Address correspondence and reprint requests to Kazumitsu Ueda, Institute for integrated Cell-Material Sciences (iCeMS), Kyoto University, Kitashirakawa-oiwake-cho, Kyoto 606-8502, Japan. E-mail: email@example.com
High cholesterol turnover catalyzed by cholesterol 24-hydroxylase is essential for neural functions, especially learning. Because 24(S)-hydroxycholesterol (24-OHC), produced by 24-hydroxylase, induces apoptosis of neuronal cells, it is vital to eliminate it rapidly from cells. Here, using differentiated SH-SY5Y neuron-like cells as a model, we examined whether 24-OHC is actively eliminated via transporters induced by its accumulation. The expression of ABCA1 and ABCG1 was induced by 24-OHC, as well as TO901317 and retinoic acid, which are ligands of the nuclear receptors liver X receptor/retinoid X receptor (LXR/RXR). When the expression of ABCA1 and ABCG1 was induced, 24-OHC efflux was stimulated in the presence of high-density lipoprotein (HDL), whereas apolipoprotein A-I was not an efficient acceptor. The efflux was suppressed by the addition of siRNA against ABCA1, but not by ABCG1 siRNA. To confirm the role of each transporter, we analyzed human embryonic kidney 293 cells stably expressing human ABCA1 or ABCG1; we clearly observed 24-OHC efflux in the presence of HDL, whereas efflux in the presence of apolipoprotein A-I was marginal. Furthermore, the treatment of primary cerebral neurons with LXR/RXR ligands suppressed the toxicity of 24-OHC. These results suggest that ABCA1 actively eliminates 24-OHC in the presence of HDL as a lipid acceptor and protects neuronal cells.
The brain contains ~ 25% of the human body's total cholesterol, despite occupying only 2% of body mass. Some fraction of cholesterol is actively converted to 24(S)-hydroxycholesterol (24-OHC) by cholesterol 24-hydroxylase, a cytochrome P450 (CYP46A1) highly expressed in a subset of neurons in the brain, and subsequently eliminated from brain tissue (Bjorkhem et al. 1997); about 0.02% of brain cholesterol in humans and 0.4% in mouse turns over each day (Dietschy and Turley 2004). Disruption of the mouse CYP46A1 gene reduced the synthesis of new cholesterol in the brain by ~ 40%, indicating at least 40% of cholesterol turnover in the brain is dependent on the conversion into 24-OHC (Lund et al. 2003). This knockout mouse exhibited severe deficiencies in spatial, associative, and motor learning, as well as in hippocampal long-term potentiation (Kotti et al. 2006), suggesting that cholesterol turnover via 24-hydroxylase is essential for brain functions, especially learning.
Because hydroxylation of the side chain of cholesterol accelerates transfer across the lipid bilayer more than 1000 times relative to unhydroxylated cholesterol (Meaney et al. 2002), it is generally accepted that after its production in the brain, 24-OHC gains access to the circulation by spontaneous diffusion across cellular membranes and the blood–brain barrier (Russell et al. 2009); however, some studies have demonstrated transporter-mediated oxysterol efflux (Tam et al. 2006; Ohtsuki et al. 2007; Terasaka et al. 2007). Tam et al. reported that ABCA1 mediates the efflux of 25-hydroxycholesterol from human embryonic kidney (HEK) cells expressing ABCA1 as well as from mouse primary embryonic fibroblasts (Tam et al. 2006); Terasaka et al. reported that ABCG1 promotes the efflux of 7-ketocholesterol and other oxysterols from macrophages onto high-density lipoprotein (HDL), and protects these cells against apoptosis induced by oxidized low-density lipoprotein (Terasaka et al. 2007).
