Comprehensive gene expression analyses of the rat prefrontal cortex after oxysterol treatment

Authors


Address correspondence and reprint requests to Dr. Wei-Yi Ong, Department of Anatomy, National University of Singapore, Singapore 117597. E-mail: wei_yi_ong@nuhs.edu.sg

Abstract

Oxysterols such as 7β-hydroxycholesterol (7β-HC) and 7-ketocholesterol (7-KC) have been linked to the pathophysiology of neurodegenerative diseases. This study was carried out to examine the effect of oxysterols on global gene expression in the rat prefrontal cortex (PFC). 7β-HC, 7-KC, or cholesterol was injected into the rat PFC and RNA was extracted from this brain region at 24-h post-injection and analyzed. Microarray analyses identified 1365 genes, whose expressions were affected by both 7β-HC and 7-KC. Among these, down-regulated genes outnumbered up-regulated genes. Pathway analysis showed that down-regulated genes had roles in carbohydrate metabolism, cell signaling and nucleic acid metabolism; and the majority of these encode G-protein coupled receptors (GPCRs). Expression of selected genes were validated by quantitative real-time PCR. Western blots confirmed down-regulation of oxytocin receptor (Oxtr) at 1 day post-7β-HC treatment. Immunohistochemical analysis showed localization of Oxtr in neurons of the PFC. Electron microscopy identified the presence of Oxtr-immunoreactivity in axon terminals. Together, these findings provide insights into molecular mechanisms through which oxysterols could exert their pathophysiological effects, and suggest that increased oxysterols may affect synaptic function by transcriptional repression of GPCRs.

Abbreviations used
7-KC

7-ketocholesterol

7α-HC

7α-hydroxycholesterol

7β-HC

7β-hydroxycholesterol

Cckar

cholecystokinin A receptor

DEG

differentially expressed gene

ERK

extracellular signal-regulated kinase

GPCR

G-protein coupled receptor

Insig

insulin-induced gene

IPA

ingenuity pathway analysis

LXR

liver X receptor

MAPK

mitogen-activated protein kinase

NF-κB

nuclear factor kappa B

Npffr1

neuropeptide FF receptor 1

Ntsr1

neurotensin receptor 1

Ntsr2

neurotensin receptor 2

Oxtr

oxytocin receptor

PBS

phosphate-buffered saline

PCGEM

parametric test based on cross-gene error model

PFC

prefrontal cortex

PKA

cAMP-dependent protein kinase

RIN

RNA integrity number

RIPA

radioimmunoprecipitation assay

RXR

retinoid X receptor

SREBP

sterol response element binding protein

The brain is the most cholesterol-rich organ in the body (Björkhem and Meaney 2004). High demand for brain cholesterol is supplied by de novo synthesis, whereas elimination of excess cholesterol is facilitated by its conversion to a more polar cholesterol oxidation products or oxysterol (Björkhem and Meaney 2004; Vejux et al. 2008). These can be generated by enzymatic action or direct oxidation of cholesterol. Enzymatic oxidation of the cholesterol side-chain generates 24-, 25-, and 27-hydroxycholesterols mainly through the action of cytochrome P450 family enzymes (Brown and Jessup 2009). In addition, oxysterols can be generated by autoxidation of cholesterol. This results in the formation of 7α-hydroxycholesterol (7α-HC) and 7β-hydroxycholesterol (7β-HC), 7-ketocholesterol (7-KC), 5α,6α-epoxycholesterol, 5β,6β-epoxycholesterol, and cholestane-3β,5α,6β-triol (Vejux et al. 2008).

Oxysterols are maintained at 103- to 106-fold lower levels compared with cholesterol under physiological conditions (Björkhem and Diczfalusy 2002; Brown and Jessup 2009). On the other hand, increased levels of oxysterols have been implicated in the pathophysiology of atherosclerosis (Brown and Jessup 1999), Alzheimer's disease (Lütjohann et al. 2000; Vaya and Schipper 2007; Leoni 2009; Leoni and Caccia 2011), age-macular degeneration (Malvitte et al. 2006), osteoporosis (Liu et al. 2005), and cancer (Jusakul et al. 2011). 7β-HC and 7-KC levels are elevated in atherosclerotic lesions and plasma of cardiovascular patients, and subjects on a high fat diet (Brown and Jessup 1999; Colles et al. 2001; Guardiola et al. 2002). Increased 7β-HC and 7-KC are also detected in the rat hippocampus after excitotoxic neuronal injury, induced by the glutamate analog, kainate (Kim et al. 2010, 2011).

Oxysterols differ from cholesterol in the presence of an additional polar moiety, which facilitates their transport through cellular membranes (Brown and Jessup 2009; Jusakul et al. 2011). Incorporation of these polar moieties into the hydrophobic regions of cell membrane induces local re-arrangement of acyl chains (Meaney et al. 2002; Jusakul et al. 2011), and oxysterols have been reported to increase the fluidity of brain synaptic plasma membrane (Wood et al. 1995). 7β-HC and 7-KC were found to induce cell death and inflammation (Vejux et al. 2008). Moreover, enhanced exocytosis was observed in PC12 cells after treatment with 7-KC, 24-hydroxycholesterol, or cholesterol 5,6β epoxide (Ma et al. 2010). These findings suggest that increased levels of oxysterols could promote cellular damage in the brain, and may be a contributing factor in excitotoxic brain injury (Ong et al. 2010).

