| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Central Clinical Laboratories, Shinshu University Hospital, 3-1-1 Asahi, Matsumoto 390-8621, Japan.
2
Department of Functional Polymer Science, Faculty of
Textile Science and Technology, Shinshu University, 3-15-1 Tokita, Ueda
386-0018, Japan.
3
Department of Laboratory Medicine, Shinshu University
School of Medicine, 3-1-1 Asahi, Matsumoto 390-8621, Japan.
a Author for correspondence. Fax 81-263-34-5316; e-mail mtozuka{at}hsp.md.shinshu-u.ac.jp
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: CSF lipoproteins were fractionated by gel filtration and ultracentrifugation, and then characterized by electrophoresis, immunoblot, electron microscopy, and analysis of apoE, total cholesterol, and phospholipid concentrations.
Results: The ratio of sialylated to nonsialylated apoE was higher in CSF than in serum. However, the fundamental forms containing apoE homodimers or heterodimers [such as apo(E-AII) and apo(AII-E2-AII) complexes] were similar in CSF and serum. apoE-containing lipoproteins were fractionated at densities of <1.006, 1.063–1.125, and 1.125–1.21 kg/L. Neither apoE nor apoAI was detected in the fraction with a density range of 1.006–1.063 kg/L. The diameters of the lipoprotein particles with densities of <1.006, 1.063–1.125, and 1.125–1.21 kg/L were 16.7 ± 3.1, 14.0 ± 3.2, and 11.6 ± 2.8 nm (mean ± SD, n = 200), respectively. All of these lipoproteins exhibited a spherical structure. The distribution profile of apoE-containing lipoproteins was affected by the apoE phenotype. A relatively large amount of apoE-containing lipoproteins was isolated from the fraction with a density >1.125 kg/L obtained from CSF associated with apoE2 or apoE3. This tendency was more obvious in CSF associated with apoE2 than in CSF without apoE2. apoE-containing lipoproteins were predominantly observed in the fraction with a density of <1.006 kg/L obtained from CSF associated with apoE4.
Conclusions: The lipoproteins in CSF have a unique composition that is different from that of the lipoproteins in plasma. However, the differences in diameter between the CSF fractions were not as large as for the serum fractions. Our data suggest that the apoE phenotype may affect the distribution profile of apoE-containing lipoproteins in the CSF. This would mean that the metabolism of apoE-containing lipoproteins depends on the apoE isoform present.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Multiple lines of evidence suggest that CSF apoE is implicated in the pathogenesis of Alzheimer disease (AD) (12)(13)(14)(15)(16)(17); however, the precise role of apoE in the development of AD is still obscure. Characterization of apoE and apoE-containing lipoproteins in the CNS is essential to clarify its involvement in AD. Recently, CSF lipoproteins have been investigated by several groups, and their characterization has been progressing gradually (9)(10)(18). However, the nature of CSF lipoproteins and their metabolic pathways are still not as clear as those of the plasma lipoproteins. One of the interesting characteristics of the plasma apoE isoforms is the preferential association of apoE3 and apoE4 with particular classes of plasma lipoproteins (apoE3 with HDL and apoE4 with triglyceride-rich lipoproteins such as VLDL) (19)(20)(21).
In this study, we characterized the apoE-containing lipoproteins in the CSF and examined the effect of the apoE phenotype on the distribution of apoE among the CSF lipoprotein fractions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
immunoblot analysis
Samples were treated with Laemmli buffer with or without
2-mercaptoethanol (22) and were loaded on 8–16%
polyacrylamide gradient gels. The separated proteins were
electrophoretically transferred onto nitrocellulose membranes (pore
size, 0.45 µm; Advantek Toyo), which were then incubated with
a blocking buffer (50 mmol/L Tris-HCl, pH 8.0, containing 20 g/L
skim milk) for 30 min at room temperature. After being washed three
times, the membranes were incubated with the primary antibodies
[rabbit anti-apoE (Dako) and mouse anti-apoAII (ICN
Pharmaceuticals)], and then washed three more times. The membranes
were incubated with horseradish peroxidase-conjugated anti-rabbit or
-mouse IgG (goat; MBL) for 1 h at room temperature. After
washing, the bands were visualized either using 3,3'-diaminobenzidine
tetrahydrochloride (Dojin Chemical) and hydrogen peroxide (Wako
Pure Chemicals) or by means of an enhanced chemiluminescence detection
kit (Amersham Life Sciences).
isoelectric focusing
Sample preparation.
