Ingemar Björkhem, Division of Clinical Chemistry, Department of Laboratory Medicine, Karolinska Institutet, Karolinska University Hospital, C1 74, 141 86 Huddinge, Sweden. (fax: +46 8 58581260; e-mail: email@example.com)
Abstract. Shafaati M, Olin M, Båvner A, Pettersson H, Rozell B, Meaney S, Parini P, Björkhem I (Karolinska University Hospital Huddinge, Huddinge, Sweden; Dublin Institute of Technology, Dublin, Ireland). Enhanced production of 24S-hydroxycholesterol is not sufficient to drive liver X receptor target genes in vivo. J Intern Med 2011; 270: 377–387.
Background. Oxysterols such as 24S-hydroxycholesterol (OHC) and 27-OHC are intermediates of cholesterol excretion pathways. In addition, they are putative endogenous agonists of the liver X receptor (LXR) class of nuclear hormone receptors and are thought to be important mediators of cholesterol-dependent gene regulation. 24S-OHC is one of the most efficient endogenous LXR agonists known and is present in the brain and in the circulation at relatively high levels.
Objectives. To explore the regulatory importance of 24S-OHC in vivo.
Design. We developed a transgenic mouse model in which human cholesterol 24-hydroxylase, the enzyme responsible for the formation of 24S-OHC, was expressed under the control of a promoter derived from the β-actin gene.
Results. Both male and female transgenic mice had elevated levels of cerebral, plasma, biliary and faecal 24S-OHC. According to the faecal excretion results, production of 24S-OHC was increased four- to sevenfold. Gene expression profiling revealed that the elevated production of 24S-OHC did not result in the anticipated activation of LXR target genes in the brain or liver.
Conclusion. In spite of the fact that 24S-OHC is a highly effective agonist of LXRs in vitro, it is not a critical activator of target genes to this nuclear receptor in vivo, either in the brain or in the liver.
Elimination of cholesterol requires the formation of bile salts. This process is dependent on sequential hydroxylation of cholesterol by members of the cytochrome P450 superfamily . Although most bile acid synthesis occurs in the liver, it may be initiated throughout the body. There is a continuous flow of mono- and dihydroxylated derivatives of cholesterol – commonly called oxysterols – from the periphery to the liver [2, 3]. Although dozens of oxysterols have been identified in the circulation, 27-hydroxycholesterol (OHC), 24S-OHC and 7α-OHC typically represent at least 85% of the total plasma oxysterol content [4, 5]. In addition to this important role as intermediates in sterol elimination pathways, oxysterols have also been reported to bind to a variety of proteins including at least two members of the nuclear hormone receptor superfamily [6, 7].
The liver X receptors (LXRs) were identified as heterodimeric partners of the retinoid X receptor [8, 9], and two forms have been described, LXRα and LXRβ. Subsequent in vitro studies determined that oxysterols, in particular those with an additional oxygen function on the side chain, bound to and activated these receptors . Numerous nonoxysterol ligands, both natural and synthetic, have been described , and whilst the importance of LXRs for maintenance of whole body and organ-specific lipid balance is well recognized, there is still some controversy concerning the identity of the native endogenous ligands . A recent study using mice in which the genes encoding the enzymes responsible for the formation of 24S-, 25- and 27-OHC were disrupted revealed a blunted response to dietary cholesterol . The corollary of these findings is that enhanced formation of one or more of these oxysterols should be associated with an increase in the associated regulatory response.
Amongst the above side-chain-oxidized oxysterols, 24S-OHC is the most effective ligand for LXRs in in vitro conditions . The current study was designed to investigate the role of this oxysterol as a regulator of LXRs in vivo too, using a mouse model engineered to stably express human cholesterol 24-hydroxylase (CYP46A1) under the control of the β-actin promoter. This resulted in an increase in the content of 24S-OHC in the brain, plasma, bile and faeces. However, little or no change in the mRNA levels of genes expected to be responsive to LXR activation was observed, and there were limited biochemical effects in the transgenic model. The results are discussed in relation to the role of oxysterols as endogenous LXR agonists.
