Cholesterol Distribution in the Golgi Complex of DITNC1 Astrocytes Is Differentially Altered by Fresh and Aged Amyloid (cid:1) -Peptide-(1– 42)*

The Golgi complex plays an important role in cholesterol trafficking in cells, and amyloid (cid:1) -peptides (A (cid:1) s) alter cholesterol trafficking. The hypothesis was tested that fresh and aged A (cid:1) -(1–42) would differentially modify Golgi cholesterol content in DINTC1 astrocytes and that the effects of A (cid:1) -(1–42) would be associated with the region of the Golgi complex. Two different methods were used to determine the effects of A (cid:1) -(1–42) on Golgi complex cholesterol. Confocal microscopy showed that fresh A (cid:1) -(1–42) significantly increased cholesterol and that aged A (cid:1) -(1–42) significantly reduced cholesterol content in the Golgi complex. Isolation of the Golgi complex into two fractions using density gradient centrifugation showed effects of aged A (cid:1) -(1–42) similar to those observed with confocal microscopy but revealed the novel finding that fresh A (cid:1) -(1–42) had opposite effects on the two Golgi fractions suggesting a specificity of A (cid:1) -(1–42) perturbation A ApoE levels Experiments conducted using fresh and aged A (cid:1) -(1–42) in immortalized DITNC1 astrocytes Two different methods were used to determine the effects of A (cid:1) -(1–42) on Golgi cholesterol content of astrocytes: 1) confocal microscopy using the fluorescent cholesterol analogue NBD-cholesterol and a fluorescent marker for the Golgi complex, BODIPY TR ceramide (30); and 2) isolation of the Golgi complex into two different fractions thought to represent different regions of the Golgi complex using sucrose density gradient centrifugation (31, 32). The results presented herein provide new insights into interaction of A (cid:1) and cholesterol. We show that fresh A (cid:1) -(1–42) has a strikingly different effect on cholesterol levels in the Golgi complex of astrocytes as compared with aged A (cid:1) -(1–42). Incubation of astrocytes with fresh A (cid:1) -(1–42) significantly increased cholesterol levels, whereas aged A (cid:1) -(1–42) significantly reduced cholesterol levels in the Golgi complex as shown by fluorescent probes and confocal microscopy. The reverse peptide A (cid:1) -(42–1), either fresh or aged, did not alter Golgi complex cholesterol levels, which argues against a nonspecific action of A (cid:1) -(1–42) . Isolation of the astrocyte Golgi complex into different fractions revealed the novel finding that the effects of fresh A (cid:1) -(1–42) on cholesterol content were opposite on the two Golgi fractions,

Several different lines of evidence point to a potentially important but not well understood interaction between Alzheimer's disease (AD) 1 and cholesterol (for reviews, see Refs. [1][2][3][4]. Apolipoprotein E4, a protein that binds and transports cholesterol and other lipids, has been identified as a risk factor for familial and sporadic AD (5,6). Patients on specific inhibitors of 3-hydroxy-3-methylglutaryl-coenzyme A reductase have a lower risk of developing AD as compared with individuals not taking those drugs (7,8). A metabolite of brain cholesterol, (24S)-hydroxycholesterol, was elevated in cerebral spinal fluid of AD patients compared with control subjects (9).
Amyloid ␤-peptide (A␤), the main component of neuritic plaques seen in brains of AD patients, interacts with cholesterol. This interaction is reciprocal. Cholesterol levels modulate amyloid precursor protein and A␤ synthesis (10 -13). Conversely A␤ alters cholesterol dynamics. Cholesterol modulates the actions of A␤ on brain membrane fluidity (14 -16). A␤ species (25-35, 1-40, and 1-42) increased the internalization of apolipoprotein E (apoE) complexed with cholesterol into neurons (17). A␤-  increased apoE levels in astrocytes (18). Cholesterol efflux from rat hippocampal neurons to cyclodextrin was enhanced by A␤- , and these results were attributed to redistribution of cholesterol to the plasma membrane (19). The Golgi complex plays an important role in cholesterol trafficking (20,21), and A␤ could act on cholesterol homeostasis in the Golgi complex. Chemical inhibitors known to act on the Golgi complex alter cholesterol trafficking (20). Furthermore it has been shown that phosphatidylcholine-phospholipase D (PC-PLD) contributes to regulation of cholesterol efflux and that PC-PLD is associated with the Golgi complex (20 -24). The A␤ peptide fragment 25-35 increased PLD activity in LA-N-2 cells (25). The effects of A␤ on cholesterol in the Golgi complex may also be dependent on whether A␤ is fresh or aged. Recently it was reported that oligomeric A␤-(1-40), but not monomeric A␤- , stimulated release of cholesterol, phospholipids, and GM1 ganglioside from neurons (26). Aged A␤-(1-40), but not fresh A␤- , preferentially binds cholesterol as compared with fatty acids or phosphatidylcholine (27) that might enhance removal of lipids from cell compartments. Fresh A␤-  was reported to be located in the hydrophobic area of synaptic plasma membrane, whereas A␤-(1-40) aged for 48 h was intercalated adjacent to the phospholipid polar head group region (28).