Despite increasing evidence that oxysterols are elevated in neurodegenerative conditions and implicated in neuronal cell death, little is known about possible effects of oxysterols on gene expression in the brain. This study was therefore carried out using microarray analyses, to elucidate comprehensive gene expression changes in the rat prefrontal cortex (PFC) after oxysterol treatment.

Methods

Animals and treatment

All procedures involving animals were approved by the Institutional Animal Care and Use Committee (IACUC) of the National University of Singapore, in accordance with National Advisory Committee for Laboratory Animal Research Guidelines. Adult male Wistar rats, weighing approximately 300 g each, were purchased from the Centre for Animal Resources (CARE), National University of Singapore. They were anesthetized with a ketamine (75 mg/kg) and xylazine (10 mg/kg) cocktail, and placed on a stereotaxic apparatus (Stoelting, Wood Dale, IL, USA). 7β-HC, 7-KC, cholesterol (Steraloids, Newport, RI, USA, 0.8 nmol in 1 μL solution, dissolved in 5% ethanol), or vehicle (5% ethanol) were stereotaxically injected into the right PFC (4.7 mm rostral to bregma, 2.5 mm lateral to midline, and 1.6 mm from surface of the cortex) (Paxinos and Watson 1996) through a small craniotomy. The needle was withdrawn 5 min after the intracortical injection, and the scalp sutured. Rats were housed under controlled conditions (23 ± 1°C, 12:12 h light-dark cycle) and were given standard laboratory chow and water ad libitum. Administration of test compounds, 7β-HC, 7-KC, or cholesterol at same concentration in the rat PFC was carried out, to compare the effects of the two autoxidatively formed oxysterols versus cholesterol on gene expression in the brain, relative to vehicle-injected controls. The total dose of test compounds (1 μL of 0.8 mM solution of each test compound) was determined based on the excess amount of oxysterol induced by kainate excitotoxicity in the rat hippocampus (approximately 2.8 ng/mg) (Kim et al. 2011) multiplied by the average estimated mass of the rat PFC (115 mg).

RNA extraction

Rats were deeply anesthetized by a ketamine (75 mg/kg) and xylazine (10 mg/kg) cocktail and killed 24 h after injection of vehicle, oxysterols, or cholesterol. The right PFC was harvested, immersed in RNAlater® (Ambion, Austin, TX, USA), snap-frozen in liquid nitrogen, and stored in the −80°C freezer until they were used for analyses. Total RNAs were extracted from each tissue sample, isolated using Trizol reagent (Invitrogen, Carlsbad, CA, USA), and purified with the RNeasy® Mini Kit (Qiagen, Valencia, CA, USA) following the manufacturer's protocol. Samples were subjected to reverse transcription using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). The conditions for reverse transcription reaction were 25°C for 10 min, 37°C for 120 min, and 85°C for 5 s.

Microarray data collection and analysis

Gene expression profiles were elucidated using the Agilent SurePrint G3 Rat GE 8x60K Microarray (Agilent Technologies, Santa Clara, CA, USA). 5 μL of total RNA of the PFC from each rat was submitted to the BFIG Core Facility Lab (National University of Singapore, Department of Paediatrics). The quality of RNA was analyzed using the Agilent 2100 Bioanalyzer, to ensure that only these with RNA integrity number (RIN) greater than 8.0 were used for analyses. cRNA generated from each sample was labeled using the one-cycle target labeling method and hybridized to a single array according to standard Agilent protocols. Data collected were exported to GeneSpring GX10 (Agilent Technologies) software for analysis, using parametric test based on cross gene error model (PCGEM). Differentially expressed genes (DEGs) were defined as genes which were altered after administration of test compounds compared with vehicle-treated rats, and were identified using one-way anova analysis. < 0.05 was considered significant.

Network analyses

DEGs were subjected to Ingenuity Pathway Analysis (IPA) (http://www.ingenuity.com; Ingenuity Systems, Mountain View, CA, USA). Each identifier mapped to its corresponding object in Ingenuity Knowledge Base was set to identify molecules whose expressions were differentially regulated. These molecules which were also known as Network Eligible Molecules, were overlaid onto a global molecular network developed from information contained in the Ingenuity Knowledge Base. Networks of Network Eligible Molecules were then algorithmically generated based on their connectivity.

Real-time PCR analyses

Real-time PCR was conducted using the 7500 Real-Time PCR system (Applied Biosystems) to verify some of the gene expression changes detected by microarray. The preparation of samples and solutions was carried out according to the manufacturer's protocol using TaqMan® Universal PCR Master Mix (Applied Biosystems) with each specific probe (Applied Biosystems), and the procedure was carried out as previously described (Kim et al. 2011). The mean and standard error were calculated, and possible significant differences among different treatment groups analyzed using one-way anova with Bonferroni's multiple comparison post-hoc test. < 0.05 was considered significant.