Materials and methods
Construction of human CYP46 overexpressor transgenic mice
Two 1.6-kb fragments (one included an HA-tag sequence) encoding for the human CYP46 cDNA were inserted into the EcoR1 restriction site in a pCAGGS expression vector (kindly provided by Prof Jun-ichi Miyazaki, Tohuku University, Miyagi, Japan). The pCAGGS vector contains the chicken β-actin promoter and rabbit β-globin poly(A) signal permitting ubiquitous overexpression of the human CYP46 cDNA in all tissues. The DNA construct was linearized and purified before microinjection. Primers for the 1.6-kb fragment with HA tag are as follows: forward, 5′ CGC GAA TTC ATG AGC CCC GGG CTG CTG CTG CTC GGC AGC GCC GTC 3′; reverse, 5′ CGC GAA TTC CTA AGC GTA GTC TGG GAC GTC GTA TGG GTA GCA GGG GGG TGG TGG GGG TG 3′. Microinjections were performed at the Karolinska Core Facility for Transgene Technologies.
Mice with an HA-tag sequence were backcrossed with C57Bl/6NCrl mice (Charles River Laboratories, Sulzfeld, Germany) for seven generations and then characterized.
Genotyping of transgenic mice
Animals were tail clipped and genomic DNA was isolated (QIAGEN DNeasy Blood & Tissue kit). The offspring were screened for the presence of transgenes by PCR analyses using specific primers. Positive transgenic mice were identified by tail DNA/PCR genotyping. PCR amplification was performed using primers specific to the human transgenic CYP46 sequence. The following conditions were used for amplification: denaturation at 94 °C for 1 min, annealing at 55 °C for 1 min, elongation at 72 °C for 45 s, repeated 30 times. A forward primer 5′-AAG ATG TAC CGT GCG CTC C-3′ and a reverse primer 5′-GCC TTG GCT TCT AGA ATC TCC-3′ resulted in a 193-bp fragment for the human cDNA transgene sequence.
A method was established to evaluate the degree of overexpression amongst positive transgenic mice using genomic tail DNA. Increasing amounts (1, 5, 10, 20 ng) of genomic DNA were analysed by RT-PCR using TaqMan Gene Expression assay. One primer/probe detected the human cDNA sequence of CYP46A1 (Hs 00198510_m1 CYP 46A1), whereas the other primer/probe detected the mouse genomic Gpbar1 sequence (Mm 00558112_s1 Gpbar1). To estimate the number of copies of hCYP46A1, the slopes obtained by the regression of cycle threshold (Ct) values and the amount of genomic DNA loaded in the reaction for the amplification of hCYP46A1 were divided by the slopes obtained from the amplification of mGpbar1.
All animal experiments received full approval from the local Animal Experimentation Ethics Committee.
Western blot analysis
Microsomes prepared from brains of both transgenic and wild-type mice were subjected to electrophoresis in three different concentrations on a 10% SDS or Bis/Tris polyacrylamide gel and transferred to nitrocellulose membranes (see Fig. 1). The membranes were incubated for 2 h at room temperature in blocking buffer (5% milk in phosphate-buffered saline, 0.05% Tween) followed by incubation overnight at +4 °C with an anti-CYP46 antibody (a generous gift from Prof D. Russell, University of Texas Southwestern Medical Center, Dallas, TX, USA). Goat anti-rabbit IgG coupled with horseradish peroxidase (Pierce, Rockford, IL, USA) was used as a secondary antibody with incubation at room temperature for 2 h. In some experiments, a human-specific rabbit polyclonal IgG antibody against CYP46 (ab36975; Abcam, Cambridge, UK) was used with goat anti-rabbit IgG (Pierce) as a secondary antibody.
The membranes were incubated in Super Signal West Dura Extended Duration Substrate [Prod#34075; Thermo Scientific, Rockford, IL, USA (Pierce)] according to the manufacturer′s instructions. The signal (around 50 kDa) was detected using Universal hood II equipment (Bio-Rad, Hercules, CA, USA).
The results were calculated from the signal of each sample in triplicate as a linear value. It is important to emphasize that the relative specificity of the CYP46 antibody against human CYP46A1 and murine Cyp46a1 is not known. When analysing human and murine brain material in parallel with this antibody, a stronger signal is obtained from the murine brain material.