The current study tested the hypotheses that A␤-(1-42) modifies Golgi complex cholesterol homeostasis in astrocytes and that the effects of A␤-  were dependent on whether A␤-  was fresh or aged. Gel electrophoresis of fresh A␤-  and A␤-(1-42) aged for 48 h showed that aged A␤-  consisted of ϳ74% of the protein as a tetramer, whereas the fresh A␤-  showed only 11% of the protein as a tetramer (28), and such structural differences could alter the behavior of A␤. The effects of A␤-(1-42) on Golgi complex cholesterol could involve activity of PC-PLD, and this possible action of A␤-(1-42) was examined. In addition, it has been reported that apoE levels were increased in astrocytes treated with aged A␤-(1-42) (18), and such effects might alter cholesterol distribution in cells. ApoE levels also were determined. Experiments were conducted using fresh and aged A␤-  in immortalized DITNC1 astrocytes (29). Two different methods were used to determine the effects of A␤-(1-42) on Golgi cholesterol content of astrocytes: 1) confocal microscopy using the fluorescent cholesterol analogue NBD-cholesterol and a fluorescent marker for the Golgi complex, BODIPY TR ceramide (30); and 2) isolation of the Golgi complex into two different fractions thought to represent different regions of the Golgi complex using sucrose density gradient centrifugation (31,32). The results presented herein provide new insights into interaction of A␤ and cholesterol. We show that fresh A␤-(1-42) has a strikingly different effect on cholesterol levels in the Golgi complex of astrocytes as compared with aged A␤- . Incubation of astrocytes with fresh A␤-(1-42) significantly increased cholesterol levels, whereas aged A␤-(1-42) significantly reduced cholesterol levels in the Golgi complex as shown by fluorescent probes and confocal microscopy. The reverse peptide A␤-(42-1), either fresh or aged, did not alter Golgi complex cholesterol levels, which argues against a nonspecific action of A␤-(1-42) . Isolation of the astrocyte Golgi complex into different fractions revealed the novel finding that the effects of fresh A␤-(1-42) on cholesterol content were opposite on the two Golgi fractions, whereas aged A␤-(1-42) had effects similar to those observed with confocal microscopy. PC-PLD activity, cell membrane cholesterol, and apoE levels of astrocytes were associated with effects of fresh A␤-(1-42) on cholesterol levels but not with effects of aged A␤-(1-42), suggesting different mechanisms for actions of fresh and aged A␤-(1-42) on Golgi complex cholesterol. Our data suggest that extracellular A␤-(1-42) specifically targets the Golgi complex and disrupts cholesterol homeostasis.
Cell Culture-DITNC1 rat astrocytes were purchased from American Type Culture Collection (Manassas, VA). These cells have been shown to have the phenotypic characteristics of type 1 astrocytes (29). Astrocyte cells were incubated in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 5% glutamate for 2 days until confluence and maintained at 37°C in 5% CO 2 and at 90% relative humidity. The cells at 80 -85% confluence were treated with medium containing 1% lipoprotein-deficient serum replacing the 10% fetal bovine serum and incubated for 16 h prior to experimentation. All experiments were done using confluent astrocytes.
