Microchip-based human serum atherogenic lipoprotein profile analysis

Analytical Biochemistry 467 (2014) 75–83

Contents lists available at ScienceDirect

Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio

Microchip-based human serum atherogenic lipoprotein profile analysis Hua Wang a, Wei Zhang b, Jun Wan b, Weiwei Liu c,d, Bo Yu e, Qinghui Jin f, Ming Guan a,⇑ a

Department of Laboratory Medicine and Central Laboratory, Huashan Hospital, Shanghai Medical College, Fudan University, Shanghai 200040, China Biomedical Research Institute, Shenzhen PKU–HKUST Medical Center, Shenzhen, Guangdong 518036, China c Department of Laboratory Medicine, Shanghai Tenth People’s Hospital, Tongji University, Shanghai 200072, China d Department of Laboratory Medicine, Shanghai Skin Disease Hospital, Shanghai 200071, China e Department of Dermatology, Shenzhen Hospital, Peking University, Shenzhen, Guangdong 518036, China f Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Science, Shanghai 200050, China b

a r t i c l e

i n f o

Article history: Received 13 June 2014 Received in revised form 24 August 2014 Accepted 25 August 2014 Available online 9 September 2014 Keywords: Microchip Gold nanoparticles NBD C6-ceramide Atherogenic lipoprotein profile

a b s t r a c t Owing to the mounting evidence of serum lipid changes in atherosclerosis, there has been increasing interest in developing new methods for analyzing atherogenic lipoprotein profiles. The separation of lipoprotein and lipoprotein subclasses has been demonstrated using a microchip capillary electrophoresis (CE) system [Chromatographia 74 (2011) 799–805]. In contrast to this previous study, the current report demonstrates that sdLDL peak efficiencies can be improved dramatically by adding gold nanoparticles (AuNPs) to the sample. Moreover, NBD C6-ceramide was identified as a satisfactory dye for specific labeling and quantitation of individual serum lipoproteins. The accuracy of the method was evaluated by comparison with ultracentrifuge separated small, dense, low-density lipoprotein (sdLDL). A high correlation was observed between these two methods for sdLDL cholesterol. Lipid levels were investigated between atherosclerotic patients and healthy controls. The variation of serum atherogenic lipoprotein profiles for atherosclerotic patients pre- and post-treatment was assessed by microchip CE. This method has potential for the rapid and sensitive detection of different lipoprotein classes as well as their subclasses and, therefore, is suitable for routine clinical applications. Microchip-based atherogenic lipoprotein profile assays will greatly improve the analysis of risk factors in atherosclerosis and will provide useful information for monitoring the effect of therapies on atherosclerotic disease. Ó 2014 Elsevier Inc. All rights reserved.

Atherosclerosis is a chronic inflammatory disease of the large arteries driven by lipids [1]. The term ‘‘atherogenic lipoprotein profile’’ has been introduced to describe a common form of dyslipidemia characterized by three lipid abnormalities: increased triglyceride-rich very low-density lipoprotein (VLDL)1 levels, decreased high-density lipoprotein cholesterol (HDL-C) concentrations, and the presence of small, dense low-density lipoprotein (sdLDL) particles [2]. Determination of the circulating levels of plasma of the LDL subclass profile as well as plasma HDL and VLDL is important in the diagnosis of primary and secondary lipid transport disorders and in the risk assessment for atherosclerosis and ⇑ Corresponding author. Fax: +86 21 62481061. E-mail address: [email protected] (M. Guan). Abbreviations used: VLDL, very low-density lipoprotein; HDL-C, high-density lipoprotein cholesterol; sdLDL, small, dense low-density lipoprotein; lLDL, large buoyant LDL particles; CE, capillary electrophoresis; AuNP, gold nanoparticle; LPDS, lipoprotein-deficient serum; DMSO, dimethyl sulfoxide; SDS, sodium dodecyl sulfate; TC, total cholesterol; LDL-C, LDL cholesterol; sdLDL-C, sdLDL cholesterol; HDL-C, HDL cholesterol; CA–IMT, carotid artery intima media thickness; RSD, relative standard deviation. 1

http://dx.doi.org/10.1016/j.ab.2014.08.031 0003-2697/Ó 2014 Elsevier Inc. All rights reserved.