Although most of hydrophobic and amphipathic compounds pass freely through the lipid bilayer, some ABC proteins, such as ABCB1 (MDR1) and ABCG2, actively transport hydrophobic and amphipathic compounds, thereby playing important roles in protecting our body by expelling such compounds into the intestinal lumen, into the bile from the liver, and into the urine from the kidney (Ueda 2011). ABCA1 and ABCG1 are involved in the efflux of cholesterol from cells. ABCA1 mediates the efflux of cholesterol to lipid-free apolipoprotein A-I (apoA-I), which serves as a lipid acceptor in the serum (Wang et al. 2000; Tanaka et al. 2001); in the case of ABCG1, HDL acts as the acceptor (Wang et al. 2004; Vaughan and Oram 2005; Kobayashi et al. 2006). Both ABCA1 and ABCG1 are expressed in the brain (Fukumoto et al. 2002; Koldamova et al. 2003; Tachikawa et al. 2005; Tarr and Edwards 2008). ApoE-containing lipoproteins, HDL-like particles, function in delivery of cholesterol from astrocytes to neurons in brain, while HDL containing apoA-I functions in reverse cholesterol transport in peripheral tissues. ABCA1 and ABCG1 expressed in astrocytes are involved in the formation of apoE-containing lipoproteins (Hirsch-Reinshagen et al. 2004; Karten et al. 2006), which stimulate axonal extension of neurons (Matsuo et al. 2012). The importance of ABCA1 and ABCG1 for the formation of apoE-containing lipoproteins is shown in results that apoE and cholesterol levels decrease in cerebrospinal fluid (CSF) of Abca1 knockout mice (Wahrle et al. 2004) and cholesterol level in astrocyte increases in Abcg1 knockout mice (Wang et al. 2008). Both proteins are also expressed in neurons, where their physiological roles remain unclear. We examined the possible roles of ABCA1 and ABCG1 in 24-OHC efflux from neuronal cells, using SH-SY5Y cells as a model. We found that both ABC proteins play roles in 24-OHC efflux.
Materials and methods
Anti-ABCA1 monoclonal antibody KM3110 was generated against the C-terminal 20 amino acids of ABCA1 in mice (Munehira et al. 2004). Rabbit polyclonal anti-ABCG1 antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse monoclonal anti-β-actin antibody was purchased from Sigma-Aldrich (St. Louis, MO, USA). 24-OHC was purchased from BIOMOL Research Laboratories (Plymouth Meeting, PA, USA). [22,23-3H] 24-OHC (specific activity: 50–60 Ci/mmol) was purchased from American Radiolabeled Chemicals, Inc. (St. Louis, MO, USA). TO901317 was purchased from ALEXIS Biochemicals (San Diego, CA, USA). 9-cis retinoic acid (RA) was purchased from Sigma. HDL from human plasma was purchased from Athens Research & Technology (Athens, GA, USA). ApoA-I from human plasma was purchased from Calbiochem (San Diego, CA, USA). Human recombinant apoE3 and apoE4 were purchased from Wako Pure Chemical Industries (Osaka, Japan). Rat cerebral neuron was purchased from DS Pharma Medical (Osaka, Japan). LDH assay kit was purchased from Promega (Madison, WI, USA).
All cell types were cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Carlsbad, CA, USA) supplemented with 100 U/mL benzylpenicillin potassium, 100 μg/mL streptomycin sulfate, and 10% fetal bovine serum (FBS). Cells were grown at 37°C in a humidified atmosphere containing 5% CO2. For differentiation of SH-SY5Y cells (Ammer and Schulz 1994), 10 μM all-trans retinoic acid was added to the culture medium a day after seeding, and cells were cultured for 5 days. Primary cerebral neurons were isolated from rat by using neuron isolation kit (DS Pharma Biomedical) and then plated on Poly-d-Lysine-coated plate at 2 × 105 cells/mL. The cells were cultured in Neurobasal (Invitrogen) supplemented with 200 mM l-glutamine and 2% B27 supplement (Invitrogen). Cells were cultured at 37°C in a humidified atmosphere containing 5% CO2.
Treatment with LXR and RXR ligands
After differentiation, SH-SY5Y cells were incubated in serum-free media in the presence of TO901317 (5 μM), a synthetic ligand of Liver X receptor (LXR), RA (5 μM), a ligand of retinoid X receptor (RXR), or the indicated concentration of 24-OHC for 16 h at 37°C. After cultured at least for 6 days, cells were treated with or without TO901317 (5 μM) and RA (5 μM) for 16 h. The cells were then treated with indicated concentrations of 24-OHC for 24 h with or without TO901317 and RA. For the determination of cell viability, LDH release was measured by using LDH assay kit (Promega).