Incubation of Cells with Fluorescent Labeled Probes and A␤-(1-42)-Astrocytes were grown on Lab-Tek TM chamber slides, incubated with NBD-cholesterol (8 M) for 1 h, and then washed three times with 1 ml of phosphate-buffered saline (PBS). Fresh A␤-(1-42) (1 M) or A␤-(1-42) (1 M) preincubated for 48 h (aged) was added to cells and incubated for different time periods (30,60, and 120 min). We have reported previously that incubation of A␤-(1-42) for 48 h resulted in ϳ74% of the peptide in a tetrameric form (28). In some experiments the reverse peptide A␤-(42-1) (fresh or aged for 48 h) was incubated with cells. Following incubation with A␤, cells were rinsed three times with PBS, and the Golgi marker BODIPY TR ceramide (2 M) was added and incubated with the cells for 1 h. Cells were then washed with PBS, fixed with 4% paraformaldehyde, and mounted for confocal microscopy using Gel/Mount (Biomeda Corp., Foster City, CA).
Laser Scanning Confocal Microscopy-Confocal fluorescence imaging was performed on an Olympus Fluoview laser scanning confocal imaging system (Olympus America Inc., New York). Images were captured using multiple photomultiplier tubes regulated by Fluoview 2.0 software (Olympus). Excitation of the fluorescent probes was accomplished using 15-milliwatt krypton-argon lasers with 5-milliwatt output. An Olympus BX 50 fluorescent microscope was used to capture the images using an oil immersion objective. NBD-cholesterol was excited at 488 nm, and emission was recorded at 540 nm. BODIPY TR ceramide was excited at 568 nm, and emission was recorded at 598 nm. The captured images for the red and green channels were merged and appeared yellow, which is indicative of colocalization. Red and green are additive, generating yellow to orange in RGB (red-green-blue) color space (33). Quantitative analysis of the colocalization of NBD-cholesterol and BODIPY TR ceramide was determined by image processing using MetaMorph imaging system V4.3 from Universal Imaging Corp. (Downingtown, PA) as described previously (33,34) and expressed as percent colocalization.
Isolation of Golgi Fractions-Isolation of the Golgi complex was accomplished using sucrose density gradient centrifugation (32). Confluent astrocytes were treated with fresh and aged A␤-(1-42) as described above. Cells were harvested and washed twice with PBS. The cells were suspended in G buffer (10 mM Tris-HCl, 0.25 M sucrose, 2 mM MgCl 2 , pH 7.4) containing 10 mM CaCl 2 . Leupeptin and phenylmethylsulfonyl fluoride were added to the cell suspension to inhibit proteolytic enzymes prior to homogenization. The cells were homogenized in a Potter-type homogenizer with 20 -30 strokes using a serrated homogenizing pestle. The homogenate was centrifuged at 5,000 ϫ g for 10 min. The pellet was suspended and homogenized in 1.4 M sucrose and overlayered with 0.8, 1.0, and 1.2 M sucrose gradients in ultracentrifuge tubes. Samples were then centrifuged at 95,000 ϫ g for 2.5 h in a SW28 rotor in a Beckman L8 -70 ultracentrifuge. Each interface was carefully removed, diluted with G buffer, and centrifuged in an SS34 rotor at 45,900 ϫ g for 20 min. Each pellet was suspended in buffer and used for the experiments. Protein concentrations of the fractions were determined using the Bradford assay. Cholesterol content in each fraction was determined enzymatically in a microassay using a diagnostic kit from Roche Molecular Biochemicals (35) and procedures reported by our laboratory (36 -38).
Galactosyltransferase Activity-A marker enzyme of the Golgi complex is galactosyltransferase, and activity of this enzyme was measured using procedures reported previously (32,39). The incubation mixture contained 7 mg/ml ovalbumin, 2 mM ATP, 200 mM MgCl 2 , 0.2% Triton X-100, and 50 mM Tris-HCl (pH 6.8). The reaction was initiated by adding [ 3 H]UDP-galactose to samples and incubated for 30 min, and the reaction was stopped by adding ice-cold 24% trichloroacetic acid. The precipitate was pelleted at 20,800 ϫ g and washed three times with 12% trichloroacetic acid. Samples were solubilized in 5% SDS, and disintegrations/min were counted using a scintillation counter.