coronary heart disease [3]. LDL particles are heterogeneous with respect to size and density of lipid composition. Two distinct phenotypes based on LDL particles have been recognized: pattern A, with a higher proportion of large buoyant LDL particles (lLDL), and pattern B, with a predominance of small, dense LDL particles (sdLDL). Intense clinical interest in the measurement of LDL subclasses stems from a strong and consistent association between a predominance of sdLDL and increased risk of coronary heart disease [4]. In addition, VLDL tends to promote atherosclerosis [5], and VLDL levels are more highly correlated with atherosclerosis lesion area in the aortic root than LDL [6]. Many epidemiological studies have confirmed the early observation of Barr and coworkers that the concentration of HDL cholesterol is inversely correlated with the risk of premature cardiovascular disease [7]. The main methods for separation and analysis of plasma lipoprotein levels are based on differences in physical properties and include ultracentrifugation, electrophoresis, and differential precipitation. However, these methods are labor-intensive, time-consuming, and costly. Nuclear magnetic resonance (NMR) is the most rapid and convenient method for determining LDL size and subfraction concentration [8]. However, it is limited by lack of

76

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83

published data on detailed procedures, calibration, and validation, which are expected when novel methods are established. A method for the separation of the predominant LDL subclasses, HDL, and VLDL using a single-step iodixanol gradient has also been described [9]. However, its routine use is still limited to the analysis of serum lipoprotein. Free zone electrophoresis in capillaries and microfluidic channels has emerged as an important separation technique because of its highly efficient separation, small injection volumes, and short analysis times [10]. Liu and coworkers demonstrated that capillary electrophoresis (CE) can be used to separate two LDL particles with significantly different charge/volume ratios given that the different ratios result in different electrophoretic mobilities [11]. Schmitz and coworkers developed a capillary isotachophoresis procedure for the separation of lipoproteins into 14 subclasses [12]. However, the addition of up to nine spacers to the running buffers was required, making this method complicated and less practical. Weiller and coworkers first used microchip CE to separate HDL and LDL [13,14]. However, HDL and LDL were not baseline resolved, and reproducibility was poor. Ping and coworkers used a polymethylmethacrylate (PMMA) chip for lipoprotein separation [15]. Baseline separation of standard lipoproteins, including HDL, LDL, and VLDL, were achieved with different selectivities. However, Ping and coworkers did not determine the subclasses of lipoprotein. We have previously shown that the lipoprotein and lipoprotein subclasses can be determined by microchip CE [16,17]. However, it is still difficult to completely separate all of the lipoprotein classes and subclasses synchronously. Recently, we demonstrated that microchip CE could be used to analyze subclasses of LDL, VLDL, and HDL [18]. However, poor sdLDL peak efficiencies were not sufficient for quantitative analysis and, thus, need to be improved. In contrast to this previous study, the current report demonstrates that sdLDL peak efficiencies can be improved dramatically by simply adding gold nanoparticles (AuNPs) to the sample buffer at a concentration of 80 nmol/L. In microchip CE analysis, lipoproteins are stained with the lipophilic dye NBD C6-ceramide and monitored by laser-induced fluorescence detection, allowing for lipoprotein analysis without prior separation from other serum proteins. However, it is unclear whether lipoprotein fractions separated by microchip CE are equivalent to those in lipoprotein fractions separated by ultracentrifugation. Although microchip CE separates serum lipoproteins into four fractions, there is still no direct evidence that these fractions represent lipoproteins rather than other serum proteins. Therefore, in the current study, we demonstrated the utility of NBD C6-ceramide as a specific stain for serum lipoproteins on microchip CE. We also confirmed that NBD C6-ceramide shows a saturation limit for lipoprotein labeling. In addition, we examined the linearity of the relation between levels of lipoprotein fractions and lipoprotein cholesterol to evaluate the utility of microchip-based lipoprotein analysis. This microchip-based method is able to detect the atherogenic lipoprotein phenotype that is most strongly correlated to atherosclerosis progression in patients.