Western blot analyses

Rats were treated with 7β-HC or vehicle, and killed at 24 h after injection. The right PFC was harvested, snap-frozen in liquid nitrogen, and stored in the −80°C freezer until analysis. Tissue samples were weighed and homogenized in radioimmunoprecipitation assay (RIPA) lysis buffer system (Santa Cruz Biotechnology,Santa Cruz, CA, USA) according to the manufacturer's protocol. Samples were centrifuged at 12 000 g and 4°C for 30 min to separate cellular protein extracts from tissue debris. The protein concentration of the supernatant was determined using the Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA). Homogenates (50 μg) were resolved in 10% SDS-polyacrylamide gel under reducing conditions for approximately 2 h and the subsequent procedures were carried out as previously described (Kim et al. 2011), with a rabbit polyclonal anti-oxytocin receptor (Oxtr) antibody (Alpha Diagnostic, San Antonio, TX, USA; diluted 1 : 500 in blocking buffer). The densities of Oxtr bands were normalized against those of β-actin, and the mean and standard error calculated. Possible significant differences between 7β-HC and vehicle treatments were analyzed using Student's t-test. < 0.05 was considered significant. Rat uterus tissue was used as a positive control.

Immunohistochemistry

Four untreated rats were deeply anesthetized and cardiac perfusion was performed with a solution of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The fixed brain tissues were removed and blocks containing the PFC were sectioned coronally at 100 μm using a vibrating microtome. Free floating sections were washed with phosphate-buffered saline (PBS) to remove traces of fixative and incubated overnight with a rabbit polyclonal anti-Oxtr antibody (diluted 1 : 25 in PBS). Subsequent procedures were carried out as previously described (Kim et al. 2011). Some sections were mounted on gelatin-coated slides and lightly counter-stained with methyl green before coverslipping, while the rest were processed for electron microscopy.

Electron microscopy

Some of the Oxtr immunolabeled brain sections were subdissected into smaller portions that included the PFC. These were osmicated, dehydrated in an ascending series of ethanol and acetone, and embedded in Araldite. Thin sections were obtained from the first 5 μm of the sections, mounted on copper grids coated with 0.3% Formvar, stained with lead citrate, and examined using a Jeol 1010EX electron microscope (Jeol, Tokyo, Japan).

Results

Microarray data collection and analysis

cDNA microarray analyses identified changes in the gene expression with more than twofold change that are both common and exclusive to 1 day post-7β-HC, -7-KC, -cholesterol, or -vehicle treatments in the injected right PFC. As shown in Fig. 1a, 7-KC treatment produced the highest number of DEGs with a total of 2024 genes. Of the DEGs regulated by 7-KC treatment, 1659 and 365 genes were down- and up-regulated, respectively. 7β-HC treatment resulted in 1588 DEGs, of which 1186 and 402 genes were down- and up-regulated, respectively. Only 39 DEGs were detected after cholesterol treatment, of which 30 and 9 were down- and up-regulated.

Figure 1.

(a) Venn diagram showing the classification of differentially expressed genes (DEGs) in the prefrontal cortex (PFC) after 1 day post-treatment with 7β-hydroxycholesterol (7β-HC), 7-ketocholesterol (7-KC), or cholesterol. The numbers in parenthesis represent the number of genes that were both down- and up-regulated with more than twofold change and < 0.05 as compared with vehicle treatment. (b) Network view of down-regulated genes common in 7β-HC and 7-KC treated groups with DEGs greater than twofold change. Ingenuity Pathway Analysis (IPA) identified these genes were involved in carbohydrate metabolism, cell signaling, and nucleic acid metabolism (as shown in Table 1) as the largest network for down-regulated genes. Data are represented as nodes displayed using various shapes that represent the functional class of the gene product. DEGs in these pathways are indicated in gray nodes. Solid and dotted lines indicate direct and indirect interactions, respectively.

DEGs exclusive to 7β-HC treatment

A total of 213 DEGs with more than twofold change were exclusive to 7β-HC treatment compared with vehicle treatment (Fig. 1a). Of these, 131 genes were down-regulated and 82 were up-regulated. Down-regulated genes with the highest fold change included: vestigial like 1 homolog (Drosophila) (Vgll1), regenerating islet-derived 3 beta (Reg3b), and predicted: hypothetical protein LOC686013 (LOC686013) (Supplementary material Table S1). Up-regulated genes with the highest fold change included: predicted: similar to preferentially expressed antigen in melanoma like 4 (LOC298613), predicted: similar to insulin-like growth factor binding protein 4 precursor (RGD1562534), and TAF7-like RNA polymerase II, TATA box binding protein (TBP)-associated factor (Taf7 l).