PC-PLD Activity-PC-PLD activity of control astrocytes and astrocytes that had been incubated with either fresh or aged A␤-(1-42) for different time periods (30,60,90, and 120 min) was quantified using the Amplex Red PC-PLD assay kit (Molecular Probes). The intensity of the Amplex Red complex formed was measured in a fluorescence microplate reader (Spectra Max Gemini XS, Molecular Devices, Sunnyvale, CA) using an excitation wavelength of 542 nm and an emission wavelength of 590 nm.
Measurement of A␤-(1-42)-Quantitative determination of A␤-(1-42) in the G1 and G2 fractions was carried out using an enzyme-linked immunosorbent assay kit from Immuno-Biological Laboratories. This kit uses a solid phase sandwich enzyme-linked immunosorbent assay protocol with anti-human A␤-(38 -42) rabbit IgG as primary antibody in a precoated plate and horseradish peroxidase-conjugated anti-human A␤-(11-28) mouse IgG as secondary conjugated antibody. Briefly, standards and test samples were diluted in a buffer consisting of 1% bovine serum albumin and 0.05% Tween 20 in PBS. Test samples as well as sample and reagent blanks were added to the precoated plate and incubated overnight at 4°C. After binding, the wells were washed (seven times) prior to adding conjugated secondary antibody and incubated for 1 h at 4°C. After washing (nine times), the chromogen containing tetramethylbenzidine was added and incubated for 30 min at room temperature. Color reaction was terminated by the stop solution, and measurement was performed at 450 nm using a Packard Fusion AlphaScreen microplate analyzer (Packard Instrument Co.). In each assay, a standard curve was constructed after subtracting absorbance of the test sample blank. Data for test samples were converted to ng of A␤-(1-42)/mg of protein.
ApoE Isolation-Media of control cells and cells treated with either fresh or aged A␤-(1-42) were removed, and phenylmethylsulfonyl fluoride was added as a protease inhibitor. The media were centrifuged at 800 ϫ g for 5 min, and the resulting supernatant was treated with 15% trichloric acid to precipitate protein, washed with cold methanol, dried, and kept at Ϫ20°C for later use. Cells were harvested with trypsin/ EDTA, and protein values were determined using the Bradford method with ovalbumin standards. Cells were treated with lysing buffer containing 8 M urea, 5% SDS, and 5% ␤-mercaptoethanol. For Western blot analysis of lysates and conditioned media, 1 l of 1 M dithiothreitol and 1 l of 0.2% bromphenol solution was added to every 50 l of lysing buffer before use, and samples were kept at 70°C for 30 min. 100 -200 g of protein were electrophoresed on a 7.5% SDS, Tricine, HCl precast gel (Bio-Rad), and proteins were transferred to a nitrocellulose membrane and incubated with monoclonal mouse anti-rat apoE (1:250) from BD Laboratory (San Diego, CA). Horseradish peroxidase-conjugated goat anti-mouse IgG (1:20,000) was used as a secondary antibody (Pierce). Immunoreactivity was visualized with SuperSignal West Pico Chemiluminescent substrate (Pierce). Band density was quantitated by densitometry using the Eagle Eye II video system and EagleSight software (Stratagene, La Jolla, CA).

A␤-(1-42) Modifies Cholesterol Distribution in the Golgi Complex as Revealed by Confocal Microscopy-Experiments in
this study tested the hypothesis that A␤-(1-42) would modify cholesterol distribution in the Golgi complex of astrocytes and that the effects of A␤-(1-42) on cholesterol distribution would be dependent on whether A␤-(1-42) was fresh or aged. Two different methods were used to determine cholesterol distribution in the Golgi complex. 1) A fluorescent cholesterol analogue, NBD-cholesterol, and a fluorescent marker for the Golgi complex, BODIPY TR ceramide, were used and imaged with confocal microscopy; and 2) sucrose density gradient centrifugation was used to isolate two different fractions of the Golgi complex, and cholesterol levels were determined enzymatically.