Materials and methods Microdevices The design of the microchip was described in a previous study [17]. The chip’s microchannels were fabricated in a quartz glass substrate (63.5  31.7  2 mm3) by a standard photolithography and chemical wet etching process. The microchannel dimensions were 21 ± 2 lm in depth and 100 ± 2 lm in width at half of the depth. Holes 2 mm in diameter were drilled at microchannel terminals and used as reservoirs. The sample injection channel length

was 28 mm, the total separation channel length was 76.9 mm, and the effective separation length was 42.5 mm. Details of the integrated microchip capillary electrophoresis detection microsystems were described previously [19]. Reagents and buffer solutions Lipoprotein-deficient serum (LPDS) was obtained from EMD Millipore (Billerica, MA, USA). According to product specifications, the original concentration of LPDS was 35 mg/ml. Standard solutions of HDL, LDL, and VLDL were obtained from Sigma Chemical (St. Louis, MO, USA). According to product specifications, the original concentrations of HDL, LDL, and VLDL were 5.9, 7.6, and 1.92 mg/ml, respectively. Fluorescent dye, NBD C6-ceramide, was purchased from Molecular Probes (Eugene, OR, USA). Ethylene glycol and dimethyl sulfoxide (DMSO) were purchased from Sigma. Citrate-stabilized colloidal gold nanoparticles (5 nm diameter, 0.01% concentration as HAuCl4, CAS no. 7440-57-5) were purchased from Sigma and served as a stock solution. The separation buffer used in this experiment was a mixture of 40 mmol/L Tricine and 40 mmol/L methylglucamine (both obtained from Sigma). The pH value of the separation buffer was adjusted to 9.8 by adding NaOH. Before electrophoresis separation, 0.1 mmol/L sodium dodecyl sulfate (SDS) and 80 nmol/L AuNPs were added to the sample buffer and 0.02 mmol/L SDS and 20 nmol/L AuNPs were added to the running buffer. Serum collection and biochemical analysis The study included 60 control subjects and 60 patients. This study was approved the ethics committees of Fudan University Huashan Hospital, and samples were collected only after the participants had given their informed consent. Fasting blood samples were collected from healthy volunteers and patients attending the university hospital. The 60 healthy subjects (32 men and 28 women) between 41 and 65 years old had no history of dyslipidemia, obesity, diabetes, or any known disease, were not taking drugs known to affect plasma lipids (e.g., hormone therapy, hypolipidemic drugs), and had only moderate alcohol and tobacco consumption. The 60 patients (35 men and 25 women), ranging from 45 to 66 years old, had carotid atherosclerosis and hypercholesterolemia. Atorvastatin was administered at 10 mg/day for 3 months, and the serum lipoprotein profiles and carotid atherosclerotic plaque were monitored pre- and post-treatment. Serum was prepared by low-speed centrifugation at 4 °C. Samples for analysis of microchip CE were stored at 4 °C, and analysis was completed 4 h after blood extraction. Aliquots of unused sample were stored at –80 °C. lLDL (d = 1.019–1.044 g/ml) and sdLDL (d = 1.044–1.063 g/ml) were separated from serum by sequential flotation in an ultracentrifuge (Himac CS120GX) with an S100AT6 rotor (both obtained from Hitachi Koki, Tokyo, Japan) according to the method of Havel and coworkers [20]. Briefly, the density of serum was raised by the addition of concentrated salt solution. The stock salt solution contained 153.0 g sodium chloride and 354.0 g potassium bromide per liter (density: 1.346). Solutions of lower density were prepared by dilution of the stock solution with 0.15 M sodium chloride solution (density: 1.005) according to the formula reported by Havel and coworkers [20]. Ultracentrifugation was carried out in the S100AT6 rotor. Serum (4 ml) was delivered into a capped plastic tube from a calibrated syringe. The appropriate volume of salt solution was added, and the tube was filled with a small amount of salt solution of the same density as the mixture. After centrifugation at 105,000g for 16 h at 12 °C, lipoproteins of less than solvent density were concentrated in a layer at the top of the tube. Beneath this was a clear colorless region occupying approximately one-fourth of the length of the tube; the remainder of the serum was stratified below

