Geometrical Separation Method for Lipoproteins Using Bioformulated

TECHNICAL NOTE pubs.acs.org/ac

Geometrical Separation Method for Lipoproteins Using Bioformulated-Fiber Matrix Electrophoresis: Size of High-Density Lipoprotein Does Not Reflect Its Density )

Mari Tabuchi,*,†,‡ Makoto Seo,§ Takayuki Inoue,† Takeshi Ikeda,† Akinori Kogure, Ikuo Inoue,^ Shigehiro Katayama,^ Toshiyuki Matsunaga,X Akira Hara,X and Tsugikazu Komoda§,O †

Department of Chemistry, Faculty of Science, Rikkyo University, 3-34-1, Nishi-Ikebukuro, Toshima-ku, Tokyo, 171-8501, Japan Research Information Center for Extremophile, Rikkyo University, 3-34-1, Nishi-Ikebukuro, Toshima-ku, Tokyo, 171-8501, Japan § Department of Biochemistry, Faculty of Medicine, Saitama Medical University, 38 Morohongo, Moroyama, Iruma-gun, Saitama 350-0495, Japan Surface Analysis Division, Shimadzu Analytical & Measuring Center, Inc. 380-1, Horiyamashita, Hadano-city, Kanagawa, 259-1304, Japan ^ Department of Endocrinology and Diabetes, Faculty of Medicine, Saitama Medical University, 38 Morohongo, Moroyama, Iruma-gun, Saitama 350-0495, Japan X Laboratory of Biochemistry, Gifu Pharmaceutical University, 5-6-1 Mitahora-higashi, Gifu 502-8585, Japan

)

‡

bS Supporting Information ABSTRACT: The increasing number of patients with metabolic syndrome is a critical global problem. In this study, we describe a novel geometrical electrophoretic separation method using a bioformulated-fiber matrix to analyze high-density lipoprotein (HDL) particles. HDL particles are generally considered to be a beneficial component of the cholesterol fraction. Conventional electrophoresis is widely used but is not necessarily suitable for analyzing HDL particles. Furthermore, a higher HDL density is generally believed to correlate with a smaller particle size. Here, we use a novel geometrical separation technique incorporating recently developed nanotechnology (Nata de Coco) to contradict this belief. A dyslipidemia patient given a 1-month treatment of fenofibrate showed an inverse relationship between HDL density and size. Direct microscopic observation and morphological observation of fractionated HDL particles confirmed a lack of relationship between particle density and size. This new technique may improve diagnostic accuracy and medical treatment for lipid related diseases.

A

nalysis of serum lipoproteins is necessary for diagnosing metabolic syndrome. Current methods for measuring or separation of lipoproteins include gradient ultracentrifugation,1,2 gel gradient electrophoresis (GGE),3 gel permeation or immunoaffinity high-performance liquid chromatography (HPLC),4,5 and capillary or microchip electrophoresis (μ-CE).6-9 With the use of these methods, four major lipoprotein fractions are typically isolated and separated, i.e., chylomicrometers (CM), high-density lipoprotein (HDL), low-density lipoprotein (LDL), and very low-density lipoprotein (VLDL), with each fraction having several well-defined subclasses. HDL is divided into two or more particles according to density, usually HDL2 (d = 1.063-1.125) and HDL3 (d = 1.125-1.210). Although HDL2 shows an inverse correlation with coronary heart disease and HDL cholesterol levels,10 higher levels of HDL are not necessarily beneficial as shown by recent findings in patients with cholesterol ester transfer protein (CETP) deficiency.11 These findings suggest that HDL particles may be more closely related to dyslipidemia than previously thought, and further detailed studies are required to define critical r 2010 American Chemical Society

diagnostic criteria. Although HDL density is generally believed to correlate with a smaller particle size, we hypothesized that a higher density does not necessarily correlate with a smaller particle size. Here, we develop a novel electrophoresis separation technique and use morphological observation to test this hypothesis. Traditional electrophoresis is a powerful tool for studying biopolymers and lipoproteins. A number of simple procedures based on electrophoresis are presently commercially available for such analyses. However, a simple and rapid separation method has yet to be developed for HDL particles, possibly due to the smaller nanoscale size of HDL and structure/charge variability. For example, morphological structural changes during lipoprotein maturation are well-known. Conventional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), the typical size-based electrophoretic Received: September 9, 2010 Accepted: December 1, 2010 Published: December 20, 2010 1131

