Effects of Lipoprotein Lipase on Uptake and Transcytosis of Low Density Lipoprotein (LDL) and LDL-associated α-Tocopherol in a Porcine in Vitro Blood-Brain Barrier Model*

During the present study the contribution of lipoprotein lipase (LPL) to low density lipoprotein (LDL) holoparticle and LDL-lipid (α-tocopherol (αTocH)) turnover in primary porcine brain capillary endothelial cells (BCECs) was investigated. The addition of increasing LPL concentrations to BCECs resulted in up to 11-fold higher LDL holoparticle cell association. LPL contributed to LDL holoparticle turnover, an effect that was substantially increased in response to LDL-receptor up-regulation. The addition of LPL increased selective uptake of LDL-associated αTocH in BCECs up to 5-fold. LPL-dependent selective αTocH uptake was unaffected by the lipase inhibitor tetrahydrolipstatin but was substantially inhibited in cells where proteoglycan sulfation was inhibited by treatment with NaClO3. Thus, selective uptake of LDL-associated αTocH requires interaction of LPL with heparan-sulfate proteoglycans. Although high level adenoviral overexpression of scavenger receptor BI (SR-BI) in BCECs resulted in a 2-fold increase of selective LDL-αTocH uptake, SR-BI did not act in a cooperative manner with LPL. Although the addition of LPL to BCEC Transwell cultures significantly increased LDL holoparticle cell association and selective uptake of LDL-associated αTocH, holoparticle transcytosis across this porcine blood-brain barrier (BBB) model was unaffected by the presence of LPL. An important observation during transcytosis experiments was a substantial αTocH depletion of LDL particles that were resecreted into the basolateral compartment. The relevance of LPL-dependent αTocH uptake across the BBB was confirmed in LPL-deficient mice. The absence of LPL resulted in significantly lower cerebral αTocH concentrations than observed in control animals.


plays a central role in lipoprotein metabolism. LPL is synthesized in most extrahepatic tissues
where it is transported to the endothelium (1, 2), a process dependent on heparan sulfate proteoglycans (HSPG) and the very low density lipoprotein (VLDL) receptor (3). There, LPL hydrolyzes triglycerides (TGs) in chylomicrons and VLDL, thereby generating free fatty acids that enter either storage or oxidative pathways. LPL also contributes to the exchange of lipids and apoproteins between different lipoprotein classes, thus affecting size and composition not only of TG-rich lipoproteins but also of low and high density lipoproteins (LDL and HDL) (1,4). In addition to its lipolytic function, it was demonstrated that proteoglycan-bound LPL directly interacts with lipoproteins, concentrating lipoprotein particles on the cell surface where they are internalized along with lipoprotein receptors (5)(6)(7)(8) or in conjunction with proteoglycans (9). This "bridging function" is independent of the enzymatic activity of LPL. In addition to lipoprotein receptor-or proteoglycan-mediated internalization of lipoproteins, LPL can contribute to selective lipid uptake. During selective uptake, originally lipoproteinassociated lipids from the surface and core are internalized by cells without concomitant endocytosis of the lipid-depleted lipoprotein (holo)particle. This pathway was described for HDL-and LDL-associated lipids to occur via pathways mediated by scavenger receptor BI (SR-BI) (10 -14) or LPL (15)(16)(17).
In contrast to other tissues, the functional significance of LPL expression in brain remains unclear. Yet there is ample evidence that LPL is functional in this organ: LPL expression was documented in brain microvessels (18) and in various brain regions (19) and is regulated by and responsive to starvation (20). Therefore, it was suggested that LPL at the bloodbrain barrier (BBB) could supply the neonatal brain with unsaturated fatty acids during embryonic development (21). On the other hand, the bridging function of LPL was proposed to contribute to the transport of lipoprotein-associated vitamins and cholesterol esters across the BBB (2). This hypothesis is supported by the fact that tissue-specific overexpression of LPL in transgenic animals results in significantly increased ␣-tocopherol (␣TocH) concentrations at the site of LPL expression (22).
During the present study we examined the pathways of LDL holoparticle and selective LDL-␣TocH uptake in primary porcine brain capillary endothelial cells (BCECs) and the modulation of these pathways by LPL. In addition, we have studied whether LPL affects transcytosis rates of LDL particles across an in vitro model of the BBB and whether targeted knock-out of the LPL gene affects cerebral ␣TocH levels.