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83

77

the clear zone. In this way, it is possible to separate lLDL and sdLDL fractions from the same serum sample. lLDL and sdLDL were stored in aliquots at 4 °C after the addition of ethylenediaminetetraacetic acid (EDTA). Serum levels of total cholesterol (TC), LDL cholesterol (LDL-C), sdLDL cholesterol (sdLDL-C), and HDL cholesterol (HDL-C) were measured by enzymatic methods. Measurements were performed with an automatic chemistry analyzer (Modular P800). Common carotid artery lesions and carotid artery intima media thickness (CA–IMT) were assessed by high-resolution B-mode ultrasonography on an LOGIQ E9 instrument (GE, USA). Derivatization and electrophoresis procedure Lipoprotein fractions or serum were stained as follows. The sample solution (1 ll) was diluted with deionized water (3 ll, 1:3, v/v) and incubated with a half-volume (2 ll) of NBD C6-ceramide solution (0.5 mg/ml in ethylene glycol/DMSO, 9:1, v/v) for 1.0 min. The solution was then mixed with 14 ll of sample buffer before separation. Prior to sample solution injection, the chip’s microchannels were flushed with 1 M NaOH for 1 min, rinsed with deionized water for 2 min, and subsequently rinsed with running buffer for 5 min. Before electrophoresis separation, 5 ll of running buffer was injected into the buffer reservoir, buffer waste reservoir, and sample waste reservoir. Then 5 ll of sample solution was injected into the sample reservoir after the chip’s microchannels were filled with running buffer. A voltage of +700 V was applied to both the sample reservoir and sample waste reservoir for 40 s to fill the intersection. Simultaneously, a voltage of +300 V was applied to the buffer reservoir and a voltage of +450 V was applied to the buffer waste reservoir to prevent the sample in the intersection from spilling into the separation channel. Next, +3000 V was applied to the buffer reservoir and buffer waste reservoir to inject the plug into the separation channel. Meantime, +300 V was applied to both the sample and waste reservoirs to push the sample back from the intersection, thereby preventing leakage of the sample into the separation channel. The equivalent separation field strength, E, was 403 V/cm. Statistical analysis The data were analyzed using the statistical package IBM SPSS Statistics 20.0. The results are presented as the means ± standard deviations. Independent-sample t tests were used to compare the mean values of variables between the atherosclerosis and control samples. Paired t tests were used to compare the mean values of variables between the atherosclerosis patients pre- and posttreatment. Correlations between variables were analyzed using Pearson’s correlation test. A level of P < 0.05 was considered to be statistically significant. The 33.3th and 66.7th percentiles were used to produce tertiles of CA–IMT. Results and discussion AuNPs improved lipoprotein peak efficiencies Recently, we demonstrated that the resolution of lipoproteins could be improved by adding AuNPs to the running buffer [18]. However, the sdLDL peak efficiencies were poor and, thus, need to be improved. To improve separation efficiency of sdLDL, we added a small amount of AuNPs to the sample. As shown in Fig. 1B, after adding 80 nmol/L AuNPs into the sample solution, the sdLDL peak underwent a focusing effect. An apparent efficiency of 3.1  104 plates (N) was obtained. This peak sharpening observed for the VLDL fraction on AuNPs addition resembles that which occurs for sdLDL. This focusing effect was not seen when

Fig.1. Influence of AuNPs on the sdLDL peak efficiencies. (A) Without AuNPs in the sample buffer. (B) With AuNPs in the sample buffer. Separation conditions: the sample solution contained 0.1 mM SDS, 80 nM AuNPs, 40 mM Tricine, and 40 mM N-methyl-D-glucamine, and the running buffer contained 0.02 mM SDS, 40 mM Tricine, 40 mM N-methyl-D-glucamine, and 20 nM AuNPs (pH 9.8, E = 403 V/cm).

AuNPs were present only in the running buffer. It is postulated that the sharp peak detected was due to intact AuNP-bound lipoprotein particles stained with NBD C6-ceramide. As expected, in the presence of AuNPs, the migration time of lLDL was shifted from 1.6 to 1.7 min, sdLDL was shifted from 1.7 to 1.8 min, and VLDL was shifted from 1.9 to 2.0 min, supporting the assumption that AuNPs have been adsorbed onto the particles, varying particle size and charge and, therefore, particle mobilities. AuNPs as the additive in the running buffer have been successfully used for electrophoretic separations and detection of DNA [21] and proteins [22]. The current report demonstrates that lipoprotein peak efficiencies can be improved dramatically by simply adding to the sample. To investigate the overall reproducibility of this assay method for serum lipoproteins, reproducibility was tested for lipoprotein fractions prepared from five replicate aliquots of the mixtures of lLDL, sdLDL, VLDL, and HDL. The relative standard deviations (RSDs) of migration times (n = 5) were less than 2.8% for run to run, 4.3% for day to day, and 4.8% for chip to chip. The RSDs of peak area (n = 5) were from 3.5 to 5.0% for run to run, from 4.8 to 5.5% for day to day, and from 5.0 to 5.6% for chip to chip. Thus, the proposed method had good stability and reproducibility.