DEGs exclusive to 7-KC treatment

A total of 649 DEGs with more than twofold change were exclusive to 7-KC treatment compared with vehicle treatment (Fig. 1a). Of these, 601 genes were down-regulated and 48 were up-regulated. Down-regulated genes with the highest fold change included: ribonuclease, RNAse A family, 1-like 1 (Rnase1 l1), resistin-like gamma (Retnlg), and predicted: similar to sentrin 15 (RGD1559489) (Supplementary material Table S2). Up-regulated genes with the highest fold change included: pentraxin related gene (Ptx3), T-cell immunoglobulin and mucin domain containing 2 (Timd2), and WW domain-containing oxidoreductase (Wwox).

DEGs common to 7β-HC and 7-KC treatments

Large number of DEGs was found in this category, which included DEGs that were altered in common after treatment with 7β-HC or 7KC. This category also accounted for the highest percentage of DEGs, i.e. 1365 DEGs out of a total of 2266 DEGs after treatment with these two test compounds versus vehicle controls, or 60.2% (Fig. 1a). Genes with the highest fold change of more than sixfold were also found in this category (Supplementary material Table S3). Of the 1365 DEGs, 1050 genes were down-regulated and 315 were up-regulated. Down-regulated genes with the highest fold change included: predicted: similar to tolloid-like 2 (LOC680012), predicted: similar to NEDD4-binding protein 1 (LOC688874), and similar to orphan sodium- and chloride-dependent neurotransmitter transporter NTT5 (RGD1562492). Up-regulated genes with the highest fold change included: peroxisomal trans-2-enoly-CoA reductase (Perc), predicted: similar to glyceraldehyde-3-phosphate dehydrogenase (LOC307263), and nudix (nucleoside diphosphate linked moiety X)-type motif 9 (Nudt9).

DEGs common to 7β-HC and cholesterol treatments

There were six DEGs in this group (Fig. 1a). Of these, two were down-regulated and four were up-regulated by 7β-HC, whereas five were down-regulated and one was up-regulated by cholesterol. Down-regulated genes by 7β-HC treatment with highest fold change included: troponin T, cardiac muscle (Tnnt2), and myelin protein zero-like 2 (Mpzl2) (Supplementary material Table S4). Up-regulated genes by 7β-HC treatment with highest fold change included: similar to nucleoside diphosphate-linked moiety X motif 16 (nudix motif 16) (LOC688828), predicted: hypothetical protein LOC100365120 (LOC100365120), and Da1-10-like (LOC100366054).

DEGs common to 7-KC and cholesterol treatments

There were six DEGs in this group (Fig. 1a). Of these, five genes were down-regulated and one was up-regulated. The down-regulated genes with highest fold change included: GDNF family receptor alpha like (Gfral), solute carrier family 9 (sodium/hydrogen exchanger), member 4 (Slc9a4), and phosphatidylinositol glycan, class N (Pign) (Supplementary material Table S5). The up-regulated gene was trypsin 10 (Try10).

DEGs common to 7β-HC, 7-KC, and cholesterol treatments

There were four genes in this group (Fig. 1a). All four genes were down-regulated by cholesterol, while one of the genes was up-regulated by both 7β-HC and 7-KC treatments. Down-regulated genes included: flavin containing monooxygenase 9 (Fmo9), insulin-like growth factor binding protein-like 1 (Igfbpl1), and wingless-type MMTV integration site family, member 9B (Wnt9b) (Supplementary material Table S6). Ankyrin repeat domain 57 (Ankrd57) was up-regulated by 7β-HC and 7-KC, but down-regulated by cholesterol.

Networks for down-regulated DEGs after treatments with 7β-HC and 7-KC

The largest category of DEGs that showed more than twofold change, and was found in common after or 7β-HC or 7-KC treatment was selected for further analysis using IPA, to identify potential functional networks. The top network with the highest number of down-regulated DEGs was related to carbohydrate metabolism, cell signaling, and nucleic acid metabolism (Table 1, Supplementary material Table S7, and Fig. 1b). This network had 30 focus molecules and included the following genes: Taar5, Oxtr, Avpr1b, Gpr152, Gpr157, P2ry4, Gpr112 l, Gpr35, Gpr142, Npvf, Gpr63, Tas1r3, Cckar, F2rl3, Lgr5, Npffr1, Ghsr, Pcp2, Tas1r2, Gpr44, Gpr156, Taar8c, Uts2r, Ccrl2, Sucnr1, Uts2d, V1rl1, Gpr132, S1pr4, and Ntsr1. Most of these genes encode G-protein coupled receptors (GPCRs).