Cells were washed with phosphate-buffered saline (PBS) and lysed in lysis buffer (20 mM Tris-HCl, pH 7.4, 1 mM EDTA, 10% glycerol and 1% Triton X-100) containing protease inhibitors (p-amidinophenyl)methanesulfonyl fluoride, 100 μg/mL; leupeptin, 2 μg/mL; and aprotinin, 2 μg/mL). Electrophoresis was performed using sodium dodecyl sulfate-polyacrylamide gradient gel (4–12%) (Invitrogen).
[3H]24-OHC efflux from differentiated SH-SY5Y cells
SH-SY5Y cells were first incubated with all-trans retinoic acid (10 μM) for 3 days. Then, SH-SY5Y cells were loaded with [3H] 24-OHC (0.5 μCi/mL) and unlabeled 24-OHC (10 μM) in DMEM containing 10% FBS and all-trans retinoic acid (10 μM) for 24 h at 37°C. Cells were incubated with DMEM containing 0.1% bovine serum albumin (BSA) for 1 h to eliminate [3H]24-OHC non-specifically attached to the cell surface and dishes. Cells were washed once with PBS containing 0.1% BSA and once with PBS alone. Cells were then incubated in serum-free DMEM containing all-trans retinoic acid (10 μM) with or without TO901317 (5 μM) and RA (5 μM) for 16 h to activate LXR/RXR pathway and induce the expression of ABCA1 and ABCG1. Next, cells were incubated for 4 h at 37°C with DMEM containing 0.02% BSA with or without HDL (50 μg/mL), apoA-I (10 μg/mL), apoE3 or apoE4 (10 μg/mL). After 4 h incubation, the medium was collected and centrifuged at 3000 g for 3 min to remove cell debris. Cells were washed twice with ice-cold PBS containing 0.1% BSA, once with PBS alone, and lipids were extracted with 3 : 2 (v/v) hexane : isopropanol. Sum of radioactivity in cells and medium of each experiment was shown in Tables S1–S4.
Reduction in ABCA1 or ABCG1 expression by RNA silencing
Control small interfering RNA (siRNA; Stealth RNAi Negative Control Medium GC Duplex #2); two siRNAs targeting human ABCA1, ABCA1#1 (ABCA1-HSS100028), and ABCA1#2 (ABCA1-HSS-100029); and two siRNAs targeting human ABCG1, ABCG1#1 (ABCG1-HSS145233), and ABCG1#2 (ABCA1-HSS-190466) were purchased from Invitrogen. SH-SY5Y cells were plated in 24-well plates at 1 × 105 cells/well, incubated for 24 h, and transfected with 10 nM siRNAs using Lipofectamine RNAiMAX (Invitrogen). Briefly, siRNA and RNAiMAX, separately diluted in OptiMEM, were mixed and incubated for 10 min at 26°C. These mixtures were added to medium containing 10% FBS and 10 μM all-trans retinoic acid, and cultured for 3 days while the cells differentiated; 24-OHC release assays were performed thereafter.
[3H] 24-OHC and cholesterol efflux from HEK293 cells, stably expressing ABCA1 or ABCG1
Human embryonic kidney (HEK293) cells, stably expressing ABCA1, ABCA1 MM (K939M and K1952M) mutant (Nagao et al. 2009), ABCG1, or ABCG1 KM (K120M) mutant (Kobayashi et al. 2006), were plated in poly-d-lysine-coated 24-well plates at 1 × 105 cells/well. Cells were loaded with [3H] 24-OHC (0.5 μCi/mL) and unlabeled 24-OHC (10 μM), or [3H]cholesterol (1 μCi/mL) and unlabeled cholesterol (10 μM), in DMEM containing 10% FBS for 24 h at 37°C. Next, cells were incubated with DMEM containing 0.1% BSA for 1 h, cells washed once with PBS containing 0.1% BSA, and once with PBS alone. HEK/ABCA1 cells were incubated for 24 h at 37°C with DMEM containing 0.02% BSA with HDL (50 μg/mL) or apoA-I (10 μg/mL). HEK/ABCG1 cells were incubated for 4 h at 37°C with DMEM containing 0.02% BSA with or without HDL (50 μg/mL). Medium was collected and centrifuged at 3000 g for 3 min to remove cell debris. The cells remaining on the dish were washed twice with ice-cold PBS containing 0.1% BSA, then with PBS, and lysed in 3 : 2 (v/v) hexane : isopropanol. Sum of radioactivity in cells and medium of each experiment was shown in Tables S5–S8.
