Sialic acid moiety of apolipoprotein E and its impact on the formation of lipoprotein particles in human cerebrospinal fluid

Clinica Chimica Acta 402 (2009) 61–66

Contents lists available at ScienceDirect

Clinica Chimica Acta j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / c l i n c h i m

Sialic acid moiety of apolipoprotein E and its impact on the formation of lipoprotein particles in human cerebrospinal fluid Kenji Kawasaki a, Naoko Ogiwara a, Mitsutoshi Sugano a, Nobuo Okumura b, Kazuyoshi Yamauchi a,⁎ a b

Department of Laboratory Medicine, Shinshu University Hospital, Japan Department of Biomedical Laboratory Sciences, School of Health Sciences, Shinshu University, Japan

a r t i c l e

i n f o

Article history: Received 5 September 2008 Received in revised form 8 December 2008 Accepted 12 December 2008 Available online 24 December 2008 Keywords: Alzheimer's disease Apolipoprotein E Sialic acid ABEE Cerebrospinal fluid

a b s t r a c t Background: Apolipoprotein (apo) E in the cerebrospinal fluid (CSF) is abundant with sialic acid (SA), and sialylation of certain proteins is known to modulate biological function. The aim of the present study was to quantify the SA content in CSF apoE and carry out the more detailed characterization of the CSF apoEcontaining lipoproteins. Methods: The method for the determination of the SA in CSF apoE was based on the conversion of SA into paminobenzoic acid ethyl ester-derivatized N-acetylmannosamine, followed by HPLC analysis. Results: The levels of CSF SA and serum SA were 25.9 ± 1.5 and 2209 ± 196 μmol/l, respectively; however, when the SA values were corrected by the total protein concentrations, CSF SA values were approximately 3.5-fold of those in the serum. The SA levels in the CSF apoE-containing lipoprotein fractions were 5.3 ± 1.3% of total CSF SA, and were correlated with the CSF apoE concentrations. However, the ratios of SA to apoE were inversely proportional to the CSF lipid concentrations. The lipoprotein particle sizes were larger when the ratios of SA to CSF apoE were greater. Conclusion: The SA moiety of the CSF apoE molecules may affect the formation of the apoE-containing lipoprotein particles and the regulation of lipid delivery in CNS. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Apolipoprotein (apo) E, synthesized primarily in the liver, circulates as a mixture of its sialylated and asialylated forms and is involved in cholesterol transport and metabolism [1,2]. In the central nervous system (CNS)—the second most active tissue in terms of apoE production [3–5]. ApoE is produced predominantly by astrocytes and microglia as a high-density lipoprotein (HDL)-like lipoprotein particle [5–9]. In both the plasma and cerebrospinal fluid (CSF), the mature form of apoE consists of 299 amino acids with a molecular weight of 35 kDa; however, CSF apoE is more extensively sialylated than plasma apoE [10–12]. Sialylated apoE, with a slightly higher molecular weight of 36–39 kDa, is produced by posttranslational modification, and is more acidic than its isoproteins. It is well known that carbohydrate modifications at the posttranslational level, such as sialylation, render proteins biologically active; therefore, the abundance of sialic acid

Abbreviations: Apo, Apolipoprotein; CSF, cerebrospinal fluid; SA, sialic acid; Aβ, βamyloid; CNS, central nervous system; AD, Alzheimer's disease; ABEE, p-aminobenzonic acid ethyl ester; ManNAc, N-acethylmannosamine; DL, detection limit; QL, quantitation limit. ⁎ Corresponding author. Department of Laboratory Medicine, Shinshu University Hospital, 3-1-1 Asahi, Matsumoto 390-8621, Japan. Tel.: +81 263 37 2805; fax: +81 263 34 5316. E-mail address: [email protected] (K. Yamauchi). 0009-8981/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2008.12.018

(SA) in the CSF apoE is expected to have some significance with regard to the biological function of SA in the CNS [13–18]. The study of the CSF apoE is essential to our understanding of not only the metabolism of lipids in the brain but also the pathophysiology of the CNS. Although several groups, including ours, have previously characterized the CSF lipoproteins [9,19,20], the detailed nature of the CSF lipoproteins (in particular, the SA in apoE-containing lipoproteins) and their metabolic pathways are still not as clear as those of the plasma lipoproteins. CSF apoE accumulates as a complex with β-amyloid (Aβ) and forms senile plaques [21–23], one of the representative characteristics of Alzheimer's disease (AD); therefore, the CSF apoE, derived from CNS, is possible causative factor in the development of AD. Understanding the effect of apoE on the formation of senile plaques and on the consequent acceleration of the development of AD can help clarify the mechanism of AD pathogenesis. To date, although several possibilities have been proposed to explain this correlation, an isoform-specific interaction of apoE with Aβ has been highlighted in particular [24–28]. However, the actual causal relationship between the CSF apoE and AD development has not been identified. Recently, we studied the interaction of apoE with Aβ from a different point of view; that is, we focused on the characteristic abundance of the SA residue in the CSF apoE [27]. In that study, we found out that the binding avidity of apoE for Aβ decreased significantly on desialylation of apoE. That finding leads us to conclude

