pro-inflammatory cytokines (interleukins: IL-6, IL-8, tumor necrosis factor α (TNFα)), growth factors (vascular endothelial growth factor (VEGF), angiopoietin-1 and −2, stromal-derived factor-1 (SDF1), fibroblast growth factor-2 (bFGF) as well as pigmented epithelium derived growth factor (PEDGF)) or nitric oxide (NO) which concentrations are changed [7, 20, 23].

Retinoid X receptor α (RXR α) is a transcription factor, which regulates gene expression of enzymes participating in lipid (FFA biosynthesis as well as oxidation) and bile acid metabolism in the liver [29]. Thus, the hepatocyte RXRα deficient mice are a unique model for studying selected symptoms of metabolic syndrome, characterized by elevation of FFA in the blood [31–33]. Mice with RXRα deficiency in hepatocytes (hRXRα ko), including young ones, demonstrate elevated serum triglycerides concentration and increased apoCIII mRNA amounts [31]. Glucose and insulin concentrations do not differ between knockout and control wild type mice, when hRXRa ko mice develop hypercholesterolemia and hyperleptinemia upon feeding them high fat (HF) diet. There is some evidence that increased glucose tolerance in hRXRα ko mice fed with HF diet is associated with the increased expression of IGF1 [33].

The aim of the study was to investigate how the selected diet-induced alterations in hRXRα ko mice metabolism such as dyslipidemia and hyperleptinemia with normal blood glucose and insulin concentrations may impact angiogenic response.

Materials and Methods

Animals and experimental conditions

The study protocol was reviewed and approved by the Local University Ethic Committee in Kraków (No. 58/OP/2003). All experiments were performed according to Polish laws and approved by the Polish Animal Inspectorate and Institutional Animal Care.

Control wild type mice (WT) and hepatocyte RXR α deficient mice (hRXRα ko) were kindly provided by Dr. Yu-Jui Yvonne Wan (University of Kansas Medical Center, Kansas City, Kansas), who generated hepatocyte specific RXRα deficient mice [32].

Male mice, which were 15 week old and weighted 29–35 g, were housed in cages at 22°C with 12-h light, 12-h dark cycle and had free access to food and water. Wild type and knockout animals were fed either standard lab chow containing about 9 energy percent (9 en%) of fat (3% weight of the diet) or high saturated fat diet (HF, coconut oil hydrogenated based, 39 en% of fat) (20% weight of the diet) (MP Biomedical Research, USA) for 7 weeks. The animals were weighted twice a week.

Blood samples were collected from the tip of tails once a week (after 4 h of starvation). Serum biochemical parameters (glucose, cholesterol, triglycerides) were measured using Cormay Diagnostic Kits (Poland). Serum leptin level was determined using Anti Mouse ELISA Kit (R&D). Insulin was measured with Elisa Kit (Linco Research).

The in vivo model of angiogenesis

The angiogenesis model was introduced in the WT as well as hRXRα ko animals following six weeks of standard or HF diet. Mice were injected subcutaneously in the dorsal region with two sterile injections of 500 µl matrigel (Becton Dickinson) containing bFGF [25 nM]. Six days later, matrigel plugs were removed under anesthesia and they were used for endothelial cell immunohistochemistry (immunostaining of CD31 PECAM1-positive cells) and analysis of gene expression in the infiltrating matrigel plug cells [30, 35].

Immunohistochemical studies were carried out in the matrigel plugs fixed in Zinc-Fixative (Becton Dickinson) and immersed in paraffin. The endothelial cells infiltrating matrigel were visualized with antibodies specific for CD31 antigen in paraffin embedded matrigel sections. The primary rat anti-mouse CD31 antibodies (anti-PECAM1, Becton Dickinson) at 1:300 dilution were used. The slides were rehydrated and incubated in 3% peroxide solution for 10 min to block endogenous peroxidase activity. The Streptavidin-Biotin (BD Pharmingen) detection system with visualization by Anti-Rat HRP Detection Kit (Becton Dickinson) was used. The matrigel sections were contra-stained with Meyer hematoxylin (DAKO, Denmark). Blood capillaries that had formed in the matrigel plugs were counted under microscope by an uninformed pathologist, who used a “hotspot” method to visually inspect five different fields in three slides taken from different parts of each plug. The previously described method of microvessel evaluation in the vasculature “hotspots” was applied [34].

Analysis of gene expression in the matrigel plug cells

Microarray

RNA from the cells that had migrated into the matrigel plugs was isolated using TRIZOL Reagent (Invitrogen Life Technologies) and purified with QIAamp RNA Blood Mini Kit for total RNA isolation (Qiagen). High quality of the isolated RNA was confirmed by its analysis on the Agilent 2100 Bioanalyzer (Agilent Technologies).