dx.doi.org/10.1021/ac102402c | Anal. Chem. 2011, 83, 1131–1136

Analytical Chemistry separation method for proteins, is usually achieved after denaturing with excess SDS. However, excess SDS leads to the decomposition of lipoprotein particles. Therefore, SDS-PAGE cannot be adopted for size separation of lipoproteins. Instead, blue native PAGE is often used, which is based on molecular weight and shape. However, the resolution between heavy and light HDL particles is somewhat poor, due to HDL size relative to the gel matrix. There is a critical need to develop a new HDL separation technique that could attain a higher resolution that would also not be affected by variation in lipoprotein charge. Minute nanoscale electrophoresis separation is possible only by using micro- or nanotechnology,7 new technology that has attracted considerable attention. We recently developed a higherresolution structural size-separation method using bionanotechnology, bioformulated-fiber matrix (BFM) by μ-CE.12-14 This allows minute differences in single strand polymorphisms (SNPs) to be detected. BFM is composed of ∼10 μm fragments of bacterial cellulose (BC) containing several nanometer to 1 μm meshes with BC rigid fibrils together with a diluted polymer solution. Since BC fibrils are more rigid and fine than polyacrylamide in a conventional high-viscosity polyacrylamide gel, BFM can attain a higher resolution separation of biopolymers. This system could potentially attain fine separation of lipoprotein particles based on size and shape, without being affected by the electric charge, when buffer conditions are optimized. There are numerous reports on the components and size of lipoproteins, lipids, and apolipoproteins (Apo) analyzed using either mass spectrometry (MS),6 microscopy,15 or metabolic analyses.16,17 However, there are only a few electrophoresis separation studies based on HDL morphology. In this study, we were able to achieve geometric separation of HDL particles utilizing BFM-μ-CE and found that density was not related to size, with high density HDL particles not necessarily having a smaller size. These findings indicate significant morphological variability of HDL particles which may also cause differences in electrophoretic mobility.

’ EXPERIMENTAL PROCEDURE Electrophoresis. A gel matrix including BFM was used for the electrophoretic separation of lipoproteins. BFM was prepared by static culture of Acetobacter in corn seep liquor-fructose (CSL-Fru) medium.14 The features of the matrix used in the electrophoresis are presented in Figure 1A. A cobweb-like matrix formed by BC was applied to the BFM.12,14 The electrophoretic buffer consisted of 0.07% [v/v] hydroxypropylmethylcellulose (HPMC; Hitachi Chemical, Hitachi Japan), 30 mM 3-morpholinopropaneesulfonic acid (MOPS; Sigma-Aldrich, Corp. Japan), 0.3 mM sodium dodecyl sulfate (SDS, Sigma-Aldrich, Corp. Japan), and 0.03% of the above BFM solution. A 1 μL aliquot of 1 μg/mL lipoprotein subfraction or serum with 1 μL of 1 mM SDS, 1 μL of stock solution of 30% (v/v) BODIPY FL C5-ceramide solution, and 7 μL of water was applied as a sample. Milli-Q water (ICN Biomedicals, Aurora, OH) was used for buffer and sample preparation. Sample dilution buffer included 0.1% SDS, and electrophoresis running buffer included 0.3% SDS, so as not to decompose the lipoprotein and to decrease the contribution of the electric charge of lipoproteins on the system. The lipoprotein charge was analyzed using the ζ-potential (Zeecom, Mcrotec Co., Ltd. Japan) or electric conductivity (Twin compact meter B173, Horiba, Co., Ltd., Japan). Electrophoresis was performed using a μ-CE system (SV1100, Hitachi Electronics Engineering,

TECHNICAL NOTE

Figure 1. Lipoprotein analyses by the BFM-μ-CE method. (A) View of the BFM. (B) The relationship between size or electric charge and migration times in the BFM-μCE-system. H, HDL including heavy HDL and light HDL; L, LDL; and V, VLDL. The b shows the lipoprotein size, and the 9 shows lipoprotein charge. The sizes were analyzed by SPM, and the electric charges were analyzed by electric conductivity. The migration times were analyzed by BFM-μ-CE.