Table 1. Down-regulated DEGs in the PFC found in common to 1 day post-7β-HC and -7-KC treatments. These are the 30 genes that were mapped in the top network related to carbohydrate metabolism, cell signaling, and nucleic acid metabolism by IPA
GeneGene symbol7β-HC7-KC
Fold changep-valueFold changep-value
Trace amine-associated receptor 5Taar5−5.350.0001−10.64<0.0001
Oxytocin receptorOxtr−4.390.0001−6.91<0.0001
Arginine vasopressin receptor 1BAvpr1b−4.16<0.0001−4.53<0.0001
Predicted: G-protein coupled receptor 152Gpr152−3.32<0.0001−4.00<0.0001
G-protein coupled receptor 157Gpr157−3.320.0009−3.290.0009
Pyrimidinergic receptor P2Y, G-protein coupled, 4P2ry4−3.260.0038−2.330.0073
Predicted: G-protein coupled receptor 112 likeGpr112 l−3.22<0.0001−5.88<0.0001
G-protein coupled receptor 35Gpr35−3.180.0001−3.860.0001
Predicted: G-protein coupled receptor 142Gpr142−3.170.0002−5.340.0002
Neuropeptide VF precursorNpvf−3.000.0002−3.180.0001
G-protein coupled receptor 63Gpr63−2.860.0056−3.550.0009
Taste receptor, type 1, member 3Tas1r3−2.780.0019−3.760.0046
Cholecystokinin A receptorCckar−2.750.0077−2.970.0005
Coagulation factor IIF2rl3−2.690.0053−3.190.0004
Leucine rich repeat containing G-protein coupled receptor 5Lgr5−2.62<0.0001−2.800.0124
Neuropeptide FF receptor 1Npffr1−2.590.0176−3.27<0.0001
Growth hormone secretagogue receptorGhsr−2.390.0024−3.49<0.0001
Purkinje cell protein 2Pcp2−2.38<0.0001−2.74<0.0001
Taste receptor type 1 member 2 precursorTas1r2−2.360.0060−2.650.0163
G-protein coupled receptor 44Gpr44−2.310.0157−3.730.0001
G-protein coupled receptor 156Gpr156−2.310.0004−2.480.0003
Trace amine-associated receptor 8cTaar8c−2.290.0001−2.71<0.0001
Urotensin 2 receptorUts2r−2.270.0050−2.380.0084
Chemokine (C-C motif) receptor-like 2Ccrl2−2.200.0011−2.91<0.0001
Succinate receptor 1Sucnr1−2.170.0001−2.790.0027
Urotensin 2 domain containingUts2d−2.120.0004−3.580.0001
Vomeronasal 1 receptor, L1V1rl1−2.110.0002−2.160.0009
G-protein coupled receptor 132Gpr132−2.090.0001−2.93<0.0001
Sphingosine-1-phosphate receptor 4S1pr4−2.070.0016−2.500.0038
Neurotensin receptor 1Ntsr1−2.060.0005−2.610.0001

The network with the second highest number of down-regulated DEGs had 25 focus molecules and was related to lipid metabolism, molecular transport, and small molecule biochemistry (Supplementary materials Table S7 and Figure S1). Many of the DEGs were connected to nuclear factor kappa B (NF-κB) and retinoid X receptor (RXR).

The third network of DEGs was related to endocrine system development and function, small molecule biochemistry, and molecular transport with 25 focus molecules (Supplementary material Table S7).

Networks for up-regulated DEGs after treatments with 7β-HC and 7-KC

The top network with the highest number of up-regulated DEGs was related to cell death, cell morphology, cellular assembly and organization (Table 2, Supplementary material Table S8, and Fig. 2). This network had 26 focus molecules and the genes involved were Kcnh2, Txnip, Tpt1, Eif3 h, Rps23, Lap3, Tsg101, Cbx3, Erc1, Rnf7, Clint1, Tcc28, Dot1 l, S100b, Stmn1, Hsp90ab1, Eef1a1, Rpap3, Ep400, Cdc25c, Hist1h4b, Irak4, Ap2b1, Fbxo33, Adrm1, and Snai1. Up-regulated genes in this network were related to NF-κB.

Table 2. Up-regulated DEGs in the PFC found in common to 1 day post-7β-HC and -7-KC treatments. These are the 26 genes that were mapped in the top network related to cell death, cell morphology, cellular assembly and organization by IPA
GeneGene symbol7β-HC7-KC
Fold changep-valueFold changep-value
Potassium voltage-gated channel, subfamily H (eag-related), member 2Kcnh29.27<0.00017.91<0.0001
Thioredoxin interacting proteinTxnip6.81<0.00015.79<0.0001
Tumor protein, translationally controlled 1Tpt16.39<0.00015.51<0.0001
Eukaryotic translation initiation factor 3, subunit HEif3 h5.91<0.00015.08<0.0001
Ribosomal protein S23Rps235.76<0.00014.67<0.0001
Leucine aminopeptidase 3Lap35.70<0.00015.400.0001
Tumor susceptibility gene 101Tsg1015.65<0.00014.82<0.0001
Chromobox homolog 3 (HP1 gamma homolog, Drosophila)Cbx35.40<0.00014.64<0.0001
ELKS/RAB6-interacting/CAST family member 1Erc14.97<0.00015.090.0002
Ring finger protein 7Rnf74.68<0.00014.02<0.0001
(HP1 gamma homolog, Drosophila)Clint14.39<0.00013.56<0.0001
Predicted: similar to TPR repeat-containing protein KIAA1043Tcc284.060.00064.52<0.0001
DOT1-like, histone H3 methyltransferase (S. cerevisiae)Dot1 l3.970.00033.740.0008
S100 calcium binding protein BS100b3.71<0.00013.18<0.0001
Stathmin 1Stmn13.63<0.00013.18<0.0001
Heat-shock protein HSP 90-beta (Heat shock 84 kDa)Hsp90ab13.12<0.00012.72<0.0001
Eukaryotic translation elongation factor 1 alpha 1Eef1a13.04<0.00012.65<0.0001
RNA polymerase II associated protein 3Rpap33.01<0.00013.25<0.0001
E1A binding protein p400Ep4002.89<0.00012.69<0.0001
Cell division cycle 25 homolog C (S. pombe)Cdc25c2.880.00022.530.0002
Histone cluster 1, H4bHist1h4b2.830.00012.870.0021
Interleukin-1 receptor-associated kinase 4Irak42.690.00232.030.0067
Adaptor-related protein complex 2, beta 1 subunitAp2b12.41<0.00012.290.0001
F-box protein 33Fbxo332.310.00012.62<0.0001
Adhesion regulating molecule 1Adrm12.220.00022.290.0004
Snail homolog 1 (Drosophila)Snai12.170.00112.19<0.0001
Figure 2.