62

K. Kawasaki et al. / Clinica Chimica Acta 402 (2009) 61–66

that the SA of the CSF apoE may be closely involved in the formation of senile plaques. In the present study, on the basis of our previous study, we thought that the quantification of the SA bound to the carbohydrate moiety of the CSF apoE would allow more detailed characterization of sialylated apoE-containing lipoprotein and consequently facilitate clarification of the mechanism of AD development. Therefore, we devised an assay method for the SA in the CSF apoE by adapting the method involving the conversion of SA residues into p-aminobenzoic acid ethyl ester (ABEE)-derivatized N-acetylmannosamine (ManNAc) followed by the high-performance liquid chromatography (HPLC) analysis [29]. We then characterized sialo-apoE containing lipoprotein particles in the CSF by measuring the SA of the CSF apoE using the present method.

described [9]. Briefly, the neuraminidase (20 kU/l)-incubated serum was added to 5 μl of 75 mmol/l dithiothreitol (Wako Pure Chemicals) for 1 h at room temperature. The prepared sample was electrophoresed on a 4.8% polyacrylamide gel containing 8 mol/l urea and 20 g/l Ampholine (pH 4–6; GE healthcare) 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 a constant voltage (200 V). The separated proteins were transferred onto a nitrocellulose membrane. After blocking the nonspecific binding sites, the membranes were incubated with apoE polyclonal antibodies. The membranes were washed and then incubated with horseradish peroxidase-conjugated anti-rabbit IgG. After washing, the bands were visualized with an enhanced chemiluminescence detection kit (GE healthcare Ltd.).

2. Materials and methods

2.9. Other assays

2.1. Subjects

The concentrations of total protein (TP), total cholesterol (TC), phospholipids (PL), and other apos (excepted, apoE) in the CSF were determined by the Pyrogallol red method (Wako Pure Chemicals), cholesterol-oxidase method (Kyowa Medex Co. Ltd., Tokyo, Japan), choline-oxidase method (Wako Pure Chemicals), and turbidimetric immunoassay (Sekisui Medical Co. Ltd., Tokyo, Japan) using a Hitachi 7170 automated analyzer, respectively. The assay conditions for TC, PL, and apos were modified by increasing the ratio between sample and reagent volumes to a value 20-fold higher that used for serum. We expressed the TC plus PL concentration as the total lipid concentration.

Fourteen CSF and sera samples were obtained from hospitalized patients without inflammatory or neurological disease. There was no erythrocyte contamination in any of the CSF samples. This study was approved by the ethics committee of Shinshu University, Japan. All the patients provided their informed consent before participation. 2.2. Reagents Neu5Ac was from Merck (Darmstadt, Germany). An ABEE sugar-chain analysis kit was from Seikagaku Corp. (Tokyo, Japan). Neuraminidase (5000 U/l) and Neu5Ac aldolase (5000 U/l) were from Nakarai Tesque (Kyoto, Japan). Acetonitrile and HPLC-grade trifluoroacetic acid (TFA) were from Wako Pure Chemicals (Osaka, Japan). Anti-apoE polyclonal antibody (rabbit) and horseradish peroxidase-conjugated anti-rabbit IgG (goat) were from Dako (Glostrup, Denmark) and MBL Co. (Nagoya, Japan), respectively.

2.8. Measurement of the CSF apoE The CSF apoE concentration was determined by the ELISA method as described previously [30]. Each measurement was carried out in triplicate.

2.10. Electron microscopy The lipoprotein particles in the fraction with a density b 1.21 kg/l isolated from the CSF by ultracentrifugation were examined under a JEOL JEM1010 electron microscope. The samples were negatively stained with 40 g/l aqueous uranyl acetate.

2.3. Isolation of the total serum lipoprotein

2.11. Statistical methods

Serum lipoproteins were isolated by the ultracentrifugation method. Briefly, serum, adjusted to a density of 1.21 kg/l with the addition of solid KBr, was centrifuged at 541,000 ×g for 20 h using an Optima TLX Ultracentrifuge (Beckman Coulter, Fullerton, CA). The isolated fractions were dialyzed against phosphate-buffered saline (PBS).

All experiments were performed at least 3 times. Data are presented as means ± SE. The relationship between total lipid and log SA/apoE was determined from the correlation coefficient obtained by linear regression analysis. Statistical analysis was performed by Welch's t test using a StatFlex (Artec, Osaka, Japan). A p b 0.05 was accepted as statistically significant.