Serum (2 µL) or CSF (15 µL) was
incubated with 5 µL of neuraminidase (20 kU/L) in 0.1 mol/L
citric acid buffer containing 10 g/L Tween 20 for 12 h at
37 °C. This was followed by treatment with 5 µL of 75 mmol/L
dithiothreitol (Wako Pure Chemicals) in 10 g/L Tween 20 for 1 h at
room temperature.
Electrophoresis.
Isoelectric focusing was carried out as
described previously (23). Briefly, 20 µL of the prepared
sample was electrophoresed on a 4.8% polyacrylamide gel containing 8
mol/L urea and 20 g/L Ampholine® (pH 4–6;
Pharmacia Biotech) using 3.3 mmol/L phosphoric acid as the anode
solution and 20 mmol/L NaOH as the cathode solution. Electrophoresis
was carried out overnight at 4 °C under constant voltage (200 V).
After electrophoresis, immunoblot analysis was performed as described
above.
lipoprotein electrophoresis
CSF (20 µL) was applied to agarose gel (Titan Gel Lipo Kit;
Helena Laboratories) and electrophoresed for 20 min under constant
voltage (90 V). The separated lipoproteins were visualized by the
following two method: (a) lipids were stained directly by
Fat Red 7B (Helena Laboratories) in methanol, and (b)
separated lipoproteins were transferred onto a nitrocellulose membrane
by capillary blotting, and apoE and apoAI were developed by the
immunoblot method described above.
fast protein liquid chromatography
CSF (500 µL) was applied to a 130-cm
Superose® 6 column (Pharmacia Biotech)
equilibrated with phosphate-buffered saline, pH 7.4. Lipoprotein
fractions were eluted at a flow rate of 0.5 mL/min.
ultracentrifugation
CSF lipoproteins were isolated by the ultracentrifugation method
described by David et al. (24) with a small modification.
CSF was centrifuged at 541 000g rpm for 1 h
using an OptimaTM TLX Ultracentrifuge (Beckman) to isolate the fraction
with a density of <1.006 kg/L. Solid KBr (Wako) was sequentially added
to CSF after each centrifugation to adjust the density to 1.006, 1.063,
1.125, and 1.21 kg/L in sequence. The isolated fractions were dialyzed
against phosphate-buffered saline.
determination of apoE, TOTAL CHOLESTEROL, AND
PHOSPHOLIPIDS IN CSF
CSF apoE was measured using an ELISA system as described
previously (25). Total cholesterol (TC) and phospholipid
(PL) in CSF were measured, respectively, by the cholesterol oxidase
method (Kyowa Medex) and the choline oxidase method (Wako), using a
Hitachi 7170 automated analyzer with a higher ratio of sample volume to
reagent volume than that is routinely used for serum samples.
electron microscopy
The lipoprotein particles in each fraction isolated by
ultracentrifugation were examined under a JEOL JEM1010 electron
microscope. Samples were negatively stained with 20 g/L aqueous
uranyl acetate. Particle diameters, measured by means of a micrometer
in each photograph, were expressed as the mean ± SD nm for a
random sample of 200 particles.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The apoE monomer, apoE homodimer, and the apo(E-AII) complex
were identified in the CSF obtained from the subject with the apoE
phenotype E3/E3. In addition, the apo(AII-E2-AII) complex was also
recognized in the CSF from the subject with the apoE phenotype E3/E2.
These complexes were identified by reducing and nonreducing SDS-PAGE
followed by immunoblotting using anti-apoAII and anti-apoE antibodies
(data not shown).
|
Bands of slightly higher molecular size than that of the apoE monomer,
which disappeared after treatment with neuraminidase, were more intense
in CSF than in serum, which indicates that the apoE monomer is more
highly sialylated in the CSF than in serum. As shown in Fig. 2
, isoelectric focusing confirmed that CSF apoE monomers existed
not only in the monosialylated form but also in the di-, tri-, or
higher sialylated forms.
|
characterization of apoE-CONTAINING LIPOPROTEINS
Lipoprotein electrophoresis.
CSF lipoproteins migrated to the
1 and pre-1 position
in agarose gel electrophoresis (Fig. 3
A). After treatment with neuraminidase, the faint
pre-1 band was shifted to the
1 position (data not shown). When the gel was
immunoblotted with anti-apoE antibody, a major band in
1 region was observed with weaker reactivity
extending into the pre-1 region (Fig. 3B
). A
band that reacted with anti-apoAI antibody was observed almost
exclusively in the pre-1 region (Fig. 3C
).
|
Ultracentrifugal analysis.