Brains were removed from wild-type and transgenic mice, fixed in neutral buffered formalin for 24 h and embedded in paraffin according to standard procedures. Paraffin sections were de-waxed and subjected to autoclaving in a 2100 Retriever machine (PickCell Laboratories, Amsterdam, the Netherlands) according to the manufacturer’s instructions using Divadecloaker solution at pH 6. Sections were blocked by 4% (v/v) normal goat serum (NGS) in Tris-buffered saline pH 7.6 (TBS). The primary rabbit anti-CYP46 antibody (a generous gift from Prof D. Russell) was diluted 1 : 4000 in TBS/NGS and sections incubated at 4 °C overnight. To detect the HA tag, we used a rabbit monoclonal anti-HA-tag antibody (C29F4; Cell Signalling Technology, Danvers, MA, USA) diluted ×400 and incubated with the sections overnight. After several rinses in TBS, the sections were exposed to a biotinylated goat anti-rabbit antibody (Dako Cytomation, Glostrup, Denmark) for 1 h at room temperature, rinsed and finally incubated with avidin-horseradish peroxidase complex for 30 min. Binding of the primary antibody was visualized with diaminobenzidine/H2O2, and nuclei were counterstained with haematoxylin.
RNA preparation and real-time RT-PCR
Total RNA was prepared using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. RNA (1 μg) was transcribed into cDNA using a high-capacity cDNA reverse transcription kit (Applied Biosystem, Carlsbad, CA, USA). The cDNA obtained was diluted 1 : 10 in RNase-free H2O. Real-time RT-PCR was performed with 5 μL cDNA and 12.5 μL SYBRGreen Mastermix (Applied Biosystem). The forward and reverse primers are shown in Table 1. All values were normalized to HPRT mRNA concentrations for hepatic analyses.
Table 1. Primer sequences used for mRNA measurements
GGT GAA AAG GAC CTC TCG AAG TG
ATA GTC AAG GGC ATA TCC AAC AAC A
CCG GCA ACA ACA AGA TCT GTG
ATG TAC AGG ATG GCG ATG CA
CTC TGT CTA TGG TTC CCT GGC T
TCC AAT CCT CTT CCC TGC C
CCC AGA GCA AAA AGC GAC TC
GGT CAT CAT CAC TTT GGT CCT TG
GCG TTC TGG AGA CCA TGG A
ACA AAG TTG CTC TGA AAA CAA ATC
Apo E Fwd
GCA GGC GGA GAT CTT CCA
Apo E Rev
CCA CTG GCG ATG CAT GTC
CYP 7A1 Fwd
CCA TGA TGC AAA ACC TCC AAT
CYP 7A1 Rev
ACC CAG ACA GCG CTC TTT GA
CYP 7B1 Fwd
CCT CTT TCC TCC ACT CAT ACA CAA
CYP 7B1 Rev
GAA CCG ATC GAA CCT AAA TTC CTT
CYP 27-OH Fwd
GCC TTG CAC AAG GAA GTG ACT
CYP 27-OH Rev
CGC AGG GTC TCC TTA ATC ACA
Applied Biosystem Gene Expression assay
Applied Biosystem Gene Expression assay
Applied Biosystem Gene Expression assay
Characterization of Cyp46Hatg seventh generation mice
Cyp46Hatg heterozygous male (n = 5) and female (n = 5) mice were housed separately according to gender and group with a regular 12-h light/12-h dark cycle and access to food and water ad libitum. We used C57/B6 N male (n = 5) and female (n = 5) (Charles River) mice as controls. The animals were killed by CO2 gas inhalation. The brain and liver were removed and plasma was collected by cardiac puncture. Tissues were snap frozen in liquid nitrogen and stored at −70 °C.
Lipid extraction and analysis
Brain was extracted according to the method of Folch with minor modifications. Approximately 10 mg brain tissue was added to 1 mL homogenization buffer (5 mmol L−1 EDTA, 50 μg mL−1 butylated hydroxytoluene in phosphate-buffered saline, pH 7.4) in a clean glass tube and the tissue was disrupted using a polytron homogenizer. An aliquot of 3 mL chloroform/methanol (2 : 1, v : v) (Folch) was added to the homogenate, and the samples were mixed by moderate shaking for 24 h at 4 °C.