The effects of HF diet on the gene expression in the cells found in the matrigel plugs removed from the knockout and control mice were screened by the microarray assays. RNA was transcribed into cDNA using Superscript II (Invitrogen Life Technologies) with a primer containing T7 promoter. cDNA was used as a template for transcription reaction (Enzo BioArray, Affymetrix). The target cRNA was purified on RNeasy columns (Qiagen), then fragmented for hybridization to Affymetrix 430A_2 GeneChips (containing 22,691 spots for estimation of 14,000 genes from mice genome). Hybridization was carried out at 45°C for 16 h in the GeneChip Hybridization Oven 640 (Affymetrix). The GeneChips were scanned with the Hewlett Packard GeneArray Scanner and results were analyzed with Affymetrix Microarray Analysis Suit. Only spots that demonstrated significant differences in signal intensities were considered (p < 0.05).

Changes in the relative gene expression in the matrigel infiltrating cells derived from the hepatocyte RXR α knockout mice compared to wild type animals (pooled material from 3 matrigel plugs) were calculated with GCOS 1.4 computer software (Affymetrix). The microarray results were presented as relative gene expression values (fold of change). Only genes in which fold of change was greater than 1.4 were used for further analysis.

Confirmation of the selected microarray indicated gene expression by the quantitative realtime PCR (qRT-PCR)

A limited number of genes, which expression was significantly altered during microarray analysis, was further monitored by qRT-PCR with Gapdh as the reference gene. The mostly affected genes, which corresponded to angiogenesis, adipogenesis, energy expenditure, and glucose metabolism, were chosen. Expression of the genes associated with angiogenesis (VEGF receptor 2: Kdr, endothelial nitric oxide synthase: Nos3, platelet endothelial cell adhesion molecule-1: Pecam1), adipogenesis (adipsin: Adn, fatty acid binding protein 4: Fabp4, lipoprotein lipase: Lpl, CCAAT/ enhancer binding protein (C/EBP) β: Cebpb, sterol regulatory element binding factor 1: Srebf1), energy expenditure (uncoupling protein 2 (mitochondrial, proton carrier): Ucp2) and glucose uptake (solute carrier family 2 (facilitated glucose transporter), member 4: Slc2a4) was confirmed by real time PCR.

One microgram of total RNA for cDNA synthesis was reverse transcribed in a total volume of 40 µl reaction buffer containing 5 × First Strand Buffer, DTT, oligo(dT) (Sigma), deoxy-NTPs (Promega) and 200 units of SUPERSCRIPT II reverse transcriptase (Invitrogen Life Technologies) at 42°C for 50 min. The mixture was heated to 70°C for 15 min and immediately chilled on ice. Then, cDNA was subjected to real time PCR in a reaction mixture containing QuantiTect SYBR Green PCR (Qiagen) mix and primers. The primers had intervening intron between the sense and antisense primers to eliminate possibility of amplifying any genomic DNA. The primers were checked for specificity by BLAST searches (Tab. 1). The thermal profile of the PCR reaction included initial denaturation for 15 min at 95°C, followed by 40 amplification cycles of denaturation for 30 s at 94°C, annealing for 30 s at 60°C, and elongation for 30 s at 72°C. Melting curve analysis was performed after PCR amplification with a temperature profile slope of 1°C/s from 35°C to 95°C. The expression rates were calculated as the normalized C_T difference between a control probe and a sample with the adjustment for the amplification efficiency relative to the expression level of the housekeeping gene Gapdh (glyceraldehyde 3-phosphate dehydrogenase). Calculations were performed using a software program Calculation Matrix for PCR Efficiency REST-XL.

Tab. 1.

List of primers used in qRT-PCR

Gene symbol	mRNA sequence	Forward primer	Reverse primer
Gapdh	NM_08084	5’tcaccaccatggagaaggc3’	5’acacccatcacaaacatgg3’
Fabp4	NM_24406	5’ggatggaaagtcgaccacaat3’	5’gtggaagtcacgcctttcata3’
Lpl	NM_08509	5’cctgaagactcgctctcagat3’	5’ggttgtgttgcttgccatt3’
Cebpb	NM_09883	5’ttcctctccgacctcttcg3’	5’ggccgaggctcacgtaac3’
Srebf1	NM_11480	5’gagcttccggcctgctat3’	5’tcagactgcgatccaggag3’
Adn	NM_13459	5’cattaacatgatgtgtgcagaga3’	5’cacgtaaccacaccttcgac3’
Ucp2	NM_11671	5’ggtcactgtgcccttaccat3’	5’ccaagcggagaaaggaag3’
Slc2a4	NM_09204	5’ggcatgggtttccagtatgt3’	5’agatgaagaagccaagcagg3’
Kdr	NM_10612	5’tgcctacctcacctgtttcc3’	5’ctctttcgcttactgttctggag3’
Nos3	NM_08713	5’tgcagagaattctggcaaca3’	5’gtggtagcgttgctgatcc3’
Pecam1	NM_08816	5’gccagtagcatcatggtcaa3’	5’ccaacaactccccttggtc3’

Statistical analysis

All results were shown as the mean value ± standard deviation (SD). Analysis of differences was performed with Student t-test and p < 0.05 was considered as significant.