Co., Ltd., Japan) equipped with an LED emitting at 470 nm.14 Polymethylmetarylrate (PMMA) chips containing microchannels 100 μm in width, 30 μm in depth, and 30 mm in length (i-chip, Hitachi Chemical, Hitachi, Japan) were utilized for the μ-CE. The migration times were normalized using marker protein and fluorescence peaks. The Lipophor system (Jookoh Co., Ltd., Japan) was used for conventional electrophoresis. Lipoproteins and Other Materials. Blood samples were obtained from 4 patients with dyslipidemia, aged 45-67 years (males/females = 2/2). Blood samples were collected from the patients on two occasions under different conditions. The first samples were taken after a fast of 12 h or longer, and the second samples collected after 1 month of treatment with fenofibrate. In another study, normal blood samples were collected for a standard sample from healthy normolipidemic volunteers after a 12 h or longer fast. All blood samples were collected into tubes containing 0.1% EDTA, pH 7.4. Serum was separated by centrifugation at 3 000 rpm for 20 min at 4 C and was then subjected to lipoprotein fractionation within 24 h of isolation. During this period, the samples were stored at 4 C. Lipoprotein particles including VLDL (d < 1.006), LDL (d = 1.006-1.063), lower density HDL (d = 1.063-1.125), and higher density HDL (d = 1.125-1.210) were isolated by sequential ultracentrifugal flotation. The method used was a slight modification of a previously described procedure.18 Briefly, a serum sample of 4 mL in 2 mL of NaBr solution (d = 1.006) was ultracentrifuged at 20 000 rpm for 30 min at 16 C in OptimaMAX (Beckman Coulter) using a MLA80 rotor (Beckman Coulter). After ultracentrifugation, the thin whitish supernatant (chylomicrometer fraction) was removed, and a 4 mL of serum sample in 2 mL of NaBr solution (d = 1.006) was ultracentrifuged at 45 000 rpm for 18 h at 16 C. After ultracentrifugation, the thin 1132

dx.doi.org/10.1021/ac102402c |Anal. Chem. 2011, 83, 1131–1136

Analytical Chemistry whitish supernatant layer containing VLDL (d < 1.006) was collected, and a 4 mL of serum sample (d = 1.006) in 2 mL of NaBr solution (d = 1.182) was ultracentrifuged at 50 000 rpm for 22 h at 16 C. After ultracentrifugation, the thin yellowish supernatant layer containing LDL (d = 1.006-1.063) was collected, and 2 mL of serum sample (d = 1.063) in 2 mL of NaBr solution (d = 1.186) was ultracentrifuged at 55 000 rpm for 22 h at 16 C. After ultracentrifugation, the thin yellowish supernatant layer containing light HDL (d = 1.063-1.125) was collected, and a 2 mL of serum sample (d = 1.125) in 2 mL of NaBr solution (d = 1.295) was ultracentrifuged at 55 000 rpm for 22 h at 16 C. After ultracentrifugation, the thin yellowish supernatant layer containing heavy HDL (d = 1.125-1.210) was collected. Lipoproteins were recovered from the tubes and dialyzed against phosphate-buffered saline (PBS) containing 3 μM EDTA. The lipoproteins were sterilized by filtration and stored at 4 C in the dark. To maintain the morphology of the lipoproteins, we employed mild conditions in the ultracentrifugation procedures and used an Optima-MAX with a MLA80 rotor (Beckman Coulter) for this purpose. In this paper, the lower density subfraction of HDL (d = 1.063-1.125) is termed “light HDL (HL)” and the higher density of HDL subfraction (d = 1.125-1.210) “heavy HDL (HH)”. SPM and Small Angle X-ray Scattering (SAXS) Analyses. The morphology and size of the lipoproteins were analyzed using an SPM-9600 (Shimadzu, Japan) under liquid conditions with neither drying nor sputter-coating. SAXS (Nano-Viewer, Rigaku, Japan) was also used for lipoprotein size determination. Safety Consideration. Gloves were worn when human blood samples were handled.