3H-compounds in cells and medium were analyzed using a Radio-HPLC system [Radio detector: Radiomatic 625TR (Perkin Elmer, Waltham, MA, USA); HPLC: Agilent 1100 series (Agilent Technologies, Santa Clara, CA, USA)]. The medium was extracted with 2 × volume of organic solvent [1 : 1 (v/v) acetonitrile : methanol]. The cells were burst by addition of distilled water, and extracted with the 2 × volume of organic solvent. Supernatants were analyzed using the Radio-HPLC with a reversed-phase column (YMC-Pack, ODS-AM-302, 150 × 4.6 mm I.D., S-5 μm, 120 A): solvent, 15 : 85 (v/v) 0.1% formic acid : acetonitrile; flow rate, 1 mL/min.
Percent efflux was calculated by dividing the radioactivity in the medium by the sum of the radioactivity in the medium and cell lysate. Unless otherwise noted, all data are presented as the mean ± SD. Experiments were done at least twice independently. The statistical significance of differences between mean values was analyzed using the non-paired t-test. Multiple comparisons were performed using the Dunnet test following anova. A value of p < 0.05 was considered statistically significant.
Induction of ABCA1 and ABCG1 expression in differentiated SH-SY5Y cells by 24-OHC
We used SH-SY5Y cells as a model to analyze the mechanism of 24-OHC elimination from neuron-like cells. SH-SY5Y cell line was established from human metastatic neuroblastoma tissue (Biedler et al. 1978) and differentiates into neuron-like cells by the treatment with all-trans-retinoic acid (Adem et al. 1987; Yu et al. 1988; Ammer and Schulz 1994). Differentiated SH-SY5Y cells have been widely used to study neuronal functions.
24-OHC, the major cholesterol metabolite in brain, is an endogenous ligand of the nuclear receptor LXR, and induces ABCA1 and ABCG1 gene transcription in macrophages and neuronal cells (Repa et al. 2000; Bjorkhem and Diczfalusy 2004). However, it was reported that ABCG1 is unresponsive to 24-OHC in undifferentiated SH-SY5Y cells (Abildayeva et al. 2006). Therefore, we first examined whether 24-OHC induces expression of ABCA1 and ABCG1 in SH-SY5Y cells. ABCA1 and ABCG1 protein were barely detectable in the absence of 24-OHC, but significantly increased in a concentration-dependent manner upon addition of 24-OHC. The effect of 24-OHC at 10 μM was comparable to that of the synthetic ligands TO901317 (TO), a ligand of LXR, and RA, a ligand of RXR (Fig. 1). However, their expression levels were much lower than those in differentiated human macrophage cell line THP-1 (data not shown). Therefore, we next examined their expression in differentiated SH-SY5Y.
SH-SY5Y cells were differentiated with all-trans retinoic acid for three days, whereupon 24-OHC was added to the medium for 16 h. ABCA1 and ABCG1 protein were faintly but significantly expressed in differentiated SH-SY5Y even in the absence of 24-OHC, and strongly increased in a concentration-dependent manner upon addition of 24-OHC, and the effect of 24-OHC at 10 μM was comparable to that of TO + RA (Fig. 1). These results suggest that ABCA1 and ABCG1 genes in SH-SY5Y cells well respond to LXR ligands after differentiation, and that exogenously added 24-OHC induces LXR-dependent genes as efficiently as the synthetic ligands.
