- Top of page
- Materials and methods
24S-Hydroxycholesterol (24S-OH-chol), a major cerebral cholesterol metabolite, is an endogenous ligand for the liver X receptor and is a potential stimulant of cholesterol release from glial cells. The elimination mechanism of 24S-OH-chol from the brain is one of the key issues for understanding cerebral cholesterol homeostasis. The purpose of the present study was to clarify the molecular mechanism of the elimination process of 24S-OH-chol across the blood–brain barrier (BBB). After an intracerebral injection in rats, [3H]24S-OH-chol was eliminated from the brain and the radioactivity derived from [3H]24S-OH-chol was detected in the plasma, while [3H]cholesterol was not significantly eliminated from the brain. Co-administration of unlabeled 24S-OH-chol significantly inhibited the [3H]24S-OH-chol elimination, while no inhibitory effect was seen at the same concentration of cholesterol. The [3H]24S-OH-chol elimination was inhibited by co-administration of probenecid, but not by benzylpenicillin. Pre-administration of digoxin completely inhibited the elimination. Xenopus laevis oocytes expressing rat oatp2 exhibited significant transport of [3H]24S-OH-chol, and this was inhibited by unlabeled 24S-OH-chol and digoxin, indicating that rat oatp2 transports 24S-OH-chol. These results are the first direct demonstration that 24S-OH-chol undergoes elimination from the brain to blood across the BBB via a carrier-mediated process, which involves oatp2 expressed at the BBB in rats.
24S-Hydroxycholesterol (24S-OH-chol) is the main cholesterol metabolite in the brain (Bjorkhem et al. 1997). Following conversion from cholesterol, 24S-OH-chol is eliminated from the brain to the circulating blood (Bjorkhem et al. 1997, 1998). This is considered to be the main elimination route of brain cholesterol. 24S-OH-chol is also an endogenous ligand for liver X receptor (LXR), and an LXR ligand has been reported to increase cholesterol release from primary glial cells associated with induction of ATP-binding cassette transporter A1 and G1 (Whitney et al. 2002). Therefore, 24S-OH-chol would act as a regulator involved in the autoregulatory mechanism for CNS cholesterol homeostasis (Pfrieger 2003).
A previous 18O-inhalation study has estimated the half-life of 24S-OH-chol in rat brain to be about 15 h (Bjorkhem et al. 1997). To undergo elimination from the brain to the circulation, 24S-OH-chol needs to cross the blood–brain barrier (BBB). However, there is no direct evidence of the brain-to-blood efflux transport of 24S-OH-chol across the BBB. In a study using erythrocytes (Meaney et al. 2002), it was hypothesized that 24S-OH-chol traverses the BBB by diffusion, as hydroxylation of the side chain of cholesterol allows oxysterol to be transferred across the lipid bilayer.
As a major part of plasma 24S-OH-chol originates from brain (Bjorkhem et al. 1998), its plasma levels have been investigated as a biomarker reflecting brain cholesterol homeostasis for the diagnosis of a number of neurodegenerative diseases, including Alzheimer’s disease. However, several reports have demonstrated that the 24S-OH-chol level in CSF is a better biomarker for Alzheimer’s disease, and the plasma levels of 24S-OH-chol showed only a weak correlation with the CSF levels (Papassotiropoulos et al. 2002; Schonknecht et al. 2002). Furthermore, in mouse, the 24S-OH-chol levels in the brain increased with aging associated with increased 24-hydroxylase (CYP46) expression in the brain, whereas its plasma levels in adults fell slightly with aging (Lund et al. 1999). These findings pose the question as to whether 24S-OH-chol can cross the BBB by diffusion.
The BBB expresses several kinds of transporters to regulate the exchange between the brain and circulating blood (Ohtsuki 2004). Organic anion transporters, such as organic anion transporter 3 (OAT3) and organic anion transporting polypeptide 2 (oatp2/oatp1a4), mediate the brain-to-blood efflux transport (Asaba et al. 2000; Mori et al. 2004). These OATs transport compounds structurally related to cholesterol, including bile acids (cholate and taurocholate) and steroid conjugates (estrone sulfate and dehydroepiandrosterone sulfate) (Noe et al. 1997; Kusuhara et al. 1999; Ohtsuki et al. 2004). However, there is no published report showing that 24S-OH-chol is a substrate of these transporters.
The purpose of the present study was, therefore, to clarify the mechanism governing the elimination of 24S-OH-chol at the BBB by direct measurement of its elimination process using the brain efflux index (BEI) method (Kakee et al. 1996) and to identify the molecules involved in the elimination process.