TECHNICAL NOTE

Figure 2. Electropherograms of a mixture of lipoproteins: (A) mixture including HDL (H) (heavy HDL and light HDL), LDL (L), and VLDL (V) isolated by ultracentrifugation analyzed by BFM-μ-CE. H/L/V = 0.5:0.6:0.1 (w/v). Normal standard lipoproteins were utilized. (B) (a) The magnified electropherogram of HDL particles. (b) Identification of a mixture of heavy HDL (HH) and light HDL (HL).

’ RESULTS AND DISCUSSION Development of a BFM-μ-CE Technique for Lipoprotein Analysis. Since the lipoproteins are heterogeneous in both size

and charge, we first investigated the separation mechanism of lipoproteins in the BFM-μ-CE system. Figure 1B shows the relation between migration time and size/charge of the lipoprotein particles. The size of lipoproteins correlated with the migration time, whereas electric charge did not affect migration times under the same buffer conditions. More detailed observation showed that though each HDL particle has its own electric charge in a water environment, it was constant under optimized buffer conditions (Supplementary Figure 1 in the Supporting Information). The detailed buffer conditions are described in the Experimental Procedure. Finally, the optimized electrophoresis conditions achieved size-separation of lipoproteins without being influenced by the lipoprotein charge. In order to clarify the correlation between the size and density in the HDL particles, we used fractions isolated by density ultracentrifugation. We first analyzed a mixture containing isolated VLDL (d < 1.006), LDL (d = 1.006-1.063), light HDL (d = 1.0631.125), and heavy HDL (d = 1.125-1.210) using BFM-μ-CE and achieved successful separation, especially clear separation between particles of heavy and light HDL (Figure 2A). Figure 2B, a shows magnified electropherograms of heavy HDL (HH) and light HDL (HL). Each HDL peak was identified by addition of the respective fraction. The corresponding peak increased as the content of light HDL increased. Since the later peak increased with the addition of light HDL, it was identified as light HDL (Figure 2B,b). The separation of light and heavy HDL particles was also successful in whole serum by BFM-μ-CE (Figure 3A), whereas

Figure 3. Separations of whole serum by (A) BFM-μ-CE system and by the (B) Lipophor system. The left lane in part B shows the same serum as in part A, and other lanes show isolated VLDL (V), LDL (L), light HDL (HL), and heavy HDL (HH).

such separation was not possible by conventional electrophoresis using the Lipophor system (Figure 3B). The band of HDL in whole serum (left lane in Figure 3B) was not split into two bands. The successful separation of HDLs by BFM is likely 1133

dx.doi.org/10.1021/ac102402c |Anal. Chem. 2011, 83, 1131–1136

Analytical Chemistry

TECHNICAL NOTE

Figure 4. Electropherograms of patient HDL (A) premedication and (B) postmedication. Each upper electropherogram shows heavy HDL (HH) and each lower electropherogram shows light HDL (HL).