Results

Effect of HF diet on serum biochemical parameters

Mice hRXRα ko fed for 7 weeks with the high saturated fat diet had significantly higher serum leptin (Fig. 1) and cholesterol in comparison to the WT mice (Fig. 2). HF diet resulted in the increase of insulin and glucose in the wild type as well as mutant animals without any significant differences between genotypes. However, hRXRα ko mice did not develop obesity on HF diet (data not shown).

Fig 1.

Serum leptin of wild type (WT) and hepatocyte RXR α−/−mice (hRXRa ko) fed standard (STD) and high saturated fat diet (HFD). Values are expressed as the mean ± SD (n = 5), statistical significance of the differences: * p < 0.05, hRXRa ko on HFD vs. WT; # p < 0.05, hRXRa ko on HFD vs. STD

Fig 2.

Serum cholesterol of wild type (WT) and hepatocyte RXR α−/− mice (hRXRa ko) fed standard (STD) and high saturated fat diet (HFD). Values are expressed as the mean ± SD (n = 5), statistical significance of the differences: * p < 0.05, hRXRa ko on HFD vs. WT; # p < 0.05, hRXRa ko on HFD vs. STD

Influence of HF diet feeding on angiogenesis

Quantitative analysis of CD31 (PECAM1) positive structures in the matrigel plugs demonstrated that hRXRα ko showed tendency for weaker angiogenic response (formation of vessels with lumen) than wild type mice, especially after the HF feeding (Fig. 3).

Fig 3.

Number of vessels with lumen in matrigel plug of wild type mice and RXRα −/− mice (hRXRa ko) fed standard (STD) and high saturated fat diet (HFD). Values are expressed as the mean ± SD. (n = 5)

Gene expression in matrigel plug cells-microarray analysis

The microarray analysis revealed down-regulation of 963 and up-regulation of 1196 genes in hRXRα ko mice, whereas 1246 genes were down-regulated and expression of 1238 genes was up-regulated in WT mice. The significantly regulated genes were associated with angiogenic response, apoptosis, inflammatory response, oxidative stress, triglyceride synthesis, gluconeogenesis and with transcription factors related to adipogenesis (Tab. 2).

Tab. 2.

Comparison of microarray changes in relative gene expression (between matrigel plug cells from mice fed standard diet (ST) and matrigel plug cells from mice fed high saturated fat diet (HF)) in RXR α deficient mice (hRXRα ko) and WT mice in pathways connected with angiogenesis, apoptosis, adipogenesis, glucose and triglycerides synthesis, eicosanoid synthesis and oxidative stress response.