HDL-dependent efflux of 24-OHC from differentiated SH-SY5Y cells
It has been reported that 25-OHC is transported from J774 mouse macrophage cells, mainly by ABCA1, in the presence of apoA-I or HDL (Tam et al. 2006). We examined whether 24-OHC was also transported from differentiated SH-SY5Y cells in the presence of apoA-I or HDL. SH-SY5Y cells were first differentiated, and expression of ABCA1 and ABCG1 was induced with TO + RA. In the absence of TO + RA treatment, ~ 5% of 24-OHC emerged into the medium within 4 h; no significant apoA-I-dependent 24-OHC efflux was observed (Fig. 2a). After treatment with TO + RA, a slight apoA-I-dependent 24-OHC efflux was observed, while no significant increase was observed by the TO + RA treatment.
We also examined 24-OHC efflux in the presence of apoE3 and the isoform apoE4 (Fig. 2b). ABCA1 mediates cholesterol efflux in the presence of apoE as efficiently as apoA-I (Remaley et al. 2001). ApoE is the major apolipoprotein in the brain and apoE isoforms and levels strongly influence Alzheimer's disease pathology and risk (Roses and Saunders 1994). However, we observed no significant 24-OHC efflux in the presence of either the apoE3 or apoE4 variant (Fig. 2b). Because 25-OHC efflux from J774 cells has been observed in the presence of HDL as well as apoA-I (Tam et al. 2006), we examined efflux in the presence of HDL; we observed significant 24-OHC efflux even without TO + RA treatment, and this efflux was further stimulated by treatment (Fig. 2c).
Efflux of 24-OHC, but not its metabolites, from differentiated SH-SY5Y cells
The majority of 24-OHC in the brain is present in non-ester form (Lutjohann et al. 1996), but the sulfate ester can also be detected (Prasad et al. 1984). To determine whether 24-OHC is metabolized in SH-SY5Y cells, and which form of 24-OHC is transported from cells, we performed Radio-HPLC analysis. The main radioactive peak in medium and in cell lysate after the efflux assay was detected at the same position (14.8 min) as the standard [3H]-24-OHC. 24-OHC sulfate, whose retention time is shorter than that of 24-OHC (Bjorkhem et al. 2001), was not detected (Fig. 3). These results suggest that 24-OHC itself is transported from differentiated SH-SY5Y cells in the presence of HDL.
Involvement of ABCA1 in HDL-dependent 24-OHC efflux from SH-SY5Y cells
Cholesterol efflux by ABCA1 is apoA-I–dependent (Wang et al. 2000; Tanaka et al. 2001), whereas efflux by ABCG1 is HDL–dependent (Wang et al. 2004; Vaughan and Oram 2005; Kobayashi et al. 2006). To determine whether 24-OHC efflux from differentiated SH-SY5Y cells in the presence of HDL is mediated by ABCG1, we suppressed expression of ABCG1 and ABCA1 using siRNAs. siRNAs against ABCG1 and ABCA1 efficiently suppressed the corresponding gene's expression in the differentiated SH-SY5Y cells (Fig. 4a). siRNAs against ABCA1 significantly decreased 24-OHC efflux from differentiated SH-SY5Y cells in the presence of HDL, relative to scrambled siRNA (Fig. 4b). In contrast, no significant decrease in 24-OHC efflux was observed after treatment with ABCG1 siRNAs. Figure 4a shows that ABCA1 expression was stimulated when ABCG1 expression was suppressed by siRNAs. Therefore, it was possible that increased ABCA1 compensated for the lost function of ABCG1. siRNA#1 against ABCA1 did not affect ABCG1 expression, whereas siRNA#2 slightly suppressed it; furthermore, the level of ABCG1 expression did not correlate with HDL-dependent 24-OHC efflux (Fig. 4c). These results suggest that ABCA1 is involved in HDL-dependent 24-OHC efflux from differentiated SH-SY5Y cells, but the involvement of ABCG1 remains unclear. Simultaneous treatment with siRNAs against ABCA1 and ABCG1 negatively affected survival of differentiated SH-SY5Y cells, especially in the presence of 24-OHC (data not shown), leading us to take an alternative approach, as described below.