2.4. Isolation of the CSF apoE-containing lipoprotein A heparin-sepharose CL6B (GE Healthcare Ltd., Buckinghamshire, UK) column was prepared according to the manufacturers instructions. Five hundred microliters of the CSF or serum lipoprotein fraction was then loaded onto this column. After washing off the unbound proteins with phosphate buffer (pH 7.4) containing 0.05 mol/l NaCl, the bound fraction was eluted with phosphate buffer (pH 7.4) containing 0.15 mol/l NaCl at a flow rate of 0.5 ml/min. We measured the absorbance of each fraction at 280 nm, and performed a dot-blot analysis to confirm the apoE-containing fraction. The purity of the fraction was confirmed by quantifying the apos (apoAI, apoAII, apoB, apoCII, apoCIII, and apoE). 2.5. Principle of ABEE assay The SA of apoE was measured by the method described by Yasuno et al. [29] with a small modification. Briefly, the sample prepared using the method described above was treated with 10 μl of a mixture of neuraminidase (1000 U/l) and N-acetylneuraminic acid aldolase (5000 U/l) for 6 h at 45 °C to cleave N-acetylneuraminic acid (= SA) from sialo-apoE, and to convert it to ManNAc. We then added 40 μl of ABEE to the synthesized ManNAc, and incubated the mixture for 1 h at 80 °C. The reactant was mixed with chloroform–water in the ratio 1:2 (v/ v). A fixed volume (10 μl) of the aqueous layer of this mixture (ABEE-derivatized ManNAc) was injected into a reversed-phase HPLC system and eluted with a linear gradient acetonitrile–water mixture in a 0.2% TFA buffer at a flow rate of 1.0 ml/min. The eluted fractions were monitored at 305 nm (for excitation wavelength) and 360 nm (for emission wavelength). The HPLC system (Shimadzu, Kyoto, Japan) was constructed from a Honenpak C18 column (i.d., 75 mm × 4.6 mm; particle size, 5 mm; Seikagaku Corp., Tokyo, Japan), an LC10AD pump, column oven CTO-10A, degasifier DGU-14A, fluorescence detector RF-10A, and sample injector 7725i (Rheodyne, Rohnert Park, CA). 2.6. Calibrators Standard solutions of Neu5Ac were prepared by dilution of Neu5Ac (Merck, Darmstadt, Germany) to 0, 32.0, 64.0, 128, 256, and 320 μmol/l using distilled water as the diluent. 2.7. Isoelectric focusing The apoE phenotypes of the subjects used in this study were determined by a combination of isoelectric focusing and immunoblotting methods, as previously

3. Results 3.1. Calibration curve and dilution linearity To determine both the detection limit (DL) and quantitation limit (QL), we prepared 3 concentrations of ABEE-derivatized ManNAc solution. The DL and QL, determined according to the following equations – recommended by the ICH [31] – were 16.8×10− 3 and 56.1×10− 3 μmol/l, respectively. DL = 3:3 σ=SL Q L = 10 σ =SL (σ, standard deviation of response; SL, slope of calibration curve) The upper limit of the assay, as determined by a series of dilutions, was 320 μmol/l. 3.2. Precision The within-run reproducibility was determined from 10 replicate measurements of the 3 pooled total lipoproteins (CV, 3.1–4.0%). The between-run reproducibility was determined from 5 measurements of the 2 pooled total lipoproteins (CV, 3.6–5.7%). 3.3. Accuracy The accuracy of this assay was confirmed by analytical recovery studies. The recovery rates from the samples to which 8.08 and 16.16 μmol/l of Neu5Ac had been added were 98.0% and 93.5%, respectively.

K. Kawasaki et al. / Clinica Chimica Acta 402 (2009) 61–66

63

Table 2 Effect of age on the sialic acid level in serum and cerebrospinal fluid Age (years) Young group (n = 7) Aged group (n = 7)

5.4 ± 2.3 72.0 ± 3.2

Cerebrospinal fluid

Serum Sialic acid (μmol/l)

Sialic acid/TP (μmol/g TP)

Sialic acid (μmol/l)

Sialic acid/TP (μmol/g TP)

2209 ± 196 2290 ± 920

38.4 ± 3.1 39.8 ± 9.2

25.9 ± 1.5a 24.5 ± 3.3a

169.2 ± 18.1b 119.4 ± 16.6b,c

The values shown are the mean ± SE. a p b 0.001 compared with serum SA values. b p b 0.001 compared with serum SA, corrected by the concentration of TP. c p = 0.077 compared with young group.