Methods
Isolation and Culture of BCECs-Porcine BCECs were isolated by sequential enzymatic digestion and centrifugation steps according to Tewes et al. (23) and characterized as described previously (24). During the first 2 days in vitro, BCECs (from one brain) were cultured in eight 75-cm 2 rat tail collagen-coated culture flasks in M199 containing 10% ox serum, 0.1 mg/ml gentamicin, 2 mM glutamine, 50 units/ml penicillin, and 50 g/ml streptomycin. After 2 days in vitro the cells were seeded on collagen-coated multiwell cell culture clusters at a minimum density of 30,000 cells/cm 2 in M199 medium containing 10% ox serum, 2 mM glutamine, 50 units/ml penicillin, and 50 g/ml streptomycin.
Isolation and Labeling of Lipoproteins-Human LDL was prepared by density gradient ultracentrifugation of plasma obtained from normolipidemic human volunteers in a TL120 tabletop ultracentrifuge (at r av 350,000 ϫ g, Beckman, Vienna) (25). LDL was recovered by direct aspiration and desalted by size exclusion chromatography on PD-10 columns. Molar concentrations were calculated from total lipoprotein mass using a molecular mass of 2500 kDa for LDL (26). Acetylation of LDL was performed as described by Basu et al. (27).
Na 125 I Labeling-Iodination of LDL was performed as described (28) using N-Br-succinimide as the coupling agent. Routinely 1 mCi of Na 125 I was used to label 3 mg of LDL protein. This procedure resulted in specific activities between 300 and 450 dpm/ng of protein. Lipidassociated activity was always less than 3% of total activity. No fragmentation of apoB-100 due to the iodination procedure could be detected by SDS or native PAGE and subsequent autoradiography or Coomassie staining.
[ 14 C]␣TocH Labeling-LDL (0.6 mg of protein, final volume 2 ml in phosphate-buffered saline) was labeled by direct addition of an ethanolic [ 14 C]␣TocH solution and incubation at 37°C (3 h, under argon, shaking water bath (24)). Non-lipoprotein-associated [ 14 C]␣TocH was removed by size exclusion chromatography on PD-10 columns. The labeling procedures did not affect the LDL integrity as verified by native PAGE.
Isolation of LPL from Bovine Milk-LPL was isolated from fresh, unpasteurized bovine milk (1 liter) exactly as described previously (29). Purity was assessed by SDS-PAGE (8%) followed by Coomassie Blue staining. Yields for LPL ranged from 500 to 1500 g of LPL/liter of milk with activities between 400 and 700 mol of free fatty acids/h/mg of LPL protein.
Determination of Tracer Uptake-BCECs were cultured on rat-tail collagen-coated multiwell cluster plates. 125 I-and [ 14 C]␣TocH-labeled LDL were added to the medium at the indicated concentrations. At the end of the incubation the cells were washed twice with ice-cold TBS containing bovine serum albumin (2 mg/ml), followed by another two washes with ice-cold TBS. Cells were lysed by treatment with NaOH (0.3 N, 25°C, 30 min). An aliquot of the cell lysate was used to determine the cellular protein concentration, and the remaining lysate was mixed with scintillation mixture to determine the radioactivity. For 125 I-LDL uptake experiments, cells were plated on 12-well cluster plates. Incubations were performed in the absence or presence of a 20-fold excess of non-labeled lipoprotein to differentiate between total and specific cell association or binding. To differentiate between degraded, internalized, and bound fractions the medium was removed and the amount of non-trichloroacetic acid-precipitable radioactivity was determined as described below. Cells were then washed and incubated in the presence of heparin (100 units/ml), and the released radioactivity is referred to as the "bound" fraction. The cells were then washed again and lysed in NaOH (0.3 N), and the remaining radioactivity is referred to as the "internalized" fraction. Degradation of 125 I-LDL was estimated by measuring the non-trichloroacetic acid-precipitable radioactivity in the medium after precipitation of free iodine with AgNO 3 (24). [ 14 C]␣TocH-LDL uptake is expressed in terms of apparent particle uptake (HDLprotein equivalents). Based on the specific activity of the labeled lipoprotein preparations, the amount of lipoprotein that would be necessary to deliver the observed amount of tracer was calculated. This is necessary to allow quantitative comparison of data between lipoproteinparticle independent (selective) lipid uptake by non-endocytotic mechanisms and lipoprotein-dependent (holoparticle) uptake (10).