Network view of up-regulated genes common in 7β-hydroxycholesterol (7β-HC) and 7-ketocholesterol (7-KC) treated groups with differentially expressed genes (DEGs) greater than twofold change. Ingenuity Pathway Analysis (IPA) identified these genes were involved in cell death, cell morphology, cellular assembly and organization (as shown in Table 2) as the largest network for up-regulated genes. Data are represented as nodes displayed using various shapes that represent the functional class of the gene product. DEGs in these pathways are indicated in gray nodes. Solid and dotted lines indicate direct and indirect interactions, respectively.

The network with the second highest number of up-regulated DEGs had 21 focus molecules and was related to cancer, cell-mediated immune response, and cellular development (Supplementary materials Table S8 and Figure S2). DEGs in this network were related to the mitogen-activated protein kinase (MAPK) and extracellular signal-regulated kinase (ERK) family such as P38 MAPK and ERK1/2.

Real-time PCR analyses

Real-time PCR was performed to validate some selected genes from the largest down-regulated network ‘carbohydrate metabolism, cell signaling, and nucleic acid metabolism’ mapped by IPA. In this study, the genes to be verified were chosen based on the availability of commercial TaqMan® probes. There was significant down-regulation of Oxtr mRNA levels following 7β-HC and cholesterol treatments by 58% and 55% as compared with vehicle treatment (= 0.0165 and 0.0410, respectively, Fig. 3a), respectively. Cckar showed down-regulation in mRNA levels after 7β-HC, 7-KC, and cholesterol treatments with 68%, 63%, and 61% reduction compared with vehicle treatment (= 0.0009, 0.0024, and 0.0036, respectively, Fig. 3b), respectively. 7β-HC treatment reduced Npffr1 mRNA levels by 55% as compared with vehicle treatment (= 0.0006) and there was also significant down-regulation of Npffr1 mRNA levels after 7β-HC treatment as compared with 7-KC (= 0.0215, Fig. 3c). Treatment with cholesterol resulted in 52% lower mRNA levels of Ntsr1 than vehicle treatment (= 0.0087, Fig. 3d). Changes in Taar5, Avpr1b, Gpr35, P2ry4, and Tas1r2 as shown by microarray analyses could not be verified by real-time PCR in this study.

Figure 3.

Real-time PCR analyses of the rat prefrontal cortex (PFC) to validate the expression changes of the four genes (a, Oxtr; b, Cckar; c, Npffr1; and d, Ntsr1) that were found in the network ‘carbohydrate metabolism, cell signaling, and nucleic acid metabolism’ (= 4 per group). *< 0.05 and **< 0.01.

Western blot analyses

Oxtr was chosen for further verification of protein expression levels in the rat PFC after 7β-HC treatment as compared with vehicle treatment based on its functional importance in the brain such as cognition, and commercially available antibody. The anti-Oxtr antibody detected two major bands at approximately 35 kDa and 50 kDa in a rat uterus sample, used as a positive control (Fig. 4a). Blots from the PFC showed a major band at approximately 35 kDa and a faint band at approximately 50 kDa (Fig. 4a). These correspond to the deglycosylated form (38 kDa) which is close to the expected theoretical molecular mass of the receptor core, and partial deglycosylated form (48 kDa) of Oxtr (Kojro et al. 1991; Breton et al. 2001), respectively. Densitometric analysis showed significantly reduced Oxtr protein expression in the rat PFC by 33% after 7β-HC treatment, compared with vehicle treated controls (= 0.0144, Fig. 4b).

Figure 4.