3.5. SA in serum and CSF

Fig. 1. HPLC separation of ABEE-derivatized ManNAc. Standard Neu5Ac, patient's serum, patient's CSF, and the fraction of the CSF apoE-containing lipoprotein were treated as described in Materials and methods. These reactants were injected into a reversedphase HPLC system, and were eluted with a linear gradient acetonitrile–water mixture in a 0.2% TFA buffer at a flow rate of 1.0 ml/min. Peak 1, free ABEE; Peak 2, ABEEderivatized ManNAc.

The biochemical profiles of the CSF used in this study are summarized in Table 1. The levels of SA in serum and the CSF (mean ± SE) were 2209 ± 196 μmol/l and 25.9 ± 1.5 μmol/l, respectively, and the levels in the CSF were around one-150th to one-50th of those in serum. In contrast, CSF SA values, corrected by the TP concentration (142.6 ± 14.4 μmol/g TP, n = 14), were significantly higher than those in serum (39.1 ± 4.7 μmol/g TP, p b 0.001, n = 14). The SA values in the CSF slightly correlated with those in serum (r = 0.566, p b 0.001). However, when the SA values in both the CSF and serum were corrected by the TP concentrations, this correlation disappeared (r = 0.254, data not shown). As shown in Table 2, no significant difference in the amounts of both serum and the CSF SA was observed between the young (n = 7, 5.4 ± 2.3 y) and aged groups (n = 7, 72.0 ± 3.2 y); however, although statistical significance was not observed, the corrected CSF SA values of the young group (169.2 ± 18.1 μmol/g TP) tended to be higher than those of the aged group (119.4 ± 16.6 μmol/g TP, p = 0.077) but not the corrected serum SA values. In addition, no significant difference in the CSF SA levels was observed between males and females, or among apoE phenotypes; however, SA levels in the CSF apoE-containing lipoprotein fractions significantly correlated with the CSF apoE concentrations, determined by ELISA method ([30], r = 0.924, p b 0.001; Fig. 2) but not the total lipid concentrations (data not shown). 3.6. The amounts of SA bound to the apoE-containing lipoprotein

3.4. HPLC separation of ABEE-derivatized ManNAc The HPLC chromatograms are shown in Fig. 1. The retention time of the synthesized Neu5Ac, derived from CSF apoE or serum apoE, was approximately 18.32 ± 0.05 min, and was completely consistent with that of the standard Neu5Ac. We also detected free ABEE at the retention time of 24.65 ± 0.11 min.

To obtain the CSF apoE-containing lipoprotein, we applied the CSF onto the heparin-sepharose affinity column as described in Materials and methods. The optical peak of the eluent, monitored at 280 nm, was detected in fraction 32 (Fig. 3-a). However, the dot-blot analysis indicated that the strongest immunoreactivity for apoE was expressed slightly after the optical peak and was detected in fractions 36–40 (Fig. 3-b). The main peaks of the SA-containing fractions consisted of

Table 1 Biochemical profiling of cerebrospinal fluids used in this study Subject no.

Sex

Age (years)

Total protein (mg/l)

apoE (mg/l)

apoE phenotype

Total lipid (mg/l)

Sialic acid in CSF (μmol/l)

Sialic acid in serum (μmol/l)

1 2 3 4 5 6 7 8 9 10 11 12 13 14

F M M F M F M M F M F M F M

8 4 9 4 6 3 4 76 69 73 76 68 70 72

156 123 266 298 100 110 110 380 190 160 130 330 320 211

2.03 1.41 2.34 2.76 2.91 2.41 2.63 2.08 ND ND ND 5.25 4.40 3.94

E3/E3 E3/E3 E3/E3 E3/E2 E3/E3 E4/E3 E3/E3 E3/E3 E3/E3 E3/E3 E4/E3 E3/E3 E3/E3 E4/E3

3.66 6.49 9.29 9.40 5.75 3.76 1.20 8.09 5.14 2.89 8.57 9.45 11.15 10.52

29.0 20.2 29.4 36.3 25.6 18.0 20.1 29.7 21.1 24.4 22.5 25.0 17.5 35.1

3285 2137 2252 1908 1642 1716 1956 3134 2318 1689 1656 1514 1776 3943

ND, we could not determine the concentrations, because of samples shortage.

Fig. 2. Comparison between apoE concentrations and SA levels in the CSF apoEcontaining lipoprotein fractions.