Construction of Recombinant SR-BI Adenovirus-The adenoviral plasmid shuttle vector (pAvCvSv) and pJM17 vectors were kindly supplied by L. Chan (Baylor College of Medicine, Houston, TX). Human SR-BI cDNA (kindly supplied by H. Hauser, Eidgenoessische Technische Hochschule, Zü rich, Switzerland), which was originally inserted into pcEXV-3 vector, was partially restricted with EcoRI, and the 2.5-kb band was eluted from the gel. This band was subcloned into pBluescript using the EcoRI site, amplified, and restricted with KpnI, and the fragment was finally partially restricted with BamHI and inserted in the plasmid shuttle vector. Adenovirus infection of BCECs was performed as described previously (30) to obtain transiently transfected cells.
Knock-out Animals-Interbreeding of heterozygous LPL knock-out mice (L1) resulted in the progeny of three different genotypes, L2 (wild type, 25%), L1 (heterozygous, 50%), and L0 (homozygous LPL knock-out mice, 25%) (31). Genotyping was performed from tail tip DNA by PCR analysis as reported previously (31). L0 animals die postnatally between 18 and 24 h after birth; therefore, L2, L1, and L0 mice were killed between 0 and 4 h after birth (using an overdose of isoflurane inhalation). Brains were removed (average weight, ϳ65 mg of wet tissue) and transferred to liquid nitrogen. The snap-frozen tissues were homogenized in liquid nitrogen, and the ␣TocH content was determined as described (22).
Statistical Analysis-Statistical differences were analyzed by using Student's t test with the level of significance set at 0.05. All data are expressed as mean Ϯ S.D. Box Whiskers plots were generated with the GraphPad prism package.

RESULTS
Immunoreactive LPL Is Detectable in BCEC Lysates-In cell lysates obtained from a BCEC population with Ն95% purity (as assessed by immunocytochemistry with anti-factor VIII and antismooth muscle cell actin IgG) immunoreactive porcine LPL comi-grated with bovine LPL and was detected at ϳ60 kDa (Fig. 1A).
Effect of Exogenous LPL on LDL Cell Association-To determine the effects of LPL on cell association of 125 I-LDL to BCECs, cells were incubated at increasing LPL:LDL molar ratios (Fig. 1B). The addition of LDL and enzymatically active LPL at increasing molar ratios dose-dependently increased the total cell association of 125 I-LDL, indicating that LPL increases binding and/or uptake of LDL. The most pronounced effect (ϳ11-fold increase, 403 versus 4529 ng/mg of cell protein) was observed at a molar ratio of 100:1 (LPL:LDL, 120 g/ml LPL and 50 g/ml LDL). At the concentration ranges tested during these experiments the effect of LPL tended to level off but did not reach a real saturation plateau.
To assess whether this increase in cell association is due to enzymatically active LPL, incubations were performed in the presence of tetrahydrolipstatin (THL), a potent inhibitor of LPL activity ( Fig. 2A). In the absence of LPL, THL and heparin were without effect on LDL association. The addition of LPL to the medium (molar ratio of LPL:LDL ϭ 100) resulted in a pronounced, 11-fold increase of cell-associated LDL that was independent of the presence of THL. This indicates that the observed LPL-mediated effect does not depend on its TG-hydrolase activity and thus results from the "bridging function" of the enzyme. More than 50% of 125 I-LDL was not heparinreleasable, indicating efficient lipoprotein particle uptake (Fig. 2B).
As LPL was discussed to contribute to the supply of the central nervous system with lipophilic vitamins, the effect of LPL on the uptake of LDL-associated [ 14 C]␣TocH was investi-gated. In principle, uptake of LDL-associated ␣TocH can occur by holoparticle uptake, selective uptake, or a combination of both pathways. The effect of LPL on either of these pathways is shown in Table I. In line with our previous results (24), ␣TocH uptake exceeded holoparticle uptake ( 14 C versus 125 I), indicating selective uptake of ␣TocH. The addition of LPL to the incubation medium resulted in significantly enhanced 125 I-and [ 14 C]␣TocH-LDL association (4-to 10-fold) in a time-dependent manner (1-5 h). The addition of LPL augmented selective ␣TocH uptake at 3 and 5 h (3-and 4.7-fold, respectively). These findings indicate that LPL is able to promote LDL holoparticle uptake and selective uptake of LDL-associated ␣TocH in endothelial cells constituting the BBB.