caused by the smaller nanoscale mesh size relative to a conventional gel. Higher Density HDL Particles Do Not Necessarily Have a Smaller Size. An impressive data set obtained using this system is presented in Figure 4. Although the migration time of heavy HDL by μ-CE was found to be faster than light HDL in a dyslipidemic patient (Figure 4A), it was reversed completely after a month of fenofibrate-treatment (Figure 4B). Since the later peak increased as the concentration of heavy HDL increased, it was identified as heavy HDL. Inversion was recognized in all fenofibrate-treated subjects that were analyzed in our experiments. Four typical data sets are shown in Table 1. The high reproducibility of this inversion demonstrates heavy HDL is on average actually larger than light HDL. Such separation is not possible using a conventional electrophoresis Lipophor system (Jookoh, Co., Ltd., Japan) due to poor resolution and low reproducibility (Supplementary Figure 2 in the Supporting Information), possibly due to the larger mesh size relative to HDL particles. The migration times of heavy HDL and light HDL were not constant for both premedication patients (Supplementary Figure 2A in the Supporting Information) and postmedication patients (Supplementary Figure 2B in the Supporting Information). Is Heavy HDL Actually Larger than Light HDL? To confirm whether higher density HDL is actually larger than light HDL, the size and the shape of HDL were measured directly by scanning probe microscopy (SPM) and small-angle X-ray scattering (SAXS). Figure 5 summarizes the data on μ-CE migration time and the average size of heavy and light HDL. To maintain the physiological environment of HDL particles, we carried out SPM under liquid conditions instead of using negatively stained transmission electron microscopy (TEM) or scanning electron microscopy (SEM) under dry conditions. In premedication patient serum, the migration time of heavy HDL was faster than that of light HDL (Figure 5A,a), according to the size correlations estimated by both SPM (Figure 5A,b,c) and SAXS (Figure 5A,d). In contrast, postmedication results were discordant (Figure 5B). Although the estimated size of heavy HDL was larger than that of light HDL by SAXS (Figure 5B,d), the opposite results were obtained using SPM (Figure 5B,b). The average diameter of heavy HDL estimated by SPM was extremely small (5.8 nm). Further observation clarified that this SPM datum was for the z-axis, while the xy-axis size correlated with mobility (Figure 5B,c). Rather, the heavy HDL xy-axis was considerably larger than that of the

Table 1. Summary of Migration Times of Lipoproteins HH pre-

post-

a

HL

L

V

P1

58.0 ( 1.0

61.4 ( 2.1

72.0 ( 1.7

93.5 ( 3.3

P2 P3

57.7 ( 2.2 58.3 ( 2.1

61.5 ( 3.3 62.1 ( 0.7

80.3 ( 5.1 75.1 ( 2.9

94.5 ( 5.8 100.3 ( 5.6 86.7 ( 5.3

P4

60.2 ( 1.5

65.0 ( 1.3

76.0 ( 2.6

P1

61.5 ( 1.4a

59.0 ( 1.1a

77.6 ( 6.8

95.5 ( 8.6

P2

65.0 ( 3.2a

57.6 ( 1.8a

71.9 ( 6.7

102.7 ( 7.9

P3

65.7 ( 2.2a

60.1 ( 2.2a

76.4 ( 2.3

93.0 ( 6.2

P4

63.7 ( 1.0

59.3 ( 2.8a

75.1 ( 2.4

101.3 ( 11.2

a

Indicates the inversion. All data shown are the average of n = 10.

z-axis and was even larger than light HDL. The xy to z ratio for heavy HDL was 1.4 and that for light HDL was 1.1 in the premedication patients, whereas the xy to z ratio for heavy HDL was 1.7 and that for light HDL was 1.0 in the postmedication patients, indicating that heavy HDL, which was fenofibratetreated, is a disklike particle. These data indicate that some heavy HDLs are actually larger than light HDLs. The diameter size by SPM analysis usually adapts z-axis data when nanoscale size small particles are analyzed. However, our data indicate that adaptation of z-axis data may cause misleading results. Geometric Analysis Strongly Supports the Size Inversion. We conducted a geometric analysis to identify the source of the disagreement in size measurements and also to determine which data correlated with the electrophoresis results. Figure 6 summarizes the shapes analyzed by SPM. Figure 6A shows HDL from a premedication patient, and Figure 6B shows HDL from a postmedication patient. Each upper panel shows the heavy HDL (parts A,a or B,a of Figure 6), and each lower panel shows the light HDL (parts A,b or B, b of Figure 6). Each left panel shows the xy-plane picture (A(i) or B(i) of Figure 6), each middle panel shows the distributions of the z-axis sizes (parts A(ii) and B(ii) of Figure 6), and each right panel shows the estimated shape (parts A(iii) and B(iii) of Figure 6). To answer this question, we identified distinct types of heavy HDL in a postmedication patient, which displayed thin petal features (Figure 6Ba(i)). In contrast, light HDL in a postmedication patient had a smaller spherical shape (Figure 6Bb(i)). Since the xy-axis of heavy HDL was larger than that of light HDL (parts Ba(iii) and Bb(iii) of Figure 6), although the z-axis of heavy HDL was smaller than that 1134