64

K. Kawasaki et al. / Clinica Chimica Acta 402 (2009) 61–66

Fig. 3. Heparin-sepharose CL6B chromatography (a) and dot-blot analysis (b) of the subfractions. Five hundred microliters of the CSF was then loaded onto this column. After the unbound proteins were washed off with phosphate buffer (pH 7.4) containing 0.05 mol/l NaCl, the bound fraction was eluted with phosphate buffer (pH 7.4) containing 0.15 mol/l NaCl at a flow rate of 0.5 ml/min. We measured the absorbance of each fraction at 280 nm, and carried out dot-blot analysis for the subfractions.

the peaks of apoE and lipids (Fig. 4). We could detect apoE, but not any apos, in these fractions by ELISA method ([30], data not shown). We also detected trace apoE in the unbound fractions by dot-blot analysis, but this concentration was not detectable by ELISA method. The content of SA in the CSF apoE-containing lipoproteins was 5.3 ± 1.3% of total SA in the CSF. We also isolated apoE-containing lipoprotein from the same patients' serum, taken on the same day, to compare the amounts of SA bound to the apoE-containing lipoprotein in the serum and CSF. The SA in serum apoE-containing lipoproteins had decreased to 0.05%. Therefore, the amount of SA bound to apoE-containing lipoprotein in the CSF was significantly higher than that in the serum (p b 0.002). 3.7. Effect of the amount of SA on the formation of the CSF apoE-containing lipoprotein We compared the relationship between the concentration of total lipid and the ratio of SA to apoE, to investigate the effect of the SA content on the formation of the CSF apoE-containing lipoprotein. As shown in Fig. 5a, the total lipid levels in apoE-containing lipoprotein

Fig. 4. The analysis of the bound fractions obtained by heparin-sepharose CL6B chromatography. The bound fractions, obtained by heparin-sepharose CL6B chromatography, were analyzed for lipids (TC plus PL, ●), SA (○), and apoE (▲) in the CSF.

Fig. 5. Analysis of the CSF apoE-containing lipoproteins particles. (a), Comparison between total lipid concentrations and the ratios of SA to apoE in the CSF apoEcontaining lipoprotein fractions. The x-axis and y-axis indicate the ratio of SA to apoE and the logarithm of total lipid (TC plus PL) concentration, respectively. (b), Negativestaining electron microscopy for apoE-containing lipoprotein particles. The ultracentrifuged subfractions with densities less than 1.21 kg/l were isolated from CSF-1, CSF-2. A 20-μl sample of each was placed on a copper grid (300 mesh) and negative stained with 40 g/l uranyl acetate. The lipoprotein particles were measured in enlarged photomicrographic prints using a micrometer. The bar indicates 100 nm.

were inversely proportional to the ratios of SA to apoE (r = 0.847, p b 0.001). Further, we examined the lipoprotein particles with a density less than 1.21 kg/l isolated from CSF-1 and CSF-2 (see Fig. 5a) under an electron microscope, and compared the particle sizes of these samples. The diameters of lipoprotein particles (mean ± SE) isolated from CSF-1 and CSF-2 were 17.5 ± 0.2 nm and 12.0 ± 0.1 nm, respectively (Fig. 5b). The differences in particle size between these fractions were significant (p b 0.001). The diameter of serum HDL lipoprotein particles, measured as a control for the electron microscopic analysis, was 11.8 ± 0.2 nm (data not shown). 4. Discussion The apoE-containing lipoprotein is one of the main CSF lipoproteins and its distinguishing characteristic is its high sialylation [10– 12]. The SA moieties of apoE are contained in its terminal carbohydrate moiety and attached only to the Thr194 residue at the core protein via an O-linked glycosylation [32–34]. It has been clarified that sialylation is not necessary for apoE secretion [32,34]; however, the exact role of the SA residues of the apoE molecule in its pathophysiological function has not been completely elucidated. The aim of the current study was to clarify the biological significance of SA bound to apoE-containing lipoprotein in the CSF by quantifying it using the method based on the ABEE conversion of SA. This method has sufficient accuracy, precision, and sensitivity to detect very small quantities of SA in the CSF or the fraction of apoEcontaining lipoprotein. The CSF SA levels (mean ± SE, 25.9 ± 1.5 μmol/l) determined by the present method were approximate to the previous results [27.2 μmol/l, determined by the method using resorcinol [35]; 51.7 ± 26.9 μmol/l, determined by the method using 1,2-diamino-4,5methylenedioxybenzene dihydrochloride [36]].