Gene/Pathway Name	Symbol	WT HF vs. ST	hRXRa ko HF vs. ST
CELL CYCLE
cyclin B1	Ccnb1		−1.87
cell division cycle 2 homolog A (S. pombe)	Cdc2a	1.74	−1.62
origin recognition complex, subunit 4-like (S. cerevisiae )	Orc4l	1.74	−2.00
origin recognition complex, subunit 6-like (S. cerevisiae )	Orc6l	n. c.	−2.00
replication protein A3	Rpa3	n. c.	−2.00
cyclin-dependent kinase 7 (homolog of Xenopus MO15 cdk-activating kinase)	Cdk7	n. c.	−3.48
DNA primase, p49 subunit	Prim1	1.52	−1.41
DNA primase, p58 subunit	Prim2	n. c.	−3.73
cyclin D3	Ccnd3	−1.74	1.32
cyclin D2	Ccnd2	−1.41	n. c.
transformation related protein 53	Trp53	−2.14	n. c.
cyclin-dependent kinase inhibitor 1A (P21)	Cdkn1a	−1.87	n. c.
minichromosome maintenance deficient 3 (S. cerevisiae )	Mcm3	1.62	n. c.
minichromosome maintenance deficient 6 (MIS 5 homolog, S. pombe) (S. cerevisiae)	Mcm6	2.30	n. c.
minichromosome maintenance deficient 7 (S. cerevisiae )	Mcm7	1.87	n. c.
cyclin A2	Ccna2	1.32	n. c.
3-monooxygenase/tryptophan 5-monooxygenase activation protein, γ polypeptide	Ywhag	2.46	−2.64
CELL ADHESION
integrin α M	Itgam	−4.59	2.30
integrin α X	Itgax	−1.62	1.74
integrin β 2	Itgb2	−3.73	1.41
integrin β 7	Itgb7	−1.32	3.73
procollagen, type XI, α1	Col11a1	n. c.	−42.22
procollagen, type III, α1	Col3a1	4.29	−1.41
procollagen, type IV, α1	Col4a1	6.50	−2.14
procollagen, type V, α 2	Col5a2	1.62	−1.52
procollagen, type VIα2	Col6a2	2.83	1.52
matrix metallopeptidase 9	Mmp9	n. c.	3.73
matrix metallopeptidase 12	Mmp12	−2.00	1.87
gap junction membrane channel protein α 1	Gja1	n. c.	−1.41
gap junction membrane channel protein α 4	Gja4	n. c.	−1.74
gap junction membrane channel protein β 5	Gjb5	n. c.	−2.46
kinase insert domain protein receptor	Kdr	n. c.	−1.87#
platelet derived growth factor receptor, β polypeptide	Pdgfrb	n. c.	−1.52
nitric oxide synthase 3, endothelial cell	Nos3	n. c.	−2.64#
integrin α 9	Itga9	−2.46	n. c.
integrin α 5(fibronectin receptor α)	Itga5	−3.48	n. c.
integrin β 1 (fibronectin receptor β)	Itgb1	−1.52	n. c.
laminin, α 2	Lama2	7.46	n. c.
laminin, α 4	Lama4	2.30	n. c.
laminin B1 subunit 1	Lamb1-1	2.64	n. c.
laminin, β 2	Lamb2	2.00	n. c.
procollagen, type IV, α 2	Col4a2	3.25	n. c.
procollagen, type V, α 1	Col5a1	2.14	1.52
procollagen, type I, α 1	Col1a1	2.00	n. c.
APOPTOSIS
perforin 1 (pore forming protein)	Prf1	−4.29	2.14
granzyme B	Gzmb	−5.66	9.85
BCL2-like 11 (apoptosis facilitator)	Bcl2l11	−1.87	1.41
nuclear factor of κ light polypeptide gene enhancer in B-cells inhibitor, ε	Nfkbie	−2.14	4.00
nuclear factor of κ light chain gene enhancer in B-cells 1, p105	Nfkb1	−1.32	1.52
nuclear factor of κ light chain gene enhancer in B-cells inhibitor, α	Nfkbia	−1.52	1.32
tumor necrosis factor receptor superfamily, member 21	Tnfrsf21	−1.52	2.00
tumor necrosis factor receptor superfamily, member 1b	Tnfrsf1b	−2.14	1.87
caspase 3	Casp3	−1.62	n. c.
caspase 6	Casp6	−1.74	n. c.
baculoviral IAP repeat-containing 5	Birc5	n. c.	−1.41
baculoviral IAP repeat-containing 4 (Birc4), mRNA	Birc4	n. c.	−1.52
OXIDATIVE STRESS RESPONSE
heme oxygenase (decycling) 1	Hmox1	−1.62	−4.92
glutamate-cysteine ligase, catalytic subunit	Gclc	n. c.	−1.74
glutathione reductase 1	Gsr	n. c.	−2.00
microsomal glutathione S-transferase 1	Mgst1	2.83	−1.74
EICOSANOID SYNTHESIS
prostaglandin-endoperoxide synthase 1	Ptgs1	n. c.	2.46
prostaglandin I2 (prostacyclin) synthase	Ptgis	−2.14	1.52
leukotriene C4 synthase	Ltc4s	−2.00	1.62
prostaglandin E synthase	Ptges	−1.87	1.62
arachidonate 15-lipoxygenase	Alox15	n. c.	3.25
arachidonate 5-lipoxygenase activating protein	Alox5ap	−1.52	n. c.
prostaglandin-endoperoxide synthase 2	Ptgs2	−3.03	n. c.
GLUCOSE METABOLISM
solute carrier family 2 (facilitated glucose transporter), member 4	Slc2a4	−2.14#	2.64
solute carrier family 2 (facilitated glucose transporter), member 1	Slc2a1	−3.25	2.30
hexokinase 1	Hk1	−1.87	2.30
hexokinase 2	Hk2	−1.74	2.64
glucose phosphate isomerase 1	Gpi1	−1.62	1.52
phosphofructokinase, muscle	Pfkm	−2.64	6.50
phosphofructokinase, liver, B-type	Pfkl	−1.32	1.74
phosphofructokinase, platelet	Pfkp	−1.52	1.52
Aldolase 1, A isoform (Aldoa), mRNA	Aldoa	−1.62	1.87
aldolase 3, C isoform	Aldoc	−2.30	n. c.
triosephosphate isomerase 1	Tpi1	−1.32	1.41
phosphoglycerate mutase 1	Pgam1	−1.41	n. c.
phosphoglycerate mutase 2	Pgam2	−4.92	34.30
pyruvate kinase, muscle	Pkm2	−1.41	n. c.
enolase 3, β muscle	Eno3	−3.25	17.15
enolase 2, neuronal	Eno2	n. c.	6.50
phosphatidylinositol 3-kinase catalytic	Pik3cd	−1.52	1.52
muscle glycogen phosphorylase	Pygm	−4.92	51.98
UDP-glucose pyrophosphorylase 2	Ugp2	1.41	1.62
phosphorylase kinase α 1	Phka1	−17.15	2.00
forkhead box O1	Foxo1	n. c.	1.52
forkhead box O3a	Foxo3a	n. c.	1.41
protein kinase C, β 1	Prkcb1	n. c.	1.62
protein kinase C, δ	Prkcd	−1.74	1.52
protein kinase C, ζ	Prkcq	n. c.	5.28
TRIGLYCERIDE SYNTHESIS
glycerol-3-phosphate dehydrogenase 1 (soluble)	Gpd1	n. c.	2.14
glyceronephosphate O-acyltransferase	Gnpat	n. c.	1.52
diacylglycerol O-acyltransferase 1	Dgat1	−1.62	n. c.
lipoprotein lipase	Lpl	3.03#	n. c.
ADIPOGENESIS
CCAAT/enhancer binding protein (C/EBP), β	Cebpb	−3.48#	1.74#
sterol regulatory element binding factor 1	Srebf1	n. c.	1.74#
adiponectin, C1Q and collagen domain	Adipoq	3.73	−6.96
resistin	Retn	n. c.	−6.06
adipsin	Adn	7.46#	−13.00#
fatty acid binding protein 4, adipocyte	Fabp4	1.62	−4.92#
lipoprotein lipase	Lpl	3.03#	n. c.