BODIPY TR ceramide, NBD-cholesterol, and confocal microscopy were used to initially examine the effects of fresh and aged A␤-(1-42) on Golgi cholesterol content of astrocytes. Cells were incubated with A␤-(1-42) for different time periods (30,60, and 120 min). Cell viability as measured by trypan blue was not altered by incubation of cells with fresh or aged A␤-(1-42) for 120 min. Means Ϯ S.E. of the percentage of viable cells for control astrocytes, astrocytes incubated with fresh A␤-(1-42), and astrocytes incubated with aged A␤-(1-42) were 96.83 Ϯ 0.58, 94.2 Ϯ 1.5, and 97.83 Ϯ 0.6, respectively. Confocal images of colocalization of NBD-cholesterol and BODIPY TR ceramide in control astrocytes and astrocytes incubated with fresh and aged A␤-(1-42) for 120 min are shown in Fig. 1. Panel C shows colocalization of the two fluorescent probes in control astrocytes. It can be seen in panel F that colocalization was greater in astrocytes treated with fresh A␤-(1-42), whereas aged A␤-(1-42) (panel I) reduced colocalization as compared with control astrocytes and astrocytes incubated with fresh A␤-(1-42). Analysis of colocalization data using MetaMorph software and expressed as percent colocalization shows that fresh A␤-(1-42) (p Յ 0.001) increased colocalization in the Golgi complex of astrocytes as compared with control astrocytes (38 versus 23%) at the 120-min incubation period (Fig. 2). The percent colocalization of the fluorescent probes was significantly (p Յ 0.01) less when astrocytes were incubated with aged A␤-(1-42) (15 versus 23%) for 120 min as compared with control astrocytes (Fig. 2). Shorter incubation times with A␤-(1-42) did not significantly alter colocalization, although it can be seen in Fig. 2  To determine whether the effects of A␤-(1-42) were specific and not attributable to a nonspecific perturbation of the Golgi complex we incubated astrocytes for 120 min with the reverse peptide A␤-(42-1) that was either fresh or aged for 48 h. Neither fresh nor aged A␤-(42-1) had an effect on the percent colocalization of BODIPY TR ceramide and NBD-cholesterol when compared with control astrocytes or with each other (Fig. 2).
UDP-galactosyltransferase Activity and Fluorescence of BODIPY TR Ceramide in Golgi Complex Fractions-Confocal microscopy revealed that A␤-(1-42) altered the colocalization of NBD-cholesterol and BODIPY TR ceramide, a marker for the Golgi complex, and these data were interpreted as A␤-  altering Golgi cholesterol levels. The Golgi complex is heterogenous in structure and function and may not be equally affected by A␤- . To further define the effects of fresh and aged A␤-(1-42) on cholesterol distribution in the Golgi complex of astrocytes, the Golgi complex was isolated using sucrose density centrifugation, and two different fractions were obtained that are thought to represent different regions of the Golgi complex (32): 1) the light Golgi fraction (G1), which is the band at the 0.8 -1.0 sucrose gradient interface; and 2) the Golgi fraction 2 (G2), which is the band at the 1.0 -1.2 sucrose gradient interface. Enrichment of UDP-galactosyltransferase activity and fluorescence of BODIPY TR ceramide have been used previously as markers for the Golgi complex (30,32,40,41), and these markers were used in the present study. The means Ϯ S.E. of UDP-galactosyltransferase activity (nmol/30 min/mg of protein) in astrocyte homogenate, G1 fraction, and G2 fraction were 0.019 Ϯ 0.004, 2.751 Ϯ 0.086, and 1.972 Ϯ 0.129, respectively. UDP-galactosyltransferase activity was significantly higher (p Ͻ 0.01) in the G1 fraction as compared with the G2 fraction and was several orders of magnitude higher in both the G1 and G2 fractions in contrast to the homogenate. Enrichment of UDP-galactosyltransferase activity in the G1 fraction as compared with the G2 fraction has been reported previously (32). BODIPY TR ceramide, a fluorescent Golgi complex marker, had a similar distribution in the G1 and G2 astrocyte fractions as was observed for UDP-galactosyltransferase activity (data not shown).