Pre-staining with NBD C6-ceramide Lipoproteins are complex nanometer-sized particles consisting of apolipoproteins, cholesterol, and a phospholipid monolayer on

78

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83

the surface. The interior of lipoproteins consists of triglycerols, cholesterol, and cholesterol esters. For lipoprotein detection, specific labeling of particular lipoproteins is very important. Weiller and coworkers reported that NBD C6-ceramide is a lipophilic dye that specifically stains lipoproteins over other serum components [13]. However, the staining mechanism of NBD C6-ceramide was not investigated in their research. In our experiment, we tested the utility of NBD C6-ceramide as a lipoprotein-specific stain in microchip CE. The electropherogram for LPDS is a smooth baseline without any signal with NBD C6-ceramide staining (Fig. 2A). Fig. 2B shows the electropherogram of a mixture of LPDS and standard LDL. As shown, standard LDL fractions were separated into two peaks by microchip CE: a major component (peak 1) with higher mobility and a minor component (peak 2) with lower mobility. Fig. 2 also shows the electropherograms of a mixture of LPDS and VLDL (panel C) and a mixture of LPDS and HDL (panel D). VLDL (peak 3) with a lower mobility than the minor LDL component (peak 3) and with a higher mobility than the HDL component (peak 4) was observed. Shown in Fig. 3 are microchip CE data for mixtures of LPDS, LDL, VLDL, and HDL using fluorescent staining. Two subclasses describe the subpopulations of LDL (peaks 1 and 2, lLDL and sdLDL), with one subclass representing VLDL (peak 3) and one subclass representing HDL (peak 4). These results indicate that NBD C6-ceramide can be used for specific labeling of the lipoproteins in our system. By considering the chemical structure of NBD C6-ceramide, we propose a specific interaction between lipoprotein and NBD

Fig.3. Electropherograms of mixtures of LPDS, LDL, VLDL, and HDL. Peaks 1 and 2: lLDL and sdLDL; peak 3: VLDL; peak 4: HDL. Separation conditions were as described in Fig. 1.

C6-ceramide. As shown in Fig. 4, NBD C6-ceramide is a fluorescent sphingolipid analogue [23] that is uncharged and lipophilic. This dye can associate with lipoprotein particles via hydrogen bonding and hydrophobic interactions with the phospholipid/cholesterol membrane. The hydrogen bonding interaction between the hydroxyl group of NBD C6-ceramide and the phosphate group of phospholipids, in addition to the hydrophobic interactions between the long alkyl chains of NBD C6-ceramide and phospholipid, can

Fig.2. Electropherograms of LPDS (A), mixtures of LPDS and LDL (B), mixtures of LPDS and VLDL (C), and mixtures of LPDS and HDL (D). Separation conditions were as described in Fig. 1.

79

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83

O2N

N

O2N

O N

A

N

NH

N+

N+

O

-O

O H

O O

O

O

N+

O

HN O P H O O O

O

HN O P O H O O

O-

O P O O

O-

O H

O

O

O N

NH

N+

HN O P O H O O

N

NH

B Hydrophilic

O2N

O N

O-

O H

O O

O

O

O

O

O

O

O

O

Hydrophobic

Fig.4. Schematic diagram of the interaction between phospholipids and NBD C6-ceramide: (A) NBD C6-ceramide; (B) phospholipid. NBD C6-ceramide associates with lipoprotein particles through hydrogen bonding and hydrophobic interactions with the phospholipid membrane.

d

O2N

N O N

O2N

N

O N

NH

m x

m x

O

O

O

O

O-

N+

N O N

O

O

O O

P

N+

O O-

O

O

O O

O

O

O O P OO

O O O P O O

O

O O H O OO P O

O

n m x

O R1 HN

NO2 N

O

*

O

NH

HN

N+

O O O O P O H O O

NH

O

O

O

O P

O

O

O-

O

H O

N+

O

O

O

O O NH

N O N

H O O

P

N+ O2N

O

O O-

O

O NH

O

O

O

O H

O O

O

O H O P OO

O

H O

O

O

O O P H O O N+

O

O

N+ OH

H

P

P

NO2 N H

N+

O

H H N O

N H

OO

N N O

O-

NO2 N

O

N

N+

NH

O O H O OO P O

O HN N+

NO2 N

NH

O

N

O

HN N O N

NO2

HN N O N

HN NO2

N O

N

NO2

(B) Sketch map of saturation

Peak area (x 100)