Microarray analysis of angiogenic gene expression

Diet high in saturated fat differentially regulated expression of the genes that affected angiogenic response associated with: cell cycle, adhesion, and matrix remodeling in ko mice and their wild type controls.

According to the recent monographic data that address proangiogenic activity of the endothelial cells, special attention was paid to the HF diet-regulated genes involved in the cell cycle and proliferation. Our study indicated that HF diet led to inhibition of the cell cycle G2/M phase gene expression (cyclins and cycline dependent kinases: Ccnb1, Cdc2a) and activation of G2/M phase inhibitor gene expression Ywhag in the matrigel plug cells from hRXRα ko mice. Inhibition of expression of the genes participating in DNA replication in G1 phase of cell cycle (Orc4L, Orc6L, Rpa3, Cdk7, Prim1, Prim2) was also observed.

In comparison, the HF diet resulted in downregulation of the cell cycle G1 phase genes (Ccnd3 and Ccnd2) and genes that encoded such inhibitors of cell cycle as p53 (Trp53) and p21 (Cdkn1a) in the cells infiltrating the matrigel plugs of WT mice. Some genes that affected DNA replication (Mcm3, Mcm6, Mcm7, Orc4L) were activated. In addition, upregulation of cyclins and kinases of G2 phase (Ccna2 and Cdc2a) as well as inhibition of expression of an inhibitor of G2 phase (Ywhag gene) was found.

Adhesion plays an important role in the contact inhibition of endothelial cell proliferation, vessel permeability and leukocyte migration. Regulation of adhesion is also required for a correct organization of new vessels in network [27]. Microarray studies of the cells from the matrigel plugs from hRXRα ko mice fed HF diet pointed to activation of genes encoding integrins participating in cell – matrix interactions (Itgam, Itgax, Itgb2, Itgb7), as well as inhibition of expression of the matrix protein genes such as collagens (Col11a1, Col3a1, Col4a1, Col5a2). Some of the genes encoding matrix metaloproteinases (Mmp9, Mmp12) were up-regulated. Moreover, inhibition of gap junction gene expression (Gja1, Gja4, Gjb5) was observed. Additionally, analysis of the cells infiltrating matrigel plugs from hRXRα ko mice maintained on the HF diet demonstrated down-regulation of the growth factor receptors genes such as VEGF receptor 2 (Kdr) and PDGF receptor (Pdgfrb), which activation played an important role in angiogenesis inducing proliferation and migration of endothelial cells. Gene expression of endothelial nitric oxide synthase (Nos3), an enzyme responsible for vasodilatation, endothelial cell protection and angiogenesis, was also inhibited.

In WT mice fed with HF diet, inhibition of genes encoding integrins (Itgam, Itgax, Itga9, Itga5, Itgb1, Itgb2) and activation of laminin (Lama2, Lama4, Lamb1-1, Lamb2) and collagen (Col3a1, Col4a1, Col4a2, Col5a1, Col5a2, Col6a2, Col1a1) gene expression was the main effect observed in the cells from the matrigel plugs.