- Top of page
- Materials and methods
Bjorkhem et al. (1997) used an in vivo18O-inhalation technique to show that the rate of conversion of cholesterol into 24S-OH-chol was about two-thirds that of cholesterol synthesis. This balance of synthesis and conversion is important for maintaining a large pool of brain cholesterol. After conversion from cholesterol, 24S-OH-chol is eliminated from the brain to maintain a steady state brain level. The present study is the first to directly demonstrate the elimination of 24S-OH-chol from the brain across the BBB (Figs 1 and 2). As shown in Fig. 2, [3H]24S-OH-chol appears to cross the BBB at least partly in the intact form. The small peak at slice 8 suggests that a part of the [3H]24S-OH-chol was metabolized in the brain and/or peripheral organs.
The half-life of 24S-OH-chol in the brain pool is about 15 h, which includes processes of release from cells into ISF and elimination from ISF to blood. (Bjorkhem et al. 1997). The elimination rate from the brain across the BBB was evaluated by the intracerebral administration of [3H]24S-OH-chol as shown in Fig. 1, and the elimination half-life was determined to be 101 min. This shorter half-life suggests that release of 24S-OH-chol from the cells into ISF is slow and is the rate-limiting step of its elimination from the brain, and that after release into ISF, 24S-OH-chol is rapidly eliminated from ISF across the BBB. 24S-OH-chol functions as an LXR ligand and also exhibits a neurotoxic effect at high concentrations (Kolsch et al. 2001). Therefore, rapid elimination from the brain ISF would play important role in preventing impairment of CNS function resulting from the accumulation of 24S-OH-chol in the brain ISF.
To date, the passage of 24S-OH-chol across the BBB has been explained in terms of diffusion, as 24S-OH-chol is transferred into erythrocytes more rapidly than cholesterol (Meaney et al. 2002). The present study shows that the elimination process is significantly inhibited by unlabeled 24S-OH-chol and several inhibitors, resulting in an increase in the percentage remaining, as shown in Table 1. These forms of inhibition indicate the involvement of a saturable process, such as carrier-mediated transport, in the brain-to-blood efflux transport of 24S-OH-chol at the BBB. In contrast, cholesterol was not significantly eliminated from the brain up to 90 min after intracerebral administration (Fig. 3), and did not inhibit the elimination of 24S-OH-chol (Table 1). These results suggest that the transport system for 24S-OH-chol at the BBB does not accept cholesterol as a substrate. This substrate specificity seems to contribute to the difference in the BBB permeability of 24S-OH-chol and cholesterol.
The effects of OAT inhibitors were examined to identify the transporters involved (Table 1). Probenecid has a broad inhibition spectrum as far as OATs are concerned; Ki = 20 μmol/L for OAT3 and 70 μmol/L for oatp2 (Sugiyama et al. 2001). Benzylpenicillin and digoxin are selective substrates/inhibitors of OAT3 (Km = 40 μmol/L) and oatp2 (Km = 0.24 μmol/L), respectively (Kusuhara et al. 1999; Sugiyama et al. 2001; Ohtsuki et al. 2004). As the injectate was diluted 30-fold in the co-administration experiment and dilution of injectate was minimal in the pre-administration experiment (Kakee et al. 1996), the concentration of inhibitors tested in Table 1 is high enough to produce an inhibitory effect. It has been reported that the elimination of homovanillic acid and 6-mercaptopurine from the brain, which is mediated mainly by OAT3 at the BBB, was inhibited by co-administration of 100 mmol/L benzylpenicillin (Mori et al. 2003, 2004). Similar treatment did not affect the elimination of [3H]24S-OH-chol (Table 1). It has also been reported that pre-administration of 100 μmol/L digoxin did not inhibit the elimination of indoxyl sulfate from the brain, which was mediated mainly by OAT3 at the BBB (Ohtsuki et al. 2002). In the present study, pre-administration of 200 μmol/L digoxin increased the percentage remaining to 122% (Table 1), which is close to the extrapolated value at 0 min in Fig. 1, suggesting that digoxin almost completely inhibited the elimination. Therefore, the brain-to-blood efflux transport of 24S-OH-chol was mainly mediated by a digoxin-sensitive transporter, i.e. oatp2.
The percentage of [3H]24S-OH-chol remaining at 0 min was extrapolated to be over 100%. A similar result has been reported in the BEI study of GABA (Kakee et al. 2001). In the BEI study, the actual volume of the injected solution was normalized based on [14C]inulin, as some part of the injectate leaked from the brain during injection. When the test compound is bound and/or taken up by neural cells immediately following injection, the compound remains in the brain in a greater amount than that estimated based on [14C]inulin, and the percentage remaining is over 100% at 0 min. Therefore, the result shown in Fig. 1 also suggests that part of the injected 24S-OH-chol was bound and/or taken up by neural cells.
Organic anion transporting polypeptide 2 has been reported to be localized at the abluminal (brain side) and luminal (blood side) membrane of brain capillary endothelial cells (Gao et al. 1999), and to mediate the brain-to-blood efflux transport of dehydroepiandrosterone sulfate and 17β-estradiol-d-17β-glucuronide (Asaba et al. 2000; Sugiyama et al. 2001). The present study indicates that rat oatp2 transports 24S-OH-chol (Fig. 4) and this transport is inhibited by unlabeled 24S-OH-chol and digoxin as well as in vivo (Figs 5 and 6).