The fractions isolated by
ultracentrifugation were applied to SDS-PAGE using an 8–16% gradient
gel followed by immunoblotting for apoE and apoAI (Fig. 4
). Both apolipoproteins were contained in each of the fractions
with densities of <1.006, 1.063–1.125, and 1.125–1.21 kg/L but not
in the fraction with a density of 1.006–1.063 kg/L. All of the former
three fractions isolated from CSF associated with apoE2 or apoE3
contained not only the apoE monomer but also apoE homo- and
heterodimers. PAGE analysis for lipoproteins showed that the molecular
size extended from 150 kDa to 600 kDa and also suggested that three
kinds of lipoprotein particles (containing predominantly apoE, apoAI,
or both apoE and apoAI, respectively) were present in each fraction
(Fig. 5
). In the fraction with a density of <1.006 kg/L, the majority
of apoAI appeared as a monomer at molecular size of 28 kDa. No
apoB was detected in any fractions (data not shown).
|
|
Gel filtration analysis.
As shown in Fig. 6
A, the main lipoproteins in the CSF samples separated by fast
protein liquid chromatography were larger than serum HDL and were
eluted at a mid-position between serum HDL and LDL (fractions 26–36).
The CSF lipoproteins contained much more PL than TC; therefore, the
lipid composition was similar to that of the serum HDL fraction. These
fractions were applied to PAGE using 8–16% gradient gel followed by
immunoblotting for apoE and apoAI (Fig. 6B
). The strongest
immunoreactivity for apoE and apoAI was observed in fractions 28–30
and 30–32, respectively. In addition to the apoE monomer, the apoE
homodimer and a small amount of the apo(E-AII) complex were detected in
fractions 28–30. The lipoproteins with densities of <1.006,
1.063–1.125, and 1.125–1.21 kg/L isolated by ultracentrifugation were
eluted in almost the same fractions by fast protein liquid
chromatography (data not shown).
|
Electron microscopy.
As shown by negative-staining electron
microscopy, the diameters of those lipoprotein particles with densities
of <1.006, 1.063–1.125, and 1.125–1.21 kg/L that were isolated from
CSF by ultracentrifugation were 16.7 ± 3.1, 14.0 ± 3.2, and
11.6 ± 2.8 nm (mean ± SD, n = 200), respectively (Fig. 7
). The differences in particle size among these fractions were
significant; however, the differences were smaller than in the case of
serum. In addition, two or more peaks were observed in the distribution
of particle sizes for each of the three fractions.
|
effect of apoE PHENOTYPE ON THE DISTRIBUTION OFapoE-CONTAINING LIPOPROTEINS
The apoE, TC, and PL concentrations for three different pools of
CSF for each of the three apoE phenotypes apoE3/E2, apoE3/E3, and
apoE4/E3 are shown in Table 1
. These CSFs were fractionated by ultracentrifugation, and the
distribution ratios for apoE, TC, and PL were determined for the five
fractions with different densities (Table 1
). The total recoveries for
apoE were 75.6–91.0%. In CSF associated with apoE2 or apoE3, the
amount of apoE recovered in the fractions with densities of
1.063–1.125 and 1.125–1.21 kg/L was much greater than in the fraction
with a density of <1.006 kg/L. In contrast, a higher ratio of
apoE-containing lipoproteins was recovered in the fraction with a
density of <1.006 kg/L than in the other two fractions for the CSF
associated with apoE4. In addition, in the subject with an E4/E3
phenotype, apoE also had a tendency to be recovered in the bottom
fraction (d >1.21 kg/L), with a higher ratio than that
observed for the other two phenotypes. The lipid compositions,
expressed as the ratio of PL/TC, were ~2.3 for the fraction with a
density of <1.006 kg/L and ~1.0 for the fractions with densities of
1.063–1.125 and 1.125–1.21 kg/L.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As reported previously (5), the CSF apoE monomer was more sialylated than the serum apoE monomer. In agarose gel electrophoresis, CSF lipoproteins showed a faint band that migrated faster than that of HDL in serum. After treatment with neuraminidase, the faint band moved to the same position as the serum HDL (data not shown). This implies that apoE affected the electrophoretic migration of the main lipoprotein in CSF. In other words, apoE is one of the main apolipoprotein components of the lipoproteins in CSF. The lipid composition of the CSF lipoproteins, in which phospholipid is more abundant than cholesterol (25), might also influence the electrical charge. It is known that recently secreted apoE is more sialylated than mature apoE (5). This, together with our results, suggests that almost all of the apoE in the CNS is in a more immature form than plasma apoE and that the sialylated form of apoE may be physiologically important in the CNS. However, the fundamental forms of apoE in CSF seemed to be almost the same as in plasma. Not only the apoE monomer but also the apoE homodimer (28) and heterodimers such as apo(E-AII) (29) and apo(AII-E2-AII) complexes (23), were observed in the CSF from subjects with the apoE phenotypes E3/E3 and E3/E2. Unlike apoE, the apoAII in the CSF appears to be derived from the plasma (4). Therefore, the quantity of apo(E-AII) and apo(AII-E2-AII) complex in the CSF would be regulated by the quantity of apoAII passing through the blood-brain barrier. Actually, variations between individual cases were observed in the immunoreactive intensity of the apo(E-AII) complex visualized by immunoblot analysis of CSF (data not shown). This implies that the apoAII concentration in the CSF may vary among individuals.