The LDL-receptor Contributes to LPL-mediated Holoparticle Binding and Turnover-The fact that BCECs express the LDLreceptor (32)(33)(34) prompted us to study the interaction of the LDL-receptor and LPL at the BBB. During these experiments (Fig. 3 To exclude the possibility that the effects described in Fig. 3 are due to up-regulation of SR-BI, the same experiments were performed with acLDL, which is a ligand for SR-BI but not for the LDL-receptor (Fig.  3). As observed with native LDL, the binding of acLDL was increased 1.5-fold in response to LPDS. However, the effects of LPL were much weaker with acLDL and led to 4-and 8-fold increases (FCS and LPDS, respectively) of acLDL binding.
To further explore whether the LDL-receptor is involved in LPL-dependent LDL particle turnover at 37°C, the next series of experiments was performed under FCS or LPDS conditions and LDL binding, internalization, and degradation were determined (Fig. 4). Under these experimental conditions, the addition of LPL led to a pronounced 3.3-fold increase of binding at the beginning of the experiment, an effect that was further increased by a preincubation in LPDS-containing medium (4.7-fold, Fig. 4A). At longer incubation times this effect was ablated, due to internalization of the originally surface-bound particles. Up-regulated LDL-receptor expression resulted in ϳ2-fold higher LDL internalization rates (Fig. 4B). The presence of LPL increased 125 I-LDL internalization between 8-and 18-fold. In line with data shown in Fig. 4 (A and B), degradation of 125 I-LDL (Fig. 4C) was considerably enhanced under LPDS conditions and the presence of exogenous LPL: when cells were preincubated in LPDS, a 1.2-fold increase in overall degradation was observed. In the presence of LPL the degradation rates were increased 3.7-fold (LPL and FCS) and 4.4-fold (LPL and LPDS), respectively. In summary, the data shown in Fig. 4 strengthen the concept that LPL-supported LDL particle turnover by BCECs is modulated at the level of LDL-receptor expression.
Mechanisms Contributing to LPL-mediated Selective Uptake of LDL-associated ␣TocH-The prime candidate receptor mediating selective uptake of lipoprotein-associated lipids is SR-BI. During preliminary experiments we have tested whether SR-BI contributes to selective uptake of LDL-associated ␣TocH in transiently SR-BI-transfected COS cells. These experiments revealed that high level SR-BI overexpression resulted in only slightly increased (1.5-fold) selective uptake of LDL-associated ␣TocH (data not shown). To further determine whether SR-BI and LPL could act synergistically on ␣TocH uptake, BCECs were transfected by an adenoviral approach (30). Along this line it is important to note that the virus concentrations needed to obtain feasible transfection rates (ϳ70%) were 100 times higher for BCECs as compared with COS cells (m.o.i. 1000 versus 10 in COS cells). Transfection of BCECs with a control virus (containing the ␤-galactosidase reporter gene) was without effect on selective ␣TocH uptake (Fig. 5), whereas adenoviral transduction of SR-BI resulted in a 2.5-fold increase in selective LDL-␣TocH uptake. Addition of LPL resulted in 5.5-and 5.9-fold higher selective LDL-␣TocH uptake in wild-type and SR-BI-transfected cells, respectively. Thus, it appears that LPL-mediated induction of selective LDL-␣TocH uptake is independent of SR-BI expression. When cells were treated with NaClO 3 (an inhibitor of proteoglycan sulfation) prior to the uptake experiments, selective uptake of ␣TocH was significantly attenuated (2-fold), indicating that HSPG mediates at least part of LPL-dependent selective uptake of LDL-

FIG. 3. Effect of LDL-receptor up-regulation on LPL-mediated LDL binding by BCECs.