It remains possible that OAT3 transports 24S-OH-chol. However, involvement of OAT3 in 24S-OH-chol elimination across the BBB would be minor, even if OAT3 transports 24S-OH-chol, as the elimination of [3H]24S-OH-chol from the brain was inhibited almost completely by 200 μmol/L digoxin (Table 1), and OAT3-mediated transport was not affected by more than 10 mmol/L digoxin (Sugiyama et al. 2001). Taking into account the in vivo and in vitro results, we consider that 24S-OH-chol in the brain ISF is taken up into brain capillary endothelial cells by oatp2 localized at the abluminal membrane.
After 24S-OH-chol is taken up from brain ISF into brain capillary endothelial cells, it must be transported from the cells to the circulating blood to cross the BBB. The present study has not identified the molecule involved in the transport of 24S-OH-chol at the luminal membrane of brain capillary endothelial cells. Oatp2 has been reported to be expressed at the luminal membrane of brain capillary endothelial cells (Gao et al. 1999) and to be a bidirectional exchanger on the basis of a trans-stimulation effect in vitro (Li et al. 2000). Multidrug resistance protein 1a (mdr1a/ABCB1), multidrug resistance-associated protein 4 (ABCC4) and breast cancer resistance protein (ABCG2) have also been reported to be expressed at the luminal membrane of brain capillary endothelial cells and to pump substrates out to the circulating blood (Schinkel et al. 1995; Wakayama et al. 2002; Hori et al. 2004; Leggas et al. 2004). Taurocholate is a low-affinity substrate of Chinese hamster MDR1 (Lam et al. 2005) and inhibits the transport activity of human multidrug resistance-associated protein 4 and breast cancer resistance protein (Imai et al. 2003; Zelcer et al. 2003). It is possible that 24S-OH-chol is a substrate of some of these transporters, as well as oatp2. Furthermore, the distribution of digoxin into the brain was enhanced in mdr1a knockout mice (Schinkel et al. 1995), indicating that digoxin is transported by mdr1a at brain capillary endothelial cells. Therefore, a contribution from an inhibitory effect of digoxin on mdr1a transport at the luminal membrane to the results shown in Table 1 can not be ruled out.
24S-Hydroxycholesterol is an endogenous potent LXR ligand. Two subtypes of LXR, α and β, are expressed in the brain, and LXRα/β double knockout mice have been reported to develop a number of abnormalities in the brain (Wang et al. 2002). 24S-OH-chol induces the expression of ABCA1 in neurons, glial cells and brain capillary endothelial cells (Fukumoto et al. 2002; Panzenboeck et al. 2002; Whitney et al. 2002; Liang et al. 2004). LXR ligand also induces apolipoprotein E secretion from astrocytoma cells (Liang et al. 2004). Therefore, the brain level of 24S-OH-chol influences CNS functions involving cholesterol homeostasis. The present findings indicate that changes in the efflux transport of 24S-OH-chol mediated by oatp2 may affect the brain levels of 24S-OH-chol.
To date, several reports have investigated the possible association between polymorphism of 24-hydroxylase, which converts cholesterol to 24S-OH-chol, and Alzheimer’s disease (Ingelsson et al. 2004; Wang et al. 2004; Golanska et al. 2005), but no final conclusion has been reached. The present study indicates that the molecule(s) responsible for the brain-to-blood efflux transport of 24S-OH-chol could be related to the risk of Alzheimer’s disease. Therefore, it is necessary to clarify the responsible molecule(s) at the BBB in humans, as the subtypes of oatp are not conserved between rodents and humans. The closest human subtype to oatp2 is OATP-A/OATP1A2 in terms of amino acid sequence (Hagenbuch and Meier 2004). OATP-A transports taurocholate and dehydroepiandrosterone sulfate, like oatp2, while OATP-A does not transport digoxin, which is an oatp2 substrate. It has also been reported that OATP-A is localized at brain capillary endothelial cells (Gao et al. 2000). Based on these findings, OATP-A appears to be a candidate transporter for 24S-OH-chol at the human BBB. Nevertheless, as the expression levels of OATP subtypes and their contributions to transport at human BBB are poorly understood, the involvement of other OATP subtypes in 24S-OH-chol elimination at the human BBB can not be ruled out.
In conclusion, the elimination of 24S-OH-chol from the brain is governed by a carrier-mediated process at the BBB in rats, and oatp2 is involved in this brain-to-blood efflux. Identification of the transporter(s) responsible for the elimination of 24S-OH-chol at the human BBB is necessary to evaluate the role of this transporter in CNS cholesterol homeostasis and CNS diseases in humans.