Samples were centrifuged at 5000 g for 5 min to permit phase separation and the lower organic phase was removed to a new tube. The aqueous phase was re-extracted once more as described above and the organic phases pooled. Extracts were dried under argon and re-dissolved in Folch and stored at −20 °C until required. Sterols were analysed by gas chromatography–mass spectrometry using deuterium-labelled internal standards as previously described [12, 13].
Analysis of bile acids
Bile acids were extracted after hydrolysis and quantified by isotope dilution–mass spectrometry as described previously .
Gene expression data are expressed as mean ± range as described by Livak et al. . Sterol determinations are expressed as mean ± standard error of the mean (SEM). Statistical comparisons were performed using the unpaired Students t-test. In the statistical evaluation of the different sterols, we used the two-tailed Student’s t-test to evaluate the significance of differences.
Generation of mice overexpressing human CYP46A1
Under normal conditions, expression of murine Cyp46a1 is restricted to neurons of the central nervous system [16, 17], and there is a continuous flux of 24S-OHC from the brain to the liver. To enhance the systemic production of 24S-OHC, we created a mouse model in which the cDNA coding for the human CYP46A1 gene was expressed under the control of a hybrid β-actin promoter identical to that previously used for the development of mice with overexpression of sterol 27-hydroxylase (CYP27A1) . The general structure of this construct is shown in Fig. 1(a). Two mouse lines were generated: one in which the construct contained the native human CYP46A1 (C46-Native) and another in which an in-frame C-terminal HA tag was fused to CYP46A1 (C46-HA). Transgenic mice were identified using PCR with primers directed against the human CYP46A1 cDNA (Fig. 1b). Estimation of the copy number of the transgene using PCR revealed that the C46-HA line had considerably higher expression of the transgene than C46-Native. The C46-HA line was selected for further study and backcrossed with C57/Bl mice for seven generations before molecular and biochemical characterization. There was no obvious difference in appearance, body weight, male/female ratio or fecundity between C46-HA mice and nontransgenic littermates, and histology revealed no abnormalities in the brain or liver. Levels of plasma cholesterol and patterns of bile acids were not significantly different from those of the wild-type mice (results not shown).
Expression of human CYP46A1 in C46-HA mice
Real-time quantitative PCR revealed significant expression of the human CYP46A1 transgene in several organs, including brain, eye, liver, lung, kidney, ovary and testis (Fig. 1c). The levels shown in the figure are related to the housekeeping gene HPRT. Because of the fact that the levels of this gene vary in the different organs, it is not possible to estimate the absolute levels of the CYP46A1 transgene and somewhat different results were obtained when cyclophilin was used as the housekeeping gene.
Western immunoblotting using both pan-reactive and human-specific antibodies revealed that significant levels of human CYP46A1 protein were present only in the brain, eye and testis and there was no correlation between mRNA and protein expression. As shown in Fig. S1, there was a low but significant level of endogenous Cyp46a1 protein in the testis of the wild-type mouse but no such protein in the eye. In the C46-HA mouse, however, significant levels of the protein were found in the eye and increased levels in the testis.
The concentration of CYP46A1 protein was considerably higher in the brain than in the eye and testis, and the presence of the protein in these organs was also confirmed by an antibody directed against the HA tag. Assuming a similar affinity of the antibody against the murine and human enzyme, the level of protein could be estimated to be more than 10-fold higher in the brain than in the other organs. Because of the unknown specificity of the antibody, however, the level could not be quantified precisely. Figure 1(d) shows Western blotting analysis of microsomal brain protein from wild-type and C46-HA mice using an antibody against both the human and murine protein. A significantly higher signal was obtained in the analysis of the material from the C46-HA mouse than from the wild-type mouse. Again, because of the unknown relative antibody specificity, the level could not be accurately quantified. Figure 1(e) shows immunohistochemical analysis of the brain Purkinje cells using an antibody against both murine (cyp46a1) and human (CYP46A1) protein. The pattern was identical in the transgenic and control mice with a tendency to stronger staining in the transgenic animals. In both cases, expression was only observed in neuronal cells with no significant expression in the glial cells. A similar pattern was also observed in the cortex. Because of the high background staining, it was not possible to use the human-specific antibody to probe the localization of the human enzyme. However, the neuronal localization of the protein was confirmed using anti-HA antibody against the HA tag of the transgene protein (Fig. 1e). The localization of the HA-tagged CYP46A1 protein was found to be the same as that of the murine endogenous cyp46a1 protein.