dx.doi.org/10.1021/ac102402c |Anal. Chem. 2011, 83, 1131–1136

Analytical Chemistry

Figure 5. Summary of changes in each data set for heavy HDL (HH) and light HDL (HL) of (A) premedication and (B) postmedication. (a) μ-CE, migration time analyzed by the BFM-μ-CE; (b) SPMz, z-axis diameter measured by SPM; (c) SPMxy, xy-axis diameter measured by SPM; and (d) SAXS, radius measured by SAXS, which was calculated by the peak distance. Typical data from a patient are listed. An / indicates inversion. Each left column shows heavy HDL (HH) and each right column shows light HDL (HL).

of light HDL (parts Ba(ii) and Bb(ii) of Figure 6), heavy HDL migrated slower than light HDL. In this paper, in order to distinguish between size and density of HDLs, the microscopic larger size of HDL is termed “large HDL” and the smaller size HDL is termed “small HDL”. Therefore, heavy but large HDL may sometimes cause a crucial inversed result by conventional electrophoresis. Prospects for the Geometrical BFM-μ-CE Method. In conventional electrophoresis, DNA and proteins assume a spherical or string shape, for which the radius of inertia concept is applicable. However, direct lipoprotein samples have various shapes and charges. In particular, the immature stage of HDL is known to be a disklike shape of variable size as mentioned above. The mesh size of a typical electrophoresis gel is about 1 μm, whereas the diameter of HDL is 5-20 nm; therefore, separation of HDLs is difficult by conventional electrophoresis gel. In contrast, our BFM has a fine three-dimensional structure including several nanometers to 1 μm meshes. Therefore, fine separation of complex HDLs based on size and shape is possible.

TECHNICAL NOTE

Figure 6. Direct size and geometrical analyses using SPM of the (A) premedication patient and (B) postmedication patient for (a) heavy HDL (HH) and (b) light HDL(HL ). (i) The xy-plain picture of HDL morphologies analyzed by SPM, (ii) distribution of z-axis diameter analyzed by SPM, and (iii) estimated xyz-shapes. Each set of numerical data on the axis shows the radius of each axis, and the numerical data on the right below each panel shows the xy/z axis value.

The electrophoresis inversion suggests that the conventional method can occasionally produce serious mistakes in distinguishing between heavy HDL and light HDL. Size determination by SEM or TEM under dry conditions is also prone to errors. Since each HDL fraction has morphological variations, the average size calculated using a specified parameter may also conflict with SAXS and light scattering results. Indeed, conflicts were recognized in both SPM and SAXS, depending on the calculation method. To conclude, we accurately analyzed lipoprotein using a new BFM-μ-CE method, for comparison with comprehensive size measurements obtained using traditional instruments. Our data strongly indicated that higher density HDL particles do not necessarily have a smaller size. By BFM-μ-CE separation, we observed that HDL particles migrate according to their geometry and not necessarily according to their density. High density HDL has been generally considered to have a smaller particle size, and disparities in density and electrophoretic results have not been reported previously other than one study which reported a slight overlap.19 Therefore, this is the first demonstration of complete inversion. As our method is inexpensive, requiring only 1 μL of serum with the analysis being completed within 60-80 s, we anticipate this technique will be suitable for medical diagnosis. 1135

dx.doi.org/10.1021/ac102402c |Anal. Chem. 2011, 83, 1131–1136

Analytical Chemistry

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

TECHNICAL NOTE

(18) Manzato, E.; Zambon, S.; Marin, R.; Baggio, G.; Crepaldi, G. J. Lipid Res. 1986, 27, 1248–1258. (19) Musliner, T. A.; Krauss, R. M. Clin. Chem. 1988, 34 (8), B78–B83.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Present Addresses

O Nihon Medical Science Institute, 11-4 Minami-Tohrimachi, Kawagoe, Saitama, 350-0045, Japan.