HDL-dependent efflux of 24-OHC by ABCA1 and ABCG1
To determine the involvement of ABCA1 and ABCG1 in 24-OHC efflux, we studied HEK293 cells stably expressing ABCA1 or ABCG1. In this system, we observed cholesterol efflux in the presence of apoA-I (Fig. 5a), but not HDL (Fig. 5b), when functional ABCA1 was expressed in HEK293 cells, as reported (Wang et al. 2000; Tanaka et al. 2001). In contrast, 24-OHC efflux from HEK/ABCA1 was clearly observed in the presence of HDL (Fig. 5d), whereas efflux in the presence of apoA-I was marginal (Fig. 5c). When functional ABCG1 was expressed in HEK293 cells, we observed cholesterol efflux in the presence of HDL as reported (Wang et al. 2004; Vaughan and Oram 2005; Kobayashi et al. 2006) (Fig. 6a); 24-OHC efflux from HEK/ABCG1 was also clearly observed in the presence of HDL (Fig. 6b). Non-functional ABCA1 or ABCG1, in which the lysine residue critical for ATP hydrolysis is replaced with methionine, exhibited neither cholesterol nor 24-OHC efflux activity. These results suggest that both ABCA1 and ABCG1 can mediate 24-OHC efflux to HDL.
Effects of LXR/RXR ligands on 24-OHC-induced cell death in rat primary neuron
24-OHC is toxic to neurons and induces apoptosis (Kolsch et al. 2001). We expected that the expression of ABCA1 and ABCG1 would protect neurons against the toxicity of 24-OHC, if they are involved in 24-OHC efflux. We examined this possibility by using rat primary cerebral neurons (Fig. 7). 24-OHC induced cell death in a concentration-dependent manner as revealed by LDH release assay. When primary neurons were pretreated with TO + RA for 16 h to induce the expression of ABCA1 and ABCG1, the neurons became significantly more resistant to 24-OHC.
24-OHC is the major cholesterol metabolite in the brain (Bjorkhem et al. 1997); high cholesterol turnover catalyzed by cholesterol 24-hydroxylase is essential for neural functions, especially learning (Kotti et al. 2006). We hypothesized that 24-OHC produced in neuronal cells is actively eliminated via transporters; in this study, we found that ABCA1 transports 24-OHC from differentiated SH-SY5Y neuron-like cells in an HDL-dependent manner.
We predicted that 24-OHC would be eliminated by the transporter(s) whose expression is induced by intracellular accumulation of 24-OHC itself. 24-OHC is an endogenous ligand of a nuclear receptor LXR; 24-OHC therefore induces transcription of ABCA1 and ABCG1 genes in macrophages and neuronal cells by activating the LXR/RXR heterodimer (Repa et al. 2000; Bjorkhem and Meaney 2004). However, it was reported that 24-OHC is not produced in SH-SY5Y cells (Ohyama et al. 2006), and indeed we could not detect 24-OHC production under our experimental conditions (data not shown). Therefore, we added 24-OHC exogenously and examined if 24-OHC induces expression of ABCA1 and ABCG1. ABCA1 and ABCG1 genes in differentiated SH-SY5Y cells well responded to exogenously added 24-OHC and synthetic ligands and 24-OHC (10 μM) induced them as efficiently as the synthetic ligands (Fig. 1). This concentration is physiologically relevant, as free 24-OHC in the brain was estimated at up to 30 μM (Lutjohann et al. 1996; Kolsch et al. 1999).
24-OHC efflux was clearly observed in the presence of HDL even without TO + RA treatment, and this efflux was further stimulated by the treatment (Fig. 2c). 24-OHC efflux in the presence of apoA-I was quite low compared to efflux in the presence of HDL. Because cholesterol efflux by ABCA1 is dependent on lipid-free apolipoproteins, and because efflux by ABCG1 is HDL-dependent, it was possible that ABCG1 is responsible for 24-OHC efflux from differentiated SH-SY5Y cells.