A␤-(1-42) Modifies Cholesterol Distribution in the Golgi Complex Fractions as Revealed by Sucrose Density
Centrifugation-We next examined the effects of A␤-(1-42) on cholesterol content of the G1 and G2 fractions. The effects of fresh and aged A␤-(1-42) on cholesterol in the G1 fraction were similar to effects we observed using confocal microscopy. Fresh A␤-(1-42) significantly (p Ͻ 0.0001) increased cholesterol, whereas aged A␤-(1-42) significantly (p Ͻ 0.001) reduced cholesterol in the G1 fraction (Fig. 3). The magnitude of effects of fresh A␤-(1-42) on cholesterol in the G1 fraction was greater than that of aged A␤-(1-42), and these results were similar to those observed with confocal microscopy. Fig. 3 shows that fresh A␤-(1-42) had an opposite effect on cholesterol in the G2 fraction as compared with the G1 fraction. There was a significant (p Ͻ 0.001) reduction in cholesterol in the G2 fraction when incubated with fresh A␤-(1-42). On the other hand, aged A␤-(1-42) had a similar effect on cholesterol in the G2 fraction as observed in the G1 fraction, resulting in a significant (p Ͻ 0.001) reduction in cholesterol (Fig. 3). Differences in effects of fresh and aged A␤-(1-42) on cholesterol levels in the Golgi complex could be attributed to differences in amounts of fresh and aged A␤-(1-42) that reach the Golgi complex. Levels of fresh A␤-(1-42) were higher in the G1 and G2 fractions (0.21 Ϯ 0.03 and 0.22 Ϯ 0.1 ng/mg of protein, respectively) as compared with aged A␤-  in the G1 and G2 fractions (0.11 Ϯ 0.05 and 0.16 Ϯ 0.04 ng/mg of protein, respectively). These differences in A␤-(1-42) levels, however, were not statistically significant.
Both colocalization data and data of the G1 fraction showed that fresh A␤-(1-42) increased cholesterol and aged A␤-(1-42) reduced cholesterol in the Golgi complex of astrocytes. The cell membrane is enriched in cholesterol and could be a source of cholesterol in cells treated with fresh A␤-(1-42). Table I shows that fresh A␤-(1-42) significantly reduced both total and free cholesterol in the cell membrane as compared with control astrocytes. Total cholesterol was reduced by 34% and free cholesterol by 30% when treated with fresh A␤-(1-42) . Aged A␤-(1-42) also significantly reduced total and free cholesterol (17 and 11%, respectively) in the cell membrane (Table I), but effects were much less than those observed for fresh A␤-(1-42) . The ratio of free cholesterol to total cholesterol was used as an estimate of esterified cholesterol. Data in Table I show that the ratios were similar among the three groups and did not differ significantly. It also was noted that cholesterol content of the astrocyte homogenate fraction did not differ among the three groups.
A␤-  Inhibits PC-PLD Activity-PC-PLD has been shown to be involved in regulating cholesterol efflux from cells (42) and in release of secretory vesicles from the Golgi complex (23,43), and an isoform of PC-PLD has been reported to be associated with the Golgi complex (20,22). To begin to address a potential mechanism whereby A␤-(1-42) alters cholesterol content in the Golgi complex, we examined the effects of fresh and aged A␤-(1-42) on activity of PC-PLD in astrocytes. Fig. 4  (panels A and B) reveals that fresh and aged A␤-(1-42) significantly (p Ͻ 0.001) inhibited PC-PLD activity at each incubation time period. There was a small diminution of effects of A␤-(1-42) on PC-PLD activity with increasing incubation time. PC-PLD activity was significantly (p Ͻ 0.001) higher after 90 and 120 min of incubation versus 30 min of incubation for fresh A␤-(1-42) and after 120 min for aged A␤-(1-42) (Fig. 4, panels  A and B). The effects of fresh A␤-(1-42) on PC-PLD activity were consistent with the increase in cholesterol in the Golgi complex, but PC-PLD activity does not explain the effects of aged A␤-(1-42) on Golgi complex cholesterol.
Effects of Fresh and Aged A␤-  on ApoE Levels in Astrocytes-It has been reported previously that A␤-  increased astrocyte activation, increased levels of apoE and apoJ in the cell, and reduced levels of the two apolipoproteins in the medium (18). The effects of activation were proposed to primarily involve apoE and its receptors. Changes in A␤-(1-42)-induced levels of apoE and release of apoE could have an effect on cholesterol content by sequestering cholesterol in the cell. Fig.  5 (inset) shows that apoE under reducing conditions can be seen as approximately a 34-kDa monomer, and this finding is consistent with apoE expression in astrocytes (18). Fresh A␤-(1-42) significantly (p Յ 0.01) increased apoE levels by 32% in astrocytes as compared with control cells (Fig. 5). Aged A␤-(1-42) appeared to reduce apoE levels, but this difference was not statistically significant when compared with control cells. ApoE was not detected in the conditioned media from control and A␤-(1-42)-treated cells (data not shown).