Influence of HF diet on metabolic gene expression

The RXRα/PPARα heterodimer regulates expression of genes responsible for fatty acid metabolism in the liver [4, 11]. The hRXRα ko animals demonstrated a decreased expression pattern of genes responsible for liver fatty acid metabolism, what could explain the increased amount of fatty acids in the circulation and subsequently lipotoxicity and apoptosis of tissue cells [3]. As expected, the HF diet led to activation of apoptosis-related genes in the matrigel cells from the hRXRα ko, but not in WT mice. Genes encoding activators of caspase 3 (Prf1, Gzmb), pro-apoptotic genes Bcl2 (Bcl2L11), NFκB pathway (Nfkbie, Nfkb1), and TNFα receptors (Tnfrsf21, Tnfrsf1b) were up-regulated, while apoptosis inhibitors genes (Birc5, Birc4) were down-regulated. Opposite effects were observed in the matrigel plug cells from WT control mice. Expression of the genes encoding caspase activators (Prf1, Gzmb), Bcl2 (Bcl2L11), NFκB (Nfkbie, Nfkbia), caspases (Casp3, Casp6), TNFα (TNfrsf21, Tnfrsf1b) was decreased.

Additionally, the HF diet down-regulated expression of genes that affected detoxification processes such as heme oxygenase 1 (Hmox1), glutamatecysteine ligase (Gclc), glutathione reductase (Gsr), microsomal glutathione S-transferase 1 (Mgst1) pointing to an increased lipotoxicity in the hRXRα ko mice. In contrast, the above genes were up-regulated in the control animals.

It was previously reported that feeding with HF diet induced expression of the eicosanoid generating enzymes [19]. Prostanoids were proposed to be potent adipogenic substances that activated of PPARγ [15], while HF diet promoted adipogenesis [17]. Our microarray studies in the cells from matrigel plugs from the hRXRα ko mice fed HF diet demonstrated upregulation of the prostaglandin-endoperoxide synthase 1 (Ptgs1), prostaglandin I₂ (prostacyclin) synthase (Ptgis), leukotriene C4 synthase (Ltc4s), prostaglandin E synthase (Ptges), as well as arachidonate 15-lipoxygenase (Alox15). On the contrary, results in the control WT mice fed with HF diet indicate down- regulation of the genes related to the eicosanoid synthesis, including prostaglandin I₂ (prostacyclin) synthase (Ptgis), prostaglandin E synthase (Ptges), leukotriene C4 synthase (Ltc4s), arachidonate 5-lipoxygenase activating protein (Alox5ap), as well as prostaglandin-endoperoxide synthase 2 (Ptgs2). Also, expression of the genes that control substrate accumulation in the cells were differentially regulated by HF in the matrigel cells from the hRXRα ko mice and their WT counterparts. In this case, glycerol-3-phosphate dehydrogenase 1 (Gpd1) and glycerol phosphate O-acyltransferase (Gnpat) genes expression related to triglyceride synthesis were stimulated in the cells harvested from the matrigel plugs of the hRXRα ko mice. The WT mice responded to the HF diet with downregulation of diacylglycerol O-acyltransferase 1 (Dgat1) and upregulation of lipoprotein lipase (Lpl).

The HF diet activated expression of transcription factors important for adipogenesis in the cells from the hRXRα ko mice. Essential for adipogenesis transcription factors [6] such as CCAAT/enhancer binding protein (C/EBP) β (Cebpb), sterol regulatory element binding factor 1 (Srebf1) were up-regulated, but markers of differentiated adipocytes, such as adiponectin (Adipoq), resistin (Retn), adipsin (Adn) and fatty acid binding protein (Fabp4), were downregulated. Surprisingly, the HF diet down-regulated gene expression of a transcription factor that affected pre-adipocyte differentiation (Cebpb), whereas genes related to differentiated adipocytes: adiponectin (Adipoq), adipsin (Adn), fatty acid binding protein 4 (Fabp4), lipoprotein lipase (Lpl) were up-regulated in the WT mice.

Genes related to glucose metabolism were also differentially regulated in our mouse models by HF diet. In the hRXRα ko mice, gene expression of glucose transporters Glut4 (Slc2a4), Glut1 (Slc2a1) and several genes involved in glycolysis, including hexokinase 1 (Hk1), hexokinase 2 (Hk2), glucose phosphate isomerase 1 (Gpi1), phosphofructokinase, muscle (Pfkm), phosphofructokinase, liver, B-type (Pfkl), platelet phosphofructokinase (Pfkp), aldolase 1, Aisoform (Aldoa), triosephosphate isomerase 1 (Tpi1), phosphoglycerate mutase 2 (Pgam2), muscle β enolase 3, (Eno3), neuronal enolase 2, γ (Eno2) were upregulated by the high saturated fat diet. Activation of PI3K genes (phosphatidylinositol 3-kinase catalytic δ polypeptide, Pik3cd), and genes encoding proteins involved in glycogen turnover, such as muscle glycogen phosphorylase (Pygm), UDP-glucose pyrophosphorylase 2 (Ugp2) and phosphorylase kinase α 1 (Phka1), was also observed. In addition, transcription factors involved in insulin signaling (Foxo1, Foxo3a) and modulators of insulin action (protein kinases c Prkcb1, Prkcd, Prkcq) were up-regulated.