Prior to the experiment, cells in 12-well trays were precultured (24 h) in medium containing 10% FCS (indicated by Ϫ) or LPDS (indicated by ϩ) to regulate LDL-receptor expression. Cells then received cold medium (4°C) containing LDL or acLDL (10 g/ml) in the absence or presence of LPL (10 g/ml). After a 1.5-h incubation at 4°C the cells were washed and lysed, and the radioactivity and the cellular protein content were determined. Results represent means Ϯ S.D. from one representative experiment performed in triplicates. The inset shows immunoblot analysis of LDL-receptor expression in BCEC homogenates. Prior to Western blot analysis with a monoclonal antibody, cells were kept in medium containing FCS (lane 1) or LPDS (lane 2). The LDL-receptor was detected at ϳ160 kDa (under reducing conditions). associated ␣TocH. Cell viability was not affected by the presence of NaClO 3 (Ͼ95% viable cells after 72 h as determined by trypan blue exclusion).
Effects of LPL on LDL Transcytosis-BCECs are polarized cells that actively regulate the diffusion and transport of cir-culating metabolites or xenobiotics into the brain parenchyma and vice versa (35). To study the effects of exogenous LPL on lipoprotein transcytosis, cells were cultured on Transwell inserts as described under "Experimental Procedures." To get an idea about the contribution of lysosomal degradation to LDL holoparticle turnover, degradation was assessed as outlined under "Experimental Procedures." The amount of trichloroacetic acid-soluble radioactivity was dependent on the presence of LPL (1756 versus 2356 ng/mg of cell protein, Fig. 6), in line with data shown in Fig. 4C. Analysis of cell-associated tracers revealed that the uptake of LDL-associated [ 14 C]␣TocH exceeded 125 I-LDL holoparticle cell association; the addition of LPL enhanced 125 I-LDL and [ 14 C]␣TocH association between 2.2-and 2.5-fold. The presence of LPL slightly (but not significantly) increased transcytosis of 125 I-LDL to the basolateral compartment (2215 Ϯ 62 versus 2788 Ϯ 610 ng of LDL/mg of cell protein). However, basolaterally resecreted LDL was depleted in [ 14 C]␣TocH tracer (38 and 48% less compared with LDL protein; in the absence or presence of LPL, respectively). The amount of trichloroacetic acid-soluble material in the basolateral compartment was ϳ80-fold lower than in the apical compartment (21 and 30 ng of LDL/mg of cell protein, absence or presence of LPL, respectively). These findings indicate that the presence of LPL contributes to enhanced uptake of LDL and LDL-associated lipids but not to increased transcytosis.
LPL Contributes to ␣TocH Uptake Across the BBB in Vivo-To test whether LPL contributes to ␣TocH uptake across the BBB in vivo, brains of control (L2), heterozygous (L1), or homozygous (L0) LPL knock-out mice were analyzed (31). The cerebral ␣TocH content of L0 animals was significantly (p Ͻ 0.01) lower as compared with L1 and L2 animals (1.28 Ϯ 0.28, 1.56 Ϯ 0.21, and 1.61 Ϯ 0.15 g/g wet tissue, respectively). In contrast, the ␣TocH concentrations in brains of L2 and L1 animals were not significantly different (Fig. 7). These findings indicate that the absence of LPL affects cerebral ␣TocH concentrations.

DISCUSSION
Although LPL mRNA is present in brain and LPL protein expression was demonstrated in brain microvessels (18,19), the functional significance of LPL at the BBB is not clear. The present study aimed at characterizing LPL-modulated uptake and transcytosis routes for LDL and LDL-associated ␣TocH. LPL is able to form a molecular bridge between proteoglycans and lipoproteins in vitro (reviewed in Refs. 1 and 2) and in vivo (36); therefore, it was suggested that LPL could facilitate transcytosis of lipoprotein-associated vitamins across the BBB (2). This could result either from enhanced LPL-mediated holoparticle transcytosis across the endothelial layer constituting the BBB, or by LPL-mediated selective uptake, as described for HDL-and LDL-associated cholesterol esters (15)(16)(17)37).