K. Kawasaki et al. / Clinica Chimica Acta 402 (2009) 61–66

The number of CSF samples in the present study was few, since we did not harvest the CSF, satisfied the criteria described in Materials and methods, and deliberately categorized the subjects into 2 groups (young and aged groups) to investigate the age effect. Thus, although the present results were obtained from small mass analyses, we could report several novel and worthwhile findings concerned with the character of the CSF apoE-lipoprotein particles. The total SA level in serum was higher than in the CSF. In contrast, the relationship between the levels of SA in serum and the CSF was reversed on correction with TP concentrations; that is, the corrected CSF SA values were significantly higher than the corrected serum SA values. Although SA is well known as an inflammation marker [37,38], we used subjects without any inflammation-related diseases in the present study. Hence, this finding indicates that CSF SA may discharge some distinctive physiological roles in the CNS, such as recognition and binding to cellular receptors. In addition, the finding that the correlation between serum SA and the CSF SA disappeared on correction with TP concentration indicated that while the CSF SA may partially originate from the free SA in blood through the bloodbrain barrier, most of it may be independently derived from the CNS and already exist in the protein-bound form. The corrected CSF SA values of the aged group appeared to be lower than those of the young group, although the difference was not statistically significant. It has been demonstrated that the SA content in platelets decreases with age and in malignant conditions [39]. On the other hand, desialylation protects low-density lipoprotein (LDL) particles from oxidative damage [40]. In our study, it was not clear whether the change in the CSF SA values would function as accelerating or protective factors in some injuries. However, our findings suggest that that this knowledge may be critical in understanding the reason behind the sialylation of proteins derived from CNS and the relationship between aging and the development of neurodegenerative diseases such as AD. The correlation between SA levels in the fractions of apoEcontaining lipoproteins and the concentrations of the CSF apoE, albeit not after correction with TP, definitely reflects that the CSF apoEcontaining lipoproteins contain SA moieties in their molecules. Indeed, the analysis using the heparin-sepharose affinity column showed that the SA-containing fractions completely overlapped with the fraction of the apoE-containing lipoprotein (the binding fractions, Fig. 4). Although the SA levels in the CSF apoE-containing lipoproteins were only b10% of the total SA levels, the relative amounts of SA bound to the CSF apoE molecules were approximately 3.5-fold of those bound to serum apoE molecule. This result was consistent with our previous results obtained by isoelectric focusing [9]. Using the surface plasmon resonance assay with a BIAcore system, we recently demonstrated that the binding avidity of asialo-apoE for dimyristoyl-phosphatidylcholine liposomes was considerably stronger than that of sialo-apoE [27]. In present study, the total lipid concentrations were inversely proportional to the ratios of SA to the CSF apoE (Fig. 5a), whereas the lipoprotein particle sizes were enlarged on an increase in the ratios of SA to the CSF apoE (Fig. 5b). These findings also indicate that the SA moiety of the CSF apoE molecules may be affecting the binding with lipids and may be involved in the formation of apoE-containing lipoprotein particles in CNS; that is, it is expected that the high sialylation of apoE may induce the formation of lipid-deficient, immature, large lipoproteins, whereas the decrease in sialylation of apoE may induce the formation of lipidrich, small, mature lipoproteins. Previous studies demonstrated [9,19,20] that the apoE-containing lipoproteins in the CSF are microheterogeneous, and that the ApoAIcontaining lipoproteins are smaller than the apoE-containing lipoproteins. In present study, even though we focused on only the apoEcontaining lipoproteins, we observed great heterogeneity of the lipoprotein particles in the electron microscopy studies (Fig. 5b). It appears that the heterogeneity may depend not only on the