(a) Western blot analysis and (b) densitometry graph of Oxtr detected in the homogenates from the prefrontal cortex (PFC) after 1 day post-7β-hydroxycholesterol (7β-HC) treatment as compared with vehicle treatment (= 4 per group). B-actin was loaded as a loading control. *< 0.05. A positive control of Oxtr was performed using homogenates from the uterus tissue of a female rat that had just delivered. (c and d) Light micrographs of Oxtr immunoreactive brain slices in the PFC at 60X (c) and 100X (d) magnifications. Dense immunoreactivity to Oxtr (arrows) was observed in the neuronal cell bodies and punctate profiles in the neuropil. Scale = 50 (c) and 30 μm (d). (e and f) Electron micrographs of Oxtr immunoreactive sections from the PFC, showing immunoreactive products (arrows) in axon terminals. The dendrite (D) forms an asymmetrical synapse (S) with axon terminal (AT) that can be identified by the presence of neurotransmitter containing small round vesicles in the axon terminals. Scale = 0.1 μm.

Immunolocalization of Oxtr in the PFC

Dense immunoreactivity to Oxtr was observed in the neuronal cell bodies and punctate profiles in the neuropil (Fig. 4c and d). Electron microscopy of Oxtr immunostained sections showed labeling in axon terminals (Fig. 4e and f). The latter were identified by the presence of neurotransmitter containing small round vesicles, characteristic of glutamatergic axon terminals (Edwards 1995).

Discussion

This study aimed to elucidate comprehensive gene expression changes in the rat PFC after treatment with oxysterols. The overall results indicate that treatment with these oxysterols exert potent effects on gene expression, compared with their parent compound, cholesterol. Many DEGs with high fold change were found in common after 7β-HC or 7-KC treatment, whereas few DEGs were found after cholesterol treatment. This study focused on gene expression changes in the PFC after treatment with oxysterols that are generated by autoxidation of cholesterol, i.e. 7β-HC or 7-KC, whereas possible effects of oxysterols that are formed by enzymatic action, such as 24S-hydroxycholesterol were not elucidated. 24S-hydroxycholesterol differs from 7β-HC and 7-KC because of the side-chain hydroxylation of cholesterol (Brown and Jessup 2009). Although levels of 24S-hydroxycholesterol, 7β-HC and 7-KC were elevated at 2-weeks post-kainate lesion, the level of 24S-hydroxycholesterol was several folds higher than 7β-HC and 7-KC (He et al. 2006; Kim et al. 2010, 2011). The relatively higher abundance of 24S-hydroxycholesterol makes it unsuitable for direct comparison with 7β-HC and 7-KC in this study. The dose of oxysterols used in this study was calculated based on levels of 7β-HC and 7-KC in the rat hippocampus, after excitotoxic brain injury induced by the glutamate analog, kainate (Kim et al. 2011). It is also similar to the concentration of these oxysterols in the visual cortex of patients with Parkinson's disease (Cheng et al. 2011).

Oxysterols are involved in transcriptional control of lipid metabolism through interaction with liver X receptors (LXRs) (Edwards et al. 2002), and the sterol response element binding protein (SREBP) family of transcription factors (Olsen et al. 2012). Both isoforms of LXR, LXRβ and LXRα, bind to 24S-hydroxycholesterol and 24, 25-epoxycholesterol with high affinity, and LXRs control cellular cholesterol efflux by regulating gene expression of cholesterol transport proteins such as ABCA1 and ABCG1 (Björkhem 2002; Björkhem and Meaney 2004). Oxysterols exert pro-inflammatory effects at transcriptional level through activation of LXR (Vejux et al. 2008; Jusakul et al. 2011). Recently, LXR targets were discovered by microarray analysis in human macrophages (Pehkonen et al. 2012), and two of these targets Rplp1 and Ranbp1 were detected in our microarray analyses, which suggest the involvement of LXR in oxysterol-induced gene expression changes. Oxysterols also bind to insulin-induced gene (Insig) proteins and suppress the activation of a membrane-bound transcriptional factor, SREBP, leading to inhibition of cholesterol synthesis (Radhakrishnan et al. 2007). In addition, oxysterols play a role in transcriptional mechanisms of the JAK2/STAT3 pathway to elevate gene expression (Romeo and Kazlauskas 2008). Changes in gene expression could also be because of epigenetic effects, such as histone modification, DNA methylation, RNA-associated silencing and nucleosome positioning (Portela and Esteller 2010). Oxysterols could affect gene transcription by recruitment of epigenetic modifying proteins such as histone acetyltransferase, histone deacetylase, and chromatin-remodeling factors, together with a transcription factor, to the gene promoter (Romeo and Kazlauskas 2008). Mutagenic and genotoxic effects of oxysterols have also been reported (Jusakul et al. 2011). Oxysterols increase the level of mitochondrial DNA damage resulting in increased production of reactive oxygen species (Gramajo et al. 2010), which could lead to changes in gene expression (Gargiulo et al. 2011).