During the present studies we have found that primary BCECs isolated from porcine brains contain endogenous LPL that was most probably transferred to BCECs from surrounding cells. The addition of exogenous LPL augmented both LDL holoparticle and selective ␣TocH uptake in BCEC monolayers. The addition of LPL resulted in significantly increased LDL holoparticle cell association and turnover via mechanisms related to LDL-receptor expression. In the presence of LPL, selective ␣TocH uptake from LDL increased up to 5-fold, and this was inhibited by adding NaClO 3 . Although adenoviral overexpression of SR-BI resulted in 2-fold increased selective ␣TocH uptake, SR-BI overexpression was without effect on ␣TocH uptake in the presence of LPL. Transwell experiments revealed that LDL holoparticle transcytosis was unaffected by LPL. However, an important observation was the significant ␣TocH depletion of lipoprotein particles that were resecreted into the basolateral compartment.
As demonstrated for BCECs obtained from other species (32,38,39), expression of the LDL-receptor in porcine BCECs is responsive to the availability of cholesterol in the culture medium. This is important because increased binding of LDL to HSPG in the presence of LPL is followed by LDL-receptormediated uptake (40,41). During the present study the effects of LPL-mediated increase in binding, holoparticle internalization, and degradation were regulated in a similar manner as LDL holoparticle turnover in the absence of LPL, indicative of an involvement of the LDL-receptor. In line with earlier results (42) binding experiments performed at 4°C strongly imply that the interaction of LPL⅐LDL complexes occurs via the LDL-receptor. When the LDL-receptors of BCECs were saturated with 125 I-LDL at 4°C and then switched to 37°C, the majority of 125 I-LDL was internalized at the earliest time point analyzed (0.5 h), which roughly corresponds to the recycling time of the LDL-receptor (up to 15 min (43)). An involvement of the LDL-receptor during endocytosis is further substantiated by experiments with 125 I-acLDL. acLDL is a ligand for the scavenger receptors class A and B, both being expressed on porcine BCECs (30,39,44), but not for the LDL-receptor (27). In accordance, the effect of exogenous LPL on acLDL binding to BCECs was much lower as compared with native LDL (4-to 8-fold increase versus 27-to 45-fold enhancement for native LDL).
Dehouck and colleagues (32) provided evidence that the LDL-receptor facilitates transcytosis of LDL across a bovine in vitro model of the BBB. This was based on two major observations: (i) a monoclonal anti-LDL-receptor antibody that interacts with the binding domain of the receptor completely blocked transcytosis and (ii) transcytosis occurred via a nondegradative pathway, obviously different from the classic LDLreceptor pathway. In BCEC monolayers LDL enters a degradative pathway that is up-regulated in response to cholesterol depletion. Moreover, in the Transwell system the vast majority of degraded LDL (ϳ99%) was detected in the apical compartment, with only 1-1.5% detectable in the basolateral compartment. Whether this indicates that only a subset of LDL-receptors undergoes transcytosis and bypasses the lysosomal compartment is currently not clear. However, from our findings it is evident that LPL has no effects on the net transcytosis rates of LDL holoparticles. These observations suggest that the majority of LDL⅐LPL complexes formed in the luminal compartment are directed to lysosomal degradation and, thus, are not available for LDL-receptor-mediated transcytosis.
The brain depends on a constant and adequate supply with ␣TocH. This is underscored by severe and characteristic neuropathologies that develop in response to ␣TocH deficiency, in Only BCEC monolayers with a TER Ն 300 ohm/cm 2 were used. Mean sucrose permeability was 4.1 ϫ 10 Ϫ6 cm/s. Cells were incubated in the presence of 125 I-LDL or [ 14 C]␣TocH-LDL (40 g per well in the upper, apical compartment) in the absence or presence of LPL (10 g/well) for 3 h at 37°C. Thereafter the insert was removed, washed twice with cold TBS (containing bovine serum albumin) and twice with TBS, and cut out, and the cells were lysed in NaOH (0.3 N). Part of the lysate was used to determine the cellular protein content, and the remaining aliquot was used to measure the cell-associated radioactivity. The medium from the basolateral compartment was removed and counted. Data shown represent mean values from one representative experiment performed in quintuplicates.