65

composition of lipoprotein particles in the apos but also on the sialylation of the CSF apoE, as described above. The variety in the ratios of SA to apoE in the CSF may be induced by the greater heterogeneity. It is also well known that the CSF apoE-containing lipoprotein regulates cholesterol recycling in the CNS [1,2,41]. Taken together with our findings, the sialylation (or desialylation) of the CSF apoE may also affect the regulation of lipid delivery in neuronal tissue. Understanding the interaction of the CSF apoE-containing lipoprotein with Aβ and its effects on the formation of senile plaques could clarify the pathogenesis of AD. A previous study [42] demonstrated that lipidated apoE expresses higher Aβ-binding affinity than delipidated apoE and the lipidation of apoE promotes its binding with Aβ. LaDu et al. [43] suggested that protection from Aβ neurotoxicity depends on the formation of apoE-Aβ complex and the clearance of the complex. We recently demonstrated that the binding avidities of apoE for Aβ are significantly decreased by the desialylation of apoE [27]. Overall, these findings lead us to the notion that the SA of the CSF apoE may be closely involved not only in the formation of apoE-containing lipoprotein particles but also in the deposition of senile plaques in brain. In conclusion, we demonstrated the possible role of the SA moiety of the CSF apoE in the formation of apoE-containing lipoprotein particles. Our findings represent important knowledge of the formation of lipoprotein particle and lipid homeostasis in CNS, and could be helpful in understanding the relationship between lipid metabolism and the development of AD. Acknowledgments The authors thank Dr. Shun'ichiro Taniguchi, Dr. Takayuki Honda and Dr. Tsutomu Katsuyama (Shinshu University School of medicine) for valuable suggestions. References [1] Mahley RW. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science 1988;240:622–30. [2] Beisiegel U, Weber W, Ihrke G, Herz J, Stanley KK. The LDL-receptor-related protein, LRP, is an apolipoprotein E-binding protein. Nature 1989;341:162–4. [3] Elshourbagy NA, Liao WS, Mahley RW, Taylor JM. Apolipoprotein E mRNA is abundant in the brain and adrenals, as well as in the liver, and is present in other peripheral tissues of rats and marmosets. Proc Natl Acad Sci U S A 1985;82:203–7. [4] Boyles JK, Pitas RE, Wilson E, Mahley RW, Taylor JM. Apolipoprotein E associated with astrocytic glia of the central nervous system and with nonmyelinating glia of the peripheral nervous system. J Clin Invest 1985;76:1501–13. [5] Pitas RE, Boyles JK, Lee SH, Hui D, Weisgraber KH. Lipoproteins and their receptors in the central nervous system. Characterization of the lipoproteins in cerebrospinal fluid and identification of apolipoprotein B, E(LDL) receptors in the brain. J Biol Chem 1987;262:14352–60. [6] Pitas RE, Boyles JK, Lee SH, Foss D, Mahley RW. Astrocytes synthesize apolipoprotein E and metabolize apolipoprotein E-containing lipoproteins. Biochim Biophys Acta 1987;917:148–61. [7] Roheim PS, Carey M, Forte T, Vega GL. Apolipoproteins in human cerebrospinal fluid. Proc Natl Acad Sci U S A 1979;76:4646–9. [8] Miida T, Yamazaki F, Sakurai M, et al. The apolipoprotein E content of HDL in cerebrospinal fluid is higher in children than in adults. Clin Chem 1999;45:1294–6. [9] Yamauchi K, Tozuka M, Hidaka H, Hidaka E, Kondo Y, Katsuyama T. Characterization of apolipoprotein E-containing lipoproteins in cerebrospinal fluid: effect of phenotype on the distribution of apolipoprotein E. Clin Chem 1999;45:1431–8. [10] Zannis VI, Kurnit DM, Breslow JL. Hepatic apo-A-I and apo-E and intestinal apo-A-I are synthesized in precursor isoprotein forms by organ cultures of human fetal tissues. J Biol Chem 1982;257:536–44. [11] Zannis VI, Breslow JL, SanGiacomo TR, Aden DP, Knowles BB. Characterization of the major apolipoproteins secreted by two human hepatoma cell lines. Biochemistry 1981;20:7089–96. [12] Zannis VI, McPherson J, Goldberger G, Karathanasis SK, Breslow JL. Synthesis, intracellular processing, and signal peptide of human apolipoprotein E. J Biol Chem 1984;259:5495–9. [13] Schauer R. Sialic acids as antigenic determinants of complex carbohydrates. Adv Exp Med Biol 1988;228:47–72. [14] Varki A. Sialic acids as ligands in recognition phenomena. FASEB J 1997;11:248–55. [15] Munday J, Floyd H, Crocker PR. Sialic acid binding receptors (siglecs) expressed by macrophages. J Leukoc Biol 1999;66:705–11. [16] Keppler OT, Horstkorte R, Pawlita M, Schmidt C, Reutter W. Biochemical engineering of the N-acyl side chain of sialic acid: biological implications. Glycobiology 2001;11:11R–18R. [17] Kannagi R. Regulatory roles of carbohydrate ligands for selectins in the homing of lymphocytes. Curr Opin Struct Biol 2002;12:599–608.