A clear signal was obtained in the analysis of brain microsomes from a C46-HA mouse but not in the corresponding analysis of material from a wild-type mouse in Western blotting experiments with the anti-HA antibody (Fig. S2). A positive signal was also obtained in Western blotting experiments with lung and kidney from the transgenic but not from the wild-type mice (results not shown).
24S-OHC levels in C46-HA mice
24S-OHC is known to exit the brain, pass through the circulation to the liver where approximately 50% is conjugated and the remainder excreted into the faeces as free sterol . In C46-HA mice, there was a significant increase (approximately twofold) in the brain content of 24S-OHC (P < 0.01; Fig. 2). Plasma levels were also significantly elevated (approximately fivefold; Fig. 2), and there was a significant positive correlation between brain and plasma levels of 24S-OHC (results not shown). There was, however, little or no increase in the 24S-OHC content of the liver, and the levels in this organ were less than 1% of those in the brain. The biliary content of 24S-OHC was increased four- to fivefold in C46-HA mice (Fig. 2). Measurement of the daily excretion of free unconjugated 24S-OHC in faeces revealed a significant four- to sevenfold increase in the transgenic mice, despite the normal content in the liver (Fig. 2). The levels of 24S-OHC in the testis and eye were <3% of those in the brain (results not shown). In addition to the increased levels of 24S-OHC, we noted a modest but significant (P < 0.05) increase in 27-OHC levels in the brain of male C46-HA mice and in the plasma of both male and female C46-HA mice (P < 0.05; Fig. 3). In the liver, there was a slight but significant (P < 0.05) decrease in the levels of 27-OHC.
Expression of genes involved in cerebral and hepatic sterol homeostasis
As 24S-OHC is an endogenous agonist of LXRα and LXRβ, we anticipated that increasing its levels would lead to a predictable response in the mRNA levels of known LXR target genes in organs either containing high levels of 24S-OHC (brain) or exposed to a high flux of 24S-OHC (liver). We found that there were relatively modest effects on gene expression in both organs (Fig. 4). Moreover, several well-recognized LXR target genes did not appear to be regulated as expected. There was a significant decrease in Cyp7a1 mRNA in the liver of male C46-HA mice instead of the expected increase. We observed a notable sexual dimorphism in some of the hepatic genes. The expression of Cyp7b1 was significantly increased in liver from male mice but decreased in females. There was also a significantly increased level of Hmgcs in liver from male but not female mice. In addition, there was a slight but significant increase in Apoe mRNA levels in the liver in females but not in males.
The changes in Cyp7b1 and Apoe in liver from female mice are consistent with an effect of LXRs, whereas the other changes are not. A sexually dimorphic response was also seen in the mRNA level of Fas which was downregulated in males (instead of the expected upregulation). There was no obvious gender difference in the expression of the different genes in the brain. There was a tendency towards decreased brain mRNA levels of Cyp27a1 and increased levels of Cyp39a1 in female mice.
The possibility that the expression of LXRα (the predominating form of this nuclear receptor in the liver) was affected by the overexpression of CYP46A1 was excluded. Thus, the LXRα mRNA levels in the C46-HA mice were not different from those in the wild-type mice (P > 0.05; results not shown).
Brain sterol homeostasis in C46-HA mice
Next, we measured the concentration of seven cholesterol precursors as well as cholesterol itself in the brain of C46-HA mice using a recently published sterol-profiling method . These studies revealed significant increases in some but not all of the sterol precursors (Fig. 5). The magnitude of this increase was approximately 1.5- to 1.7-fold, and there was a significant positive correlation (r = 0.85; P < 0.05) between brain lathosterol content (an index of the cholesterol synthesis rate) and brain content of 24S-OHC in C46-HA mice (Fig. 6). The latter correlation was higher in the overexpressed mice than in the controls. The possibility that LXRβ (the predominant form of the nuclear receptor in the brain) was affected by the overexpression of CYP46A1 was excluded. Thus, the LXRβ mRNA levels in the C46-HA mice were not different from those in the wild-type animals (P > 0.05; results not shown).