DISCUSSION
Several different lines of evidence show that A␤ and cholesterol interact (for reviews, see Refs. 4 and 44 -47). This interaction is reciprocal given that cholesterol modulates expression of amyloid precursor protein and A␤ (10 -13) and cholesterol dynamics such as cholesterol trafficking are affected by A␤ (15,(17)(18)(19). The Golgi complex plays an important role in cholesterol trafficking (21,48), and we proposed that A␤-(1-42) would alter cholesterol trafficking in the Golgi complex and that the effects of A␤-(1-42) would be dependent on whether the peptide was fresh or aged. Two different methods, confocal microscopy and isolation of the Golgi complex by sucrose density centrifugation, were used to test the hypothesis that fresh and aged A␤-(1-42) would have different effects on cholesterol in the Golgi complex of astrocytes. Results of this study showed   Fresh A␤-(1-42) increased astrocyte cholesterol in the G1 fraction by 59% but reduced cholesterol in the G2 fraction by 29%. Earlier work has indicated that the G1 fraction may represent the trans-Golgi region and that the G2 fraction is representative of the medial and cis regions (31,32). Fresh A␤-(1-42) may stimulate movement of cholesterol from the cis and medial regions of the Golgi to the trans-Golgi region. Mechanisms of anterograde intra-Golgi transport have focused on cisternal maturation/progression and vesicular/tubular transport, but the principal mechanism of intra-Golgi transport including transport of lipids has not been resolved (for a review, see Ref. 49). Movement of cholesterol from the G2 to the G1 fraction induced by fresh A␤-(1-42) would account for only 29% of the increase in cholesterol observed in the G1 fraction. Our data showing that there was a significant reduction in cholesterol content in the cell membrane of astrocytes treated with fresh A␤-(1-42) (34%) as compared with control astrocytes suggest that fresh A␤-(1-42) stimulates transport of cholesterol from the cell membrane to the trans-Golgi region. There is evidence that cycling of lipids including cholesterol and proteins occurs between the cell surface membrane and the Golgi complex that is clathrin-independent (50,51). It has been suggested that this cell membrane-Golgi pathway may regulate lipid raft distribution and function with recycling occurring continuously (51). Amyloid precursor protein and A␤ are associated with cholesterol-rich low density membrane domains (52)(53)(54)(55). Extracellular A␤ that is not highly aggregated may target lipid rafts, promoting cycling of lipid rafts to the Golgi complex.
Aged A␤-(1-42) reduced cholesterol levels in the Golgi complex but also reduced cholesterol levels in the cell membrane by 16% as compared with the 34% reduction by fresh A␤- . The destinations and mechanisms of the cell membrane cholesterol removed by fresh and aged A␤-(1-42) may differ. Aged A␤-(1-42) may complex with cholesterol in the cell membrane and transport cholesterol to the cell interior. Support for this notion are data showing that indeed aged A␤-(1-40), but not fresh A␤- , complexed with lipids with cholesterol Ͼ Ͼ stearic acid Ͼ PC (27). Once removed, the aged A␤-(1-42)cholesterol complex may act as a type of lipid droplet in the cell cytosol. Synucleins that had been chemically cross-linked formed oligomers within cells that complexed with lipid droplets and cell membranes (56). It is clear that fresh A␤-(1-42) and aged A␤-(1-42) are acting differently on cholesterol homeostasis.