On the contrary, the HF diet in the matrigel cells harvested from the WT mice caused an overall down-regulation of the genes encoding glucose transporters Glut4 (Slc2a4), Glut1 (Slc2a1), genes involved in glycolysis (hexokinase 1 (Hk1), hexokinase 2 (Hk2), phosphofructokinase, muscle (Pfkm), phosphofructokinase, platelet (Pfkp), aldolase 1, A isoform (Al-doa), aldolase 1, C isoform (Aldoc), phosphoglycerate mutase 2 (Pgam2), phosphoglycerate mutase 1 (Pgam1), enolase 3, β muscle (Eno3), pyruvate kinase, muscle (Pkm2)) and catabolism of glycogen (Pygm, Phka1). The genes associated with signal transduction from insulin receptor was also downregulated in WT mice by HF diet.

Quantitative real time-PCR results

Direction of changes in expression of the adipogenesis and angiogenesis related genes in response to the HF diet in the cells from matrigel plugs from the two mouse models was in agreement with the microarray screening. The qRT-PCR results confirmed inhibition of the critical for angiogenesis markers of endothelial cells proliferation genes (Kdr, Nos3) in the cells from the hRXRα ko mice fed HF diet. Also, inhibition of expression of the gene markers for differentiated adipocytes (Fabp4, Adn) and up-regulation of the genes related to early differentiation of adipocytes (Cebpb, Srebf1) were observed (Tab. 2, Fig. 4).

Fig 4.

Real time PCR confirmation of microarray changes in relative gene expression (between matrigel plug cells from mice fed standard diet (ST) and matrigel plug cells from mice fed high saturated fat diet (HF)) in RXR α deficient mice (hRXRα ko) and WT mice in pathways connected with angiogenesis (VEGF receptor 2: Kdr, endothelial nitric oxide synthase: Nos3. platelet endothelial cell adhesion mole-cule-1: Pecam1), adipogenesis (adipsin Adn, fatty acid binding protein 4: Fabp4, lipoprotein lipase: Lpl, CCAAT/ enhancer binding protein (C/EBP) β Cebpb, sterol regulatory element binding factor 1: Srebf1), energy expenditure (uncoupling protein 2 (mitochondrial proton carrier): Ucp2) and glucose uptake (solute carrier family 2 (facilitated glucose transporter), member 4: Slc2a4). Gapdh was used as reference gene

Regulation of the gene expression in the matrigel plug cells of the WT mice fed with HF diet was in accordance with the microarray results concerning angiogenesis (down-regulation of Kdr and Nos3), adipogenesis (up-regulation of Lpl, down-regulation of Cebpb), glucose uptake (inhibition of glucose transporter Glut4 (Slc2a4)) (Tab. 2).

Discussion

In the presented study, the hRXRα ko mice maintained on the HF diet demonstrated significantly elevated amounts of serum cholesterol and leptin without developing obesity, what is characteristic for this experimental model and remains in contrast to the WT mice [4, 29, 31–33]. Our observations might be due to peripheral lipotoxicity [28], increased energy expenditure (enhanced UCP2 mRNA level, observed also in our RT-PCR results) and/or decreased angiogenesis [7, 18, 23]. Our results indicated that the hRXRα ko mice on the HF diet, despite increased amounts of proangiogenic leptin [21] in serum, demonstrated tendency to decreased angiogenic response and capillary network differentiation (lumen formation) in our matrigel model of angiogenesis. The microarray results confirmed inhibition of the important angiogenesisrelated genes in the cells isolated from these mice. On the contrary, activation of the genes related to the early steps of adipogenesis but not the final adipocyte differentiation, was observed in the matrigel cells of the hRXRα ko animals fed with the HF diet. For example, down-regulation of Kdr, the receptor for VEGF, and inhibition of endothelial nitric oxide synthase (Nos3) genes expression were observed in the microarray and qRT-PCR studies. Both VEGFA and NO play important roles in angiogenesis by promoting endothelial cells proliferation and vascular network formation [1, 5, 23].

Inhibition of the early angiogenic response was also supported by the microarray studies, where decreased expression of the genes encoding gap junction proteins, matrix remodeling enzymes along with up-regulation of the genes indicative of lipotoxicity and/or apoptosis activation were found.