66

K. Kawasaki et al. / Clinica Chimica Acta 402 (2009) 61–66

[18] Crocker PR. Siglecs: sialic-acid-binding immunoglobulin-like lectins in cell–cell interactions and signalling. Curr Opin Struct Biol 2002;12:609–15. [19] Guyton JR, Miller SE, Martin ME, Khan WA, Roses AD, Strittmatter WJ. Novel large apolipoprotein E-containing lipoproteins of density 1.006–1.060 g/ml in human cerebrospinal fluid. J Neurochem 1998;70:1235–40. [20] LaDu MJ, Gilligan SM, Lukens JR, Cabana VG, Reardon CA, Van Eldik LJ, Holtzman DM. Nascent astrocyte particles differ from lipoproteins in CSF. J Neurochem 1998;70:2070–81. [21] Namba Y, Tomonaga M, Kawasaki H, Otomo E, Ikeda K. Apolipoprotein E immunoreactivity in cerebral amyloid deposits and neurofibrillary tangles in Alzheimer's disease and kuru plaque amyloid in Creutzfeldt-Jakob disease. Brain Res 1991;541:163–6. [22] Gunzburg MJ, Perugini MA, Howlett GJ. Structural basis for the recognition and cross-linking of amyloid fibrils by human apolipoprotein E. J Biol Chem 2007;282:35831–41. [23] Xu F, Vitek MP, Colton CA, et al. Human apolipoprotein E redistributes fibrillar amyloid deposition in Tg-SwDI mice. J Neurosci 2008;28:5312–20. [24] Strittmatter WJ, Saunders AM, Schmechel D, et al. Apolipoprotein E: high-avidity binding to β-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease. Proc Natl Acad Sci U S A 1993;90:1977–81. [25] LaDu MJ, Pederson TM, Frail DE, Reardon CA, Getz GS, Falduto MT. Purification of apolipoprotein E attenuates isoform-specific binding to β-amyloid. J Biol Chem 1995;270:9039–42. [26] Malek G, Johnson LV, Mace BE, et al. Apolipoprotein E allele-dependent pathogenesis: a model for age-related retinal degeneration. Proc Natl Acad Sci U S A 2005;102:11900–5. [27] Sugano M, Yamauchi K, Kawasaki K, et al. Sialic acid moiety of apolipoprotein E3 at Thr194 affects its interaction with β-amyloid(1–42) peptides. Clin Chim Acta 2008;388:123–9. [28] Yamauchi Y, Deguchi N, Takagi C, et al. Role of the N- and C-terminal domains in binding of apolipoprotein E isoforms to heparan sulfate and dermatan sulfate: a surface plasmon resonance study. Biochemistry 2008;47:6702–10. [29] Yasuno S, Kokubo K, Kamei M. New method for determining the sugar composition of glycoproteins, glycolipids, and oligosaccharides by high-performance liquid chromatography. Biosci Biotechnol Biochem 1999;63:1353–9. [30] Yamauchi K, Tozuka M, Nakabayashi T, et al. Apolipoprotein E in cerebrospinal fluid: relation to phenotype and plasma apolipoprotein E concentrations. Clin Chem 1999;45:497–504.

[31] International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use, ICH Harmonised Tripartite Guideline, Validation of Analytical Procedures: Text and Methodology, Q2 (R1); 2005. [32] Reardon CA, Driscoll DM, Davis RA, Borchardt RA, Getz GS. The charge polymorphism of rat apoprotein E. 1986;261:4638–4645. [33] Zanni EE, Kouvatsi A, Hadzopoulou-Cladaras M, Krieger M, Zannis VI. Expression of ApoE gene in Chinese hamster cells with a reversible defect in O-glycosylation. Glycosylation is not required for apoE secretion. J Biol Chem 1989;264:9137–40. [34] Wernette-Hammond ME, Lauer SJ, Corsini A, Walker D, Taylor JM, Rall Jr SC. Glycosylation of human apolipoprotein E. The carbohydrate attachment site is threonine 194. J Biol Chem 1989;264:9094–101. [35] Kakari S, Avgoustatos G, Ferderigos AS, et al. Total and lipid-bound sialic acid in the cerebrospinal fluid of patients with brain tumors. Anticancer Res 1984;4:313–6. [36] Hayakawa K, De Felice C, Watanabe T, et al. Determination of free Nacetylneuraminic acid in human body fluids by high-performance liquid chromatography with fluorimetric detection. J Chromatogr 1993;620:25–31. [37] De Graaf TW, Van der Stelt ME, Anbergen MG, van Dijk W. Inflammation-induced expression of sialyl Lewis X-containing glycan structures on alpha 1-acid glycoprotein (orosomucoid) in human sera. J Exp Med 1993;177:657–66. [38] Reganon E, Vila V, Martínez-Sales V, et al. Sialic acid is an inflammation marker associated with a history of deep vein thrombosis. Thromb Res 2007;119:73–8. [39] Goswami K, Koner BC. Level of sialic acid residues in platelet proteins in diabetes, aging, and Hodgkin's lymphoma: a potential role of free radicals in desialylation. Biochem Biophys Res Commun 2002;297:502–5. [40] Dousset N, Dousset JC, Taus M, et al. Effect of desialylation on low density lipoproteins: comparative study before and after oxidative stress. Biochem Mol Biol Int 1994;32:555–63. [41] Poirier J. Apolipoprotein E, cholesterol transport and synthesis in sporadic Alzheimer's disease. Neurobiol Aging 2005;26:355–61. [42] Bentley NM, Ladu MJ, Rajan C, Getz GS, Reardon CA. Apolipoprotein E structural requirements for the formation of SDS-stable complexes with β-amyloid-(1–40): the role of salt bridges. Biochem J 2002;366:273–9. [43] Ladu MJ, Reardon C, Van Eldik L, Fagan AM, Bu G, Holtzman D, Getz GS. Lipoproteins in the central nervous system. Ann N Y Acad Sci 2000;903:167–75.