apoE-containing lipoproteins in the CSF may be microheterogeneous in size, as suggested by LaDu et al. (9) and Guyton et al. (18). However, our result differed from those of the above groups in two ways. First, the novel large apoE-containing lipoprotein with a density of 1.006–1.060 kg/L, which was isolated by Guyton et al. using density gradient ultracentrifugation, (18) seemed to us to actually be a lower density lipoprotein with a density of <1.006 kg/L or close to 1.006 kg/L. The fraction isolated under the natural density of CSF (<1.006 kg/L) contained both apoE and lipid, but we could not identify these components in the fraction with a density of 1.006–1.063 kg/L. Second, previous data (9)(18) suggested that apoE-containing lipoproteins exist as large particles, whereas apoAI and apoAII exist as smaller lipoprotein particles. However, our data indicate that the large lipoproteins contained both apoE and apoAI. In addition, apo(E-AII) and apo(AII-E2-AII) complexes were detected in those lipoproteins. As demonstrated previously by Pitas et al. (7) and Borghini et al. (10), three kinds of lipoproteins predominantly containing apoAI, apoE, or both apoAI and apoE, respectively, were observed in the present study in all subfractions obtained by ultracentrifugation except for that with a density of 1.006–1.063 kg/L. Immunoblotting using non-SDS-PAGE suggested that the apoAI-containing lipoproteins are smaller than the apoE-containing lipoproteins, and there was variation in mean diameters between the fractions obtained by ultracentrifugation. Our data are similar to those of Guyton et al. (18) except that ours revealed a greater heterogeneity. The histogram we plotted for the size distribution of CSF lipoproteins indicated that two or more kinds of particles might exist in each subfraction obtained by ultracentrifugation. These particles may differ in terms of their apolipoprotein composition, as described above.
Weisgraber (21) has demonstrated that the preference of apoE3 for HDL and that of apoE4 for VLDL or intermediate-density lipoproteins results from a difference in the charge at residue 112 of the apoE molecule and whether they can form an apo(E-AII) complex. Although we could not investigate CSF with subjects homozygous for apoE2 or apoE4, it is quite possible that the distribution of apoE-containing lipoproteins in CSF is affected by the apoE phenotype. The apoE in CSF associated with apoE2 or E3 preferentially existed in the fractions with densities >1.063 kg/L. This tendency was more obvious in the subject with apoE2 than in the subject without apoE2, the difference being significant for the fraction with a density range of 1.125–1.21 kg/L. In contrast, in the CSF associated with apoE4, a large amount of apoE was recovered in the fraction with a density of <1.006 kg/L. This difference might be induced by cysteine-arginine interchanges at residues 112 and/or 158 of the apoE molecules and the formation of apo(E-AII) and apo(AII-E2-AII) complexes in CSF as well as in plasma. Although the presence of apoE in the bottom fraction could be an artifact of the ultracentrifugation method, as suggested previously (16)(30), it is obvious that the amount of apoE in the bottom fraction was greater in CSF associated with apoE4 than in CSF associated with the other phenotypes. One possible reason is that lipoproteins constructed with the apoE monomer would be unstable and that the formation of apoE homo- or heterodimers would be needed to stabilize these lipoproteins.