In the hRXRα ko mice, impaired fatty acid utilization in the liver and subsequently endothelial exposure to dyslipidemia lead to endothelial dysfunction, lipotoxicity and oxidative stress [8, 13]. It was shown, that free fatty acids (FFAs), PPARs’ activators, inhibited angiogenesis in in vitro studies [12]. Our analysis of the gene expression in the cells in the matrigel plug of hRXRα ko animals revealed up-regulation of the genes involved in the triglyceride synthesis (Gpd1, Gnpat). An excess of FFAs may lead to lipotoxicity of endothelium by increased de novo synthesis of diacylglycerol [14], which activates isoforms of protein kinase C (PKC) and leads to stimulation of endothelial superoxide production by NAD(P)H oxidase, inhibition of endothelial NO synthase activity and activation of NF-κB and proinflammatory pathway [9]. In our microarray study, we indeed observed activation of PKC gene expression parallel with inhibition of the antioxidant genes (Hmox1, Gdc, Gsr, Mgst1), inhibition of Nos3 and up-regulation of proinflammatory pathways NFκB (Nfkbie) and TNFα receptors (Tnfrsf21, Tnfrsf1b) in the knockout mice on the HF diet. FFA induced endothelial lipotoxicity and utilization of glucose as a source of energy as evidenced by up-regulation of the genes encoding glycogenolysis enzymes, Glut4 (Slc2a4) and Glut1 (Slc2a1) transporters and activation of glucose/cell cycle related transcription factors (Foxo1, Foxo3).

Additionally, our microarray analysis verified activation of the pro-apoptotic genes that might impair angiogenic response in the cells of the knockout mice on the HF diet. Increase of caspase 3 activity through Prf1, Gzmb, up-regulation of proapoptotic protein Bcl2 (Bcl2L11) and inhibition of anti-apoptotic proteins gene expression (Birc4, Birc5) argued for activated apoptosis [26]. On the contrary, expression of the apoptotic genes such as: caspase activators (Prf1, Gzmb), caspases (Casp3, Casp6) was down-regulated by the HF diet in the control mice.

Microarray analysis as well as quantitative real time PCR results also indicated that beside downregulation of the pro-angiogenic genes, the HF diet activated transcription factors related to early adipogenesis such as Srebf1, Cebpb [6] in the cells that formed primitive vessel networks in the matrigel from the hRXRα ko mice, what was also in agreement with the increased amounts of serum leptin in these mice. However, this adipokine was also reported to inhibit maturation of the adipocytes [2].

Scarfo et al. [24] documented that arachidonic acid and its metabolites induced genes of enzymes involved in eicosanoid synthesis, what was associated with formation of cellular lipid droplets. The expression of the enzymes participating in the eicosanoid synthesis was up-regulated in the presented study. The amounts of mRNA encoding prostaglandin-endoperoxide synthase 1 (Ptgs1), prostaglandin I₂ (prostacyclin) synthase (Ptgis) and arachidonate 15-lipoxygenase (Alox15) were significantly increased in the cells migrating to the hRXRα ko animals. Previously, it was shown that eicosanoids potentiated adipocyte differentiation by activating PPARγ [15]. The well known effect of angiogenesis inhibition by PPARγ activators [12] was in agreement with the obtained results.

In summary, HF diet resulted in dyslipidemia and hyperleptinemia, inhibition of the angiogenic response, and activation of adipogenesis in the cells migrating to the matrigel plugs in the hepatocyte RXR α deficient mice. Inhibition of angiogenesis might also result from increased apoptosis of the new capillary network forming cells. These results may add to the multifactorial background of pathological angiogenesis observed in diabetes.

Abbreviations

Adn

adipsin

apoCIII

apolipoprotein CIII

bFGF

fibroblast growth factor-2

Cebpb

CCAAT/enhancer binding protein (C/EBP) β

DNA

deoxyribonucleic acid

Fabp4

fatty acid binding protein 4

FFAs

free fatty acids

Gapdh

glyceraldehyde 3-phosphate dehydrogenase

HF

high fat diet

hRXRα ko

hepatocyte retinoid X receptor α deficient

IGF-1

insulin-like growth factor 1

Kdr

VEGF receptor 2

MS

metabolic syndrome

Lpl

lipoprotein lipase

LXR α

liver X receptor α

NFκB

nuclear factor κB

NO

nitric oxide

Nos3

endothelial nitric oxide synthase

ox LDL

oxidized low density lipoproteins

PCR

polymerase chain reaction

Pecam1

platelet endothelial cell adhesion molecule-1

PEDGF

pigmented epithelium derived growth factor

PPARs

peroxisome proliferator activated receptors

qRT-PCR

quantitative real-time PCR

RNA

ribonucleic acid

RXRα

retinoid X receptor α

SDF1

stromal-derived factor-1

Slc2a4

solute carrier family 2 (facilitated glucose transporter), member 4

Srebf1

sterol regulatory element binding factor 1

TNFα

tumor necrosis factor α

Ucp2

uncoupling protein 2

VEGF

vascular endothelial growth factor

WT

wild type