The results of the present study add to previous findings that, in the presence of enhanced oxysterol production, a systemic increase in levels of 27-OHC did not have a major effect on overall cholesterol homeostasis. Markedly increased levels of 27-OHC as a consequence of overexpression of CYP27A1  or knockout of the gene coding for the metabolizing enzyme Cyp7b1  also had little or no effect on cholesterol turnover. In mice with knockout of Cyp27a1, there is an upregulation of cholesterol and bile acid synthesis because of the lack of bile acids . This upregulation is a consequence of the reduced formation of bile acids and there is a return to normal when the mice are treated with bile acids .
24S-OHC is a more efficient activator of LXR than 27-OHC in vitro , and we considered it would be important to determine whether an increased production of 24S-OHC is able to affect LXR-regulated genes in vivo.
Expression of human CYP46A1 under the control of a modified β-actin promoter led to increased 24S-OHC content in brain, and plasma, but not in liver. Faecal excretion of the free form of 24S-OHC was increased four- to sevenfold. Immunohistochemical analysis revealed that transgene expression in the brain was restricted to neurons. Taken together, these data indicate that our model represents an enhanced version of the normal situation in which the majority of circulating 24S-OHC is produced by the brain and eliminated via hepatic metabolism . We thus considered this model suitable for investigating a possible role of 24S-OHC as a critical endogenous ligand for LXR.
It is interesting that, after the brain, the highest levels of CYP46A1 protein and 24S-OHC were observed in the eye and in the testis. Similar to the brain, these two organs are protected by barriers: the blood–retina and blood–testis barriers. The contribution of these two organs to the production of 24S-OHC in the transgenic mice, however, must be very limited in relation to the production by the brain. The very marked difference between CYP46A1 mRNA expression and levels of CYP46A1 protein and 24S-OHC in all organs other than the brain suggests that organ-specific factors are critical for the translation and/or stability of the protein.
LXR is recognized as a critical regulator of lipid turnover, and pharmacological activation results in an increase in the expression of genes including Cyp7a1, Srebp1c, Abca1, Abcg5 and Abcg8  and a decrease in Cyp7b1 . Enhanced production of one of the most potent putative endogenous ligands of LXR would be expected to lead to an analogous transcriptional response. However, contrary to these expectations, only modest changes were observed. Hepatic expression of Cyp7a1 was reduced rather than increased, whereas Cyp7b1 was increased in male mice and decreased in female mice. The similarity in hepatic levels of 24S-OHC in C46-HA and wild-type mice provides some explanation for the lack of induction, although it does not explain the apparent decrease in hepatic 7α-hydroxylation capacity, a key step in the formation of bile acids. The decreased Cyp7a1 mRNA levels in the liver of transgenic male mice are not likely to be associated with major effects on cholesterol homeostasis as plasma cholesterol levels and bile acid patterns were not significantly different from those of control mice.
The failure of C46-HA mice to accumulate 24S-OHC in the liver, despite high levels in both pre- and posthepatic compartments, is puzzling. Recent evidence from Temel et al. using a biliary diversion model suggests that sterols do not have to undergo biliary secretion to reach the intestine . Free, unmodified 24S-OHC may thus pass directly to the faeces, whilst 24S-OHC entering the liver will be subjected to sulphation, glucuronidation, ω-oxidation and conversion into bile acids. The increased levels of 24S-OHC in bile of the transgenic mice demonstrate that the liver is exposed to significantly increased levels of 24S-OHC. It is possible that the ABCG5 and ABCG8 transporters (ATP-binding cassette subfamily G member 5 and 8, respectively) may be important for the elimination of 24S-OHC from the liver, in spite of our failure to demonstrate an increase in the corresponding mRNA levels. Very recent data from Xu et al.  indicate that 25-OHC-3-sulphate antagonizes LXR and SREBP signalling in rat hepatocytes, apparently by decreasing nuclear protein levels. As 24S-OHC is known to be sulphated to some extent during hepatic metabolism, the possibility must be considered that high levels of 24S-OHC sulphate may blunt any stimulatory effect mediated by the free sterol. Theoretically, the effects of the overexpression on some of the hepatic LXR target genes that were opposite to those expected may have been because of an inhibitory effect on the signalling by a metabolite of 24S-OHC, e.g. sulphate. As LXR mRNA levels in the liver were not affected by the overexpression, it is evident that there were no effects on the signalling machinery at the transcriptional level.