FIG. 7. ␣TocH content in brains of control, heterozygous, and homozygous LPL knock-out mice. Newborn animals (L2, n ϭ 13; L1, n ϭ 29; L0, n ϭ 18) were killed between 0 and 4 h after birth. Genotyping was performed from tail tip DNA by PCR analysis. Brains were removed, snap-frozen, and homogenized in liquid nitrogen. Following Folch extraction, the lipid extracts were analyzed by HPLC with fluorescence detection. Statistical significance was analyzed by Student's t test, and data are presented as Box Whiskers plots. a manner reminiscent of Friedreich's ataxia (45,46). In fasting humans circulating ␣TocH is roughly equally distributed between the LDL and HDL fractions (22). Therefore, LDL holoparticle uptake or selective uptake of LDL-␣TocH at the BBB and subsequent transcytosis would be an attractive pathway to facilitate ␣TocH uptake into the brain. As demonstrated in one of our earlier reports (24), BCECs are capable of selective ␣TocH uptake. The present study revealed that the basal capability of BCECs for selective ␣TocH uptake is increased ϳ5-fold in the presence of exogenous LPL. As has been reported for LDL-associated cholesterol ether (17), and HDL-associated cholesterol esters (15,16), LPL-dependent selective ␣TocH uptake was independent of LPL catalytic activity as determined in the presence of THL. Treatment of the cells with NaClO 3 (30 mM, higher concentrations affected cell viability) led to a 50% reduction in LPL-mediated selective uptake, indicating that proteoglycans are involved in this process. These findings are in line with other reports: Seo et al. (17) have demonstrated that LPL-mediated selective uptake of LDL-cholesterol ether was significantly reduced in fibroblasts after blockage of proteoglycan sulfation and in proteoglycan-deficient Chinese hamster ovary cells. HSPG was also implicated in the uptake of VLDL by macrophages (9) and LPL-mediated LDL uptake by HepG2 cells and fibroblasts (40).
The present study revealed that ␣TocH uptake into brain is modulated by LPL in vivo. Brains obtained from newborn homozygous LPL knock-out mice (31) had a significantly lower absolute ␣TocH content as compared with heterozygous (L1) and control (L2) animals. At birth the lipoprotein profiles of L2, L1, and L0 mice are not significantly different. Immediately after birth TGs of L0 mice are mildly increased and HDL levels are not significantly different from controls (31). This ensures that the reduced ␣TocH concentration in brain of LPL knockout animals observed during the present study are due to the absence of LPL and not a result of altered lipoprotein profiles. Previously, we have shown that selective uptake of HDL-associated ␣TocH is mediated via SR-BI (30) and most probably bypasses LPL-dependent selective uptake. Consistent with this finding, it was demonstrated that the cerebral ␣TocH content of SR-BI (but not LDL-receptor) knock-out mice is significantly reduced (by ϳ70% (47)).
The potential role of SR-BI on selective LDL-␣TocH uptake was studied by transient adenoviral transduction of human SR-BI into BCECs. SR-BI is a receptor with multiligand specificity and mediates selective lipid uptake also from LDL (14). During the present study SR-BI transduction resulted in increased selective LDL-␣TocH uptake, however, this was only 50% of the value obtained in the presence of LPL. Most importantly the effect of SR-BI overexpression was ablated when LPL was present during the incubation. Along this line it is interesting to note that HDL turnover by BCECs was much less affected by the presence of exogenous LPL. 2 The presence of LPL in the apical compartment of Transwell inserts had no effect on LDL holoparticle transcytosis (see above) and, thus, no effect on ␣TocH accumulation in the basolateral compartment. Probably the most intriguing observation during the present study was a significant ␣TocH depletion (ϳ50% relative to protein) of the LDL particles that were resecreted into the basolateral compartment. This would indicate that part of ␣TocH (and probably other lipids) is shed from non-degraded LDL particles during transcytosis across the in vitro BBB. A similar phenomenon of selective lipid depletion that was reported for HDL undergoing retroendocytosis in hepatocytes (48,49) could occur at the BBB. In this hepatic, SR-BI-depend-ent process, HDL traffics through the endosomal recycling compartment, a cholesterol-and cholesterol ester-depleted HDL particle is resecreted at the basal cell surface, while polarized secretion of HDL-derived cholesterol occurs at the apical membrane. During this pathway HDL particles can lose as much as 75% of their original cholesterol ester tracer. This pathway was termed selective transcytosis of lipoprotein cholesterol (49). Presently it is not clear whether ␣TocH that has been shed from LDL particles during transcytosis generates a "backup pool" of endothelial cell-associated ␣TocH, which is transported into the brain by mechanisms independent of LDL transcytosis.