The pathways that were altered in response to oxysterol treatment in the brain were determined by IPA. Interestingly, many DEGs in the top network of down-regulated genes related to carbohydrate metabolism, cell signaling, and nucleic acid metabolism were found to encode GPCRs. The latter is the largest family of plasma membrane proteins with seven membrane-spanning α-helical domains, and participates in many forms of information processing (Rosenbaum et al. 2009). They are involved in modulation of signaling proteins [e.g. cAMP-dependent protein kinase (PKA) and MAPK] and enzymatic effectors, as well as transcriptional regulation of target genes (Shaywitz and Greenberg 1999; Neves et al. 2002; Rosenbaum et al. 2009). GPCR signaling is regulated by phosphorylation status of the receptor for rapid and precise temporal control, and transcriptional and post-transcriptional mechanisms for long term regulation (Collins et al. 1991; Pierce et al. 2002). Transcriptional down-regulation of GPCRs in the PFC has been reported in the aging human brain, and could be the result of accumulative oxidative damage to DNA, lipids, and proteins (Erraji-Benchekroun et al. 2005).

Other down-regulated genes mapped in the second largest network were related to lipid metabolism, molecular transport, and small molecule biochemistry. Genes in this network were connected to NF-κB and RXR, and one possibility is that oxysterols could induce cellular oxidative stress, which leads to NF-κB- and associated LXR activation (Robbesyn et al. 2004). In addition, the largest network generated from the up-regulated genes associated with cell death, cell morphology, cellular assembly and organization with 26 focus molecules, also demonstrated relationship with NF-κB.

In view of the down-regulation of DEGs as the major trend in gene expression induced by oxysterol treatments and the importance of GPCRs, several genes from the largest network of down-regulated genes were selected for further verification by real-time PCR. Treatment with oxysterol resulted in reduced mRNA expression of Oxtr, Cckar, and Npffr1 while Ntsr1 was only reduced by cholesterol treatment. Cholecystokinin A receptor (Cckar), neuropeptide FF receptor 1 (Npffr1), and neurotensin receptor 1 (Ntsr1) are neuropeptide receptors which are GPCRs. Cckar is one of the two known receptors for the binding of cholecystokinin (Koefoed et al. 2009; Wilson et al. 2012). Cholecystokinin is a neuropeptide transmitter that coexists with dopamine in the mesolimbic system and modulates dopamine release (Hökfelt et al. 1980; Skirboll et al. 1981; Li et al. 1995). It has strong relation to schizophrenia and bipolar affective disorder (Asherson et al. 1998; Christoforou et al. 2007). Npffr1 binds to neuropeptide FF (Bonini et al. 2000; Elshourbagy et al. 2000) and can be activated by neuropeptide VF precursor (Liu et al. 2001). Ntsr1 is suggested to be involved in working memory in both humans (Li et al. 2011) and rodents (Tirado-Santiago et al. 2006). mRNA levels of Ntsr1 and neurotensin receptor 2 (Ntsr2) are markedly reduced in the temporal gyrus of patients with Alzheimer's disease (Gahete et al. 2010).

The reduction in Oxtr expression by oxysterol is a novel observation, and is of interest because of its emerging role in cognition. Oxtr is a receptor for oxytocin which is associated with social cognition (Landgraf and Neumann 2004; Kirsch et al. 2005; Kosfeld et al. 2005; Hollander et al. 2007), autism (Hollander et al. 2007; Kosaka et al. 2012), and anxiety (Heinrichs et al. 2003; Scantamburlo et al. 2007). Decreased Oxtr mRNA expression as a result of hypermethylation of the gene promoter has been found in temporal cortex tissue and peripheral blood lymphocytes of patients with autism spectrum disorders (Gregory et al. 2009). Loss of oxytocin or Oxtr in rodents results in decreased social recognition (Takayanagi et al. 2005; Lee et al. 2008), elevated aggressive behavior (Takayanagi et al. 2005), increased anxiety (Mantella et al. 2003) and reduced reciprocal social interactions (Pobbe et al. 2012). As with its mRNA expression, the deglycosylated form of Oxtr protein was significantly down-regulated in the PFC after 7β-HC treatment. Previous study has shown that glycosylation status at the N-terminal would not affect the activity and characteristics of Oxtr (Kimura et al. 1997). Oxtr was localized to neurons in the PFC, consistent with the results of a previous study (Adan et al. 1995), and electron microscopy showed Oxtr immunoreactivity in axon terminals. One possibility is that increased oxysterols during neuronal injury could result in down-regulation of Oxtr at nerve terminals, thus affecting neuronal plasticity.

In conclusion, the present discovery of oxysterol induced gene expression changes provides insight into possible molecular mechanisms through which increased level of autoxidatively formed oxysterols could exert a pathophysiological effect in the brain. Further study will be required to verify other genes in the PFC identified by the present microarray analyses, as they may be important downstream targets of oxysterols, and to elucidate whether increased oxysterols could contribute to cognitive dysfunction after excitotoxicity.

Acknowledgments

This study was supported by the National Research Foundation of Singapore under its Competitive Research Programme (CRP Award No. NRF-CRP 3-2008-1). All authors have no conflict of interest to declare.

Ancillary