It is known that lecithin cholesterol acyl transferase (31) and cholesteryl ester transfer protein (32) are synthesized in the CNS. LaDu et al. (9) suggested that the lower density lipoproteins in the CSF contain much more nonesterified cholesterol than the higher density lipoproteins. This may support the operation of lecithin cholesterol acyl transferase and/or cholesteryl ester transfer protein. Although additional studies of the lipid composition of each fraction are required, the differences in lipid composition might also affect the distribution of apoE.
It has been demonstrated that subjects with apoE4 tend to show higher
plasma cholesterol concentrations than those with apoE3
(33). The different effects of apoE3 and apoE4 on
lipoprotein distribution is one of the reasons for this tendency. apoE4
displays a preference for triglyceride-rich lipoprotein particles in
plasma; therefore, the remnants would be expected to be cleared more
effectively in subjects with apoE4 than in those with apoE3. Indeed,
apoE homo- and heterodimers, which were formed by apoE3 and apoE2, but
not by apoE4, have been assumed to be inactive forms of apoE in terms
of binding to the LDL receptor. Consequently, the down-regulation of
hepatic LDL receptor would be more effectively induced by apoE4, and
this would raise the cholesterol concentration. In contrast, we
reported previously that CSF associated with apoE4 has a lower
cholesterol concentration than CSF associated with apoE3 or apoE2
(25). The TC/apoE ratio in CSF is 
1:60 of that in
serum. Thus, the cholesterol concentration in CSF is unlikely to
be affected by the same mechanism as that in plasma. apoE4-containing
lipoproteins, which do not include apo(E-AII) and apo(AII-E2-AII)
complexes, would be metabolized more effectively than apoE3- or
apoE2-containing lipoproteins.
apoE in the CNS is involved in the development of AD, and the frequency is significantly higher in subjects with apoE4 (13)(14)(15). Although various reasons for this correlation have been considered, such as isoform-specific binding of ß-amyloid (16)(17), the actual relationship between cause and effect remains to be determined. Mulder et al. (34) have suggested that an alteration of lipid homeostasis in the CNS might be related to AD. Our results suggest that the actual apoE isoform may influence lipid homeostasis in the CNS. If so, this would mean that the individual's phenotype plays an important role in the relationship between apoE and AD.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  I. Dolev and D. M. Michaelson A nontransgenic mouse model shows inducible amyloid-{beta} (A{beta}) peptide deposition and elucidates the role of apolipoprotein E in the amyloid cascade PNAS, September 21, 2004; 101(38): 13909 - 13914. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. L. TUMA and A. L. HUBBARD Transcytosis: Crossing Cellular Barriers Physiol Rev, July 1, 2003; 83(3): 871 - 932. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Suzuki, M. Tozuka, K. Yamauchi, M. Sugano, T. Nakabayashi, N. Okumura, H. Hidaka, T. Katsuyama, and K. Higuchi Predominant Apolipoprotein J Exists as Lipid-poor Mixtures in Cerebrospinal Fluid Ann. Clin. Lab. Sci., October 1, 2002; 32(4): 369 - 376. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. I. Posse de Chaves, D. E. Vance, R. B. Campenot, R. S. Kiss, and J. E. Vance Uptake of Lipoproteins for Axonal Growth of Sympathetic Neurons J. Biol. Chem., June 23, 2000; 275(26): 19883 - 19890. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Demeester, G. Castro, C. Desrumaux, C. De Geitere, J. C. Fruchart, P. Santens, E. Mulleners, S. Engelborghs, P. P. De Deyn, J. Vandekerckhove, et al. Characterization and functional studies of lipoproteins, lipid transfer proteins, and lecithin:cholesterol acyltransferase in CSF of normal individuals and patients with Alzheimer's disease J. Lipid Res., June 1, 2000; 41(6): 963 - 974. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and PL (•), each of which is
expressed in mg/L. The fraction numbers are shown on the
x-axis. A 30-µL aliquot of each of the gel-filtration
subfractions was treated with nonreducing Laemmli buffer and loaded
onto an 8–16% polyacrylamide gradient gel. After electrophoresis,
separated proteins were transferred electrophoretically onto
nitrocellulose membranes, followed by immunoblotting using anti-apoE
(a) or anti-apoAI (b) antibodies. The
bands were visualized by means of an enhanced chemiluminescence kit, as
described in Materials and Methods.
); 1.063–1.125 kg/L ,14.0 ± 3.2 nm (
); and
1.125–1.21 kg/L, 11.6 ± 2.8 nm (
).