To examine if 24-OHC efflux to HDL was mediated by ABCG1, we suppressed the expression of ABCG1 and ABCA1 using siRNAs. Surprisingly, siRNA against ABCG1 had no significant effect, whereas siRNA against ABCA1 decreased 24-OHC efflux from differentiated SH-SY5Y cells in the presence of HDL. Because ABCA1 expression was induced when ABCG1 expression was suppressed, we speculate that increased levels of ABCA1 compensated for the lower levels of ABCG1 (Fig. 4a). However, when ABCA1 expression was suppressed, the 24-OHC efflux did not correlate with ABCG1 expression. Therefore, the involvement of ABCG1 in 24-OHC efflux from differentiated SH-SY5Y cells remains unclear. To determine the involvement of ABCA1 and ABCG1 in 24-OHC efflux, we studied HEK293 cells stably expressing ABCA1 or ABCG1 (Figs 5 and 6). 24-OHC efflux from HEK/ABCA1 and HEK/ABCG1 was clearly observed in the presence of HDL, whereas efflux in the presence of apoA-I was marginal. Taken together, these results suggest that both ABCA1 and ABCG1 mediate 24-OHC efflux to HDL.
Because cholesterol efflux by ABCA1 is apoA-I–dependent (Wang et al. 2000; Tanaka et al. 2001), it was surprising to observe 24-OHC efflux by ABCA1 in the presence of HDL, but not apoA-I. However, we previously reported (Nagao et al. 2009) that the function of ABCA1 does not depend on apoA-I, and that ABCA1 can transport cholesterol in the presence of bile salts in the medium. We have also proposed that lipid accumulation within the extracellular domain via ATP hydrolysis-dependent lipid transport causes conformational changes that generate apoA-I–binding site(s) on the surface of the extracellular domain of ABCA1, and that apoA-I bound to these site(s) is directly loaded with lipids by ABCA1 (Nagao et al. 2012). The amphipathic molecule 24-OHC may not be reserved in the extracellular domain of ABCA1, but may escape to the medium and bind to HDL as a lipid acceptor in the medium.
Because 24-OHC induces apoptosis of neuronal cells by generating free radicals (Kolsch et al. 2001), 24-OHC must be eliminated from cells as soon as possible after its production by 24-hydroxylase. Our study is the first to demonstrate that 24-OHC is actively eliminated by ABCA1. Treatment with ABCA1 siRNAs stimulated ABCG1 expression; simultaneous treatment with siRNAs against ABCA1 and ABCG1 negatively affected the survival of differentiated SH-SY5Y cells, especially in the presence of 24-OHC. We observed 24-OHC efflux from HEK/ABCG1 cells in the presence of HDL; therefore, ABCG1 may also be involved in 24-OHC elimination. The treatment of primary cerebral neurons with LXR/RXR ligands suppressed the toxicity of 24-OHC, suggesting that expression of ABCA1 and ABCG1 prevent apoptosis of neuronal cells. This should be also examined using neuronal cells from the ABCA1/ABCG1 double knockout mouse. ABCG4 was not analyzed in this study, since no antibody, which recognizes endogenously expressed ABCG4, was available. Because ABCG4 is expressed in brain and is required for sterol transport (Wang et al. 2008), it is possible that ABCG4 is also involved in 24-OHC efflux.
The G-395C polymorphism in the promoter of ABCA1 is implicated in low serum HDL levels (Probst et al. 2004) and reduces the CSF concentration of 24-OHC (Kolsch et al. 2006). This reduction may be caused by decreased 24-OHC efflux by ABCA1, or by lower levels of HDL, which serves as a lipid acceptor in the CSF. Specific inactivation of mouse brain ABCA1 leads to motor and sensorimotor behavioral and synaptic changes (Karasinska et al. 2009). The results reported in this study will enhance our understanding of the roles of ABC proteins and HDL in cholesterol homeostasis in the brain.
Our study was supported by a Grant-in-aid for Scientific research (S) from the Ministry of Education, Culture, Sports, Science and Technology of Japan; by the Program for Promotion of Basic and Applied Researches for Innovations in Bio-oriented Industry (BRAIN) of Japan; and by the World Premier International Research Center Initiative, MEXT, Japan. We have no conflict of interest to declare.