’ ACKNOWLEDGMENT This work was supported by the Frontier Project “Adaptation and Evolution of Extremophile” from the Ministry of Education, Culture, Sports, Science and Technology of Japan. We greatly appreciate the assistance in the experiments of Ms. Noriko Fukushima at Saitama Medical University and Ms. Maki Watanabe, Mr. Takeshi Kato, and Mr. Yukihiro Atsuta at Rikkyo University. We greatly thank Dr. Yoshio Iwasaki and Ms. Yayoi Taniguchi in Rigaku Corporation for the SAXS analyses. We also thank Drs. Kayoko Tanaka and Masumi Akita at Saitama Medical University for their instruction in electron microscopic analysis. ’ REFERENCES (1) Pietzsch, J. S. S.; Nitzsche, S.; Leonhardt, W.; Schentke, K.-U.; Hanefeld, M. Biochim. Biophys. Acta 1995, 1254, 77–88. (2) Chapman, M. J.; Goldstein, S; Lagrange, D.; Laplaud, P. M. J. Lipid Res. 1981, 22, 339–358. (3) Warnick, G. R.; McNamara, J. R.; Oggess, C. N.; Clendenen, F.; Williams, P. T.; Landolt, C. C. Clin. Lab. Med. 2006, 26, 803–846. (4) Usui, S.; Hara, Y.; Hosaki, S.; Okazaki, M. J. Lipid Res. 2002, 43, 805–814. (5) Haginaka, J.; Yamaguchi, Y.; Kunitomo, M. J. Chromatogr., B 2001, 751 (1), 161–167. (6) Macfarlane, R. D.; Bondareako, P. V.; Cockrill, S. L.; Cruzado, I. D.; Koss, W.; McNeal, C. J.; Spiekerman, A. M.; Wakins, L. K. Electrophoresis 1997, 18, 1796–1806. (7) Weiller, B. H.; Ceriotti, L.; Shibata, T.; Rein, D.; Roberts, M. A.; Lichtenberg, J.; German, J. B.; Rooij, N. F.; Verpoorte, E. Anal. Chem. 2002, 74, 1702–1711. (8) Ceriotti, L.; Shibata; Folmer, B.; Weiller, B. H.; Roberts, M. A.; de Rooij, N. F.; Verpoorte, E. Electrophoresis 2002, 23, 3615–3622. (9) Ping, G.; Zhu, B.; Jabasini, M.; Xu, F.; Oka, H.; Sugihara, H.; Baba, Y. Anal. Chem. 2005, 77, 7282–7287. (10) Anderson, D. W.; Nichols, A. V.; Pan, S. S.; Lindgren, F. T. Atherosclerosis 1978, 29, 161–179. (11) Koizumi, J.; Mabuchi, H.; Yoshimura, A.; Michishita, I.; Takeda, M.; Itoh, H.; Sakai, Y.; Sakai, T.; Ueda, K.; Takeda, R. Atherosclerosis 1985, 58, 175–186. (12) Tabuchi, M. Nat. Biotechnol. 2007, 25, 389–390. (13) Tabuchi, M.; Kobayashi, K.; Fujimoto, M.; Baba, Y. Lab Chip 2005, 5, 1412–1415. (14) Tabuchi, M.; Baba, Y. Anal. Chem. 2005, 77, 7090–7093. (15) Andorson, D. W.; Nichols, A. V.; Forte, T. M.; Lindgren, F. T. Biochim. Biophys. Acta 1977, 493, 55–68. (16) Barlage, S.; Boettcher, D.; Boettcher, A.; Dada, A.; Schmitz, G. Cytometry, Part A 2006, 69A, 196–199. (17) Kontush, A.; Therond, P.; Zerrad, A.; Couurier, M.; NegreSalvayre, A.; de Souza, J. A.; Chanepie, S.; Chapman, M. J. Arterioscler. Thromb. Vasc. Biol. 2007, 27, 1843–1849. 1136

dx.doi.org/10.1021/ac102402c |Anal. Chem. 2011, 83, 1131–1136