In principle, one would not expect any effects of 24S-OHC on the target genes in the liver in the absence of significantly increased levels of oxysterol. From this point of view, the failure to obtain the expected response of LXR target genes in the transgenic mice is not surprising. If the liver is protected from an oxysterol-induced signalling by increased metabolism and possibly also by an inhibitory effect of one of its metabolites, it is not consistent with the notion that 24S-OHC is of major importance as a regulator of cholesterol homeostasis in the liver.
In contrast to the situation in the liver, the levels of 24S-OHC were significantly increased in the brain in the transgenic mice. Thus, the model seems to be more appropriate for evaluating the effects on brain than on liver. However, once again the transcriptional response was very limited and there were no significant changes in the LXR-regulated genes. Expression of Cyp27a1 was slightly reduced, whereas there was no effect on mRNA levels of Cyp7b1. In spite of this, there was an increase in brain 27-OHC content. In preliminary experiments, evidence has been obtained that this may be because of an effect of 24S-OHC on the metabolism of 27-OHC (unpublished data). The lack of effect of overexpression on the levels of Apoe mRNA in the brain is of interest in relation to the demonstration by Abildayeva et al.  that 24S-OHC induces Apoe transcription, protein synthesis and secretion in cultured astrocytes via an LXR-controlled mechanism. If this mechanism is important in vivo, an upregulation of apoe mRNA levels would have been expected.
Studies of the Cyp46a1 null mouse revealed that when brain cholesterol elimination was reduced, there was a compensatory decrease in cholesterol synthesis. Overexpression of CYP46A1 would then be expected to lead to the opposite effect, i.e. an increased consumption of cholesterol with a compensatory increase in the rate of cholesterol synthesis. In accordance with this, we found significantly increased levels of several cholesterol precursors in the brain of C46-HA mice. There was a slight increase in the cholesterol content in the brain of overexpressed male mice. It is interesting that there was a significant correlation between levels of 24S-OHC and levels of lathosterol both in wild-type and transgenic mice, suggesting a close coupling between cholesterol elimination by the Cyp46a1 pathway and cholesterol synthesis.
The present results are consistent with those recently reported by Hudry et al.  who used a viral approach to specifically overexpress CYP46A1 locally in cortex and hippocampus of female APP23 mice. No changes were found in cortical or hippocampal cholesterol levels as a consequence of the overexpression but a slight upregulation of Hmgcr was observed. There was no upregulation of LXR target genes in these regions of the brain.
The sexual dimorphism observed in the response of some of the hepatic genes to the expression of C46-HA is intriguing but cannot be explained at the present time. Gender differences are known to be present both in steroid-synthesizing and steroid-metabolizing systems in rodents. Theoretically, a metabolite could have an effect on LXR signalling that is opposite to the effect of the unmetabolized steroid. At present, we cannot exclude the possibility that some of the gender differences in response to the overexpression may reflect gender differences in metabolism. Further work is needed to clarify this.
To summarize, the current data do not support the contention that side-chain oxysterols are major endogenous ligands of the LXR class of nuclear receptors. We failed to find a general activation of LXR signalling genes in the brain of C46-HA mice, in spite of increased levels of 24S-OHC. As the steady-state levels of 24S-OHC in the liver were not increased in C46-HA mice, it is difficult to draw firm conclusions about the potential of oxysterols to activate LXR target genes in this organ. If the low levels of 24S-OHC in the liver are because of an efficient metabolism, the possibility must be considered that this represents protection against potential adverse effects on cholesterol homeostasis. In any case, and even if side-chain-oxidized oxysterols are of regulatory importance under specific conditions , it is evident that they are not the master regulators as previously suggested by Kandutsch et al. .
This work was supported by a Pfizer – Karolinska Institutet collaborative grant, the Swedish Science Council, Swedish Brain Power, the Söderberg Foundation and the Swedish Alzheimer’s Foundation.