Cholesterol levels were increased in the Golgi complex by incubation with fresh A␤- , and this increase may be due in part to cycling of cholesterol from the cell membrane to the Golgi complex. We also propose that the elevation in cholesterol levels in the Golgi complex involves PC-PLD function. PC-PLD plays a role in cholesterol trafficking, it is located in the Golgi complex, and PC-PLD is also involved in vesicular trafficking from the Golgi complex (20 -23). Our findings that fresh A␤-(1-42) increased cholesterol content in the Golgi complex and inhibited activity of PC-PLD are in agreement with the purported role of PC-PLD in cholesterol trafficking. Inhibition of PC-PLD reduces cholesterol efflux (22,42). Activation of PC-PLD enhanced the release of secretory vesicles from the trans-Golgi complex (43). Inhibition of PC-PLD by fresh A␤-(1-42) may in turn reduce release of secretory vesicles with their lipid and protein cargoes. We found that fresh A␤-(1-42) inhibited PC-PLD activity, but another study reported that the A␤ peptide fragment 25-35 increased PLD activity in LA-N-2 cells (57). Differences in effects of A␤ on PLD activity between that study and the present study may have resulted from dissimilarities in peptide structure and peptide concentration (A␤-(1-42) at 1 M versus A␤- (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) at 100 M). Inhibition of PC-PLD activity by fresh A␤-(1-42) was in agreement with our findings that cholesterol levels were elevated in the Golgi complex in the presence of fresh A␤- . However, we also observed that aged A␤-(1-42) significantly inhibited PC-PLD activity but that cholesterol levels were reduced in the Golgi complex of astrocytes incubated with aged A␤- . Both fresh and aged A␤-(1-42) inhibit PC-PLD activity, but the mechanism of effects on cholesterol levels in the Golgi complex may differ as discussed in the preceding paragraph. cantly increased apoE levels in astrocytes by ϳ32% as compared with control cells. On the other hand, aged A␤-(1-42) reduced apoE levels, but this reduction did not differ significantly from that seen in control cells. Previously it was found that aged A␤-(1-42) increased apoE levels in rat astrocytes from primary cell culture (18). In the same study it also was reported that aged A␤-(1-42) reduced levels of apoE in the medium (18). We did not detect apoE in the media of control astrocytes and cells incubated with either fresh or aged A␤- . Two major differences between our study and the earlier report were the incubation time of 2 versus 12 h and the A␤-(1-42) concentration of 1 versus 10 M, respectively. An important observation of the earlier study was that A␤-(1-42) did not alter steady-state levels of apoE mRNA, and that may indicate that A␤-(1-42) was not acting on transcriptional regulation but could have been acting post-translationally on apoE turnover (18). Whether the observed increase in apoE levels in astrocytes incubated with fresh A␤-(1-42) plays a role in effects of the peptide on cholesterol in the Golgi complex is unclear. ApoE is associated with the Golgi complex (58), and changes in apoE levels may modify intracellular cholesterol distribution.
We have shown using two different methods that A␤-(1-42) modifies the cholesterol content of the Golgi complex of astrocytes. The effects of A␤-(1-42) on cholesterol levels were dependent on whether astrocytes were treated with fresh or aged A␤- . Fresh A␤-(1-42) increased cholesterol levels of the Golgi complex, whereas aged A␤-(1-42) reduced cholesterol levels. However, the direction of effects of fresh A␤-(1-42) on Golgi cholesterol levels was contingent on the region of the Golgi complex suggesting that fresh A␤-(1-42) may also be acting on intra-Golgi cholesterol transport. The Golgi complex has been described as a major site for protein and lipid sorting and release of lipid rafts (59,60). Either increasing or decreasing cholesterol levels could impact on important Golgi complex functions such as protein sorting, sphingomyelin synthesis, and lipid trafficking resulting in cellular pathophysiology. Our data suggest that the recycling of cholesterol between the cell membrane and the Golgi complex was altered and that such changes could modify membrane structure and function. Fresh A␤ may inhibit the formation of lipid rafts in the Golgi complex including caveolin-organized rafts that are transported to the plasma membrane and form caveolae. A␤ peptides are located in detergent-insoluble membrane domains that are thought to be caveolae (54,61) and in cholesterol domains not containing caveolin (52). PC-PLD activity, cell membrane cholesterol, and apoE expression of astrocytes were associated with effects of fresh A␤-(1-42) on Golgi complex cholesterol but not with effects of aged A␤-(1-42), arguing against a common mechanism. There is a reciprocal synergy between A␤ and cholesterol. Changes in cholesterol content alter A␤ levels, and A␤ acts on cholesterol homeostasis. The dynamic interaction of A␤ and cholesterol that includes targeting of the Golgi complex may be an important factor that contributes to cellular dysfunction occurring with AD.