(A) Sketch map of unsaturation

O H

O

O O

O O H O O

HN N+

OH

N+

O O O

O

* NH

O

R2

N

O2N

O

O

O O

R3 NH

O

O

n

O

N+

O

F N+

O

O

N+

O

O H

O

O

O-

O P O O H O H N

O

O

O H O O P O O-

N+

N N O

N H

H N O H

O

n

N H

H N

O2N

x

m

n

n

n m

x

x

m

n

x

n m

O P

O

x

O

HN

O

O

O

H

O-

NO2

P

H O

O

m

O

O H O

O O

n

O

O

N O

O

x

NH

O

P

O O

O O

n

O

O

O

m

O O O O P O H O

O

m

N+

x

O

* NH

O

R2

n

HN

x

NH

O N N O

O-

O

O

m

x

n m x

N+

O

O

O

x

n

P

O O O

H

O

O

m

O

O

O

O

f

n

c

O

O O

R3 O R1 *

O

m

N+

O

O

O

n

x

O

O

O O P O O-

O

O H O O P O -O H O

N+

O-

m

n

n m

N

N+

x

n

m

O

O

O

n

x

n

O

O

O

n

m

HN

O

O

N H

n

O

P

x

O

O

O

H O

O

H N

O2N

n

NH

H -O O O P O H O

x

n

O -O

O N N

x

O

O

N H

N+

O

O

P

H O

O

N+

OO P O O

O O

O -O

H N

O O

n

N

m

O

O

m

n

NO2 N+

O O

x

n

m

m

N O2N

H

O H

O

O N+

N N O

O O

O O P O -O

N H

H N O

m

O

O

NO2

P

H O

m

m

m

O

x

O

O

O O

x

HN

O x

x

x

O

O

O

O

O

O

m

O

O

O

O

O O

O

O

O

O

H

O

H O

O

H O

n

O

O

N+

O-

O

x

P

O m

n

n

-O

x

n

e

n

n m x

m

H O

N+

m

O O

O

N+

x

a

x

O

O

N H

O

NH

N+

O H -O O P O O

n

n

O

P

H O

O

O

n

O O -O P O H

O HN

O

m

O -O

H N

N+ NH

x

n

N

-O O O P O

O

O

O O

O H

m

N

O

N O

N

O O

P

O

HN

x

O

O

O

O2N

N+ O

x m

m

O N+ N

N

O

O m

m

HN

n

x

O

O

O

O

O P

O

m

O

O O

O x

x

O

-O

HN

O H O -O P O H O

x

O

O2N

O O

O

O

O

O

O

N+

O

n

O O

O O O -O P O

O

-O

m

N+

O

P

O -O P O O H O

x

O -O P O O

-O

O

N+

n

O -O P O O

N+

NO2

NH

N+

OH

m

b

N+

OH

N+

x

N+

NH

N O N

O 2N

(C) NBD C6- ceramide (mg/ml) Fig.5. Saturation of NBD C6-ceramide as a lipoprotein dye. (A, B) A schematic diagram of unsaturation (A) and saturation (B) between phospholipids and NBD C6-ceramide is shown. (C) A certain amount of serum was incubated with increasing amounts of the indicated dye and subsequently analyzed by microchip electrophoresis. a: NBD C6-ceramide; b: phospholipid; c: apolipoproteins; d: cholesterol; e: cholesterol esters; f: triglycerols.

80

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83

Fig.6. Correlation analysis of HDL and sdLDL cholesterol levels determined by microchip electrophoresis and enzymatic methods (n = 53). (A) HDL cholesterol was determined by enzymatic methods. (B) sdLDL (d = 1.044–1.063 g/ml) was separated from serum by sequential flotation in an ultracentrifuge and measured by enzymatic methods.

Table 1 Characteristics of study subjects. Variable

Age (years) Males/Females Atherosclerosis risk factors Hypertension Tobacco use Diabetes mellitus Familial hypercholesterolemia Medications Statins Other hypocholesterolemic agents Beta blockers Past medical history Myocardial infarction Cerebral infarction

Atherosclerosis patients

Control subjects

(n = 60)

(n = 60)

60.0 ± 9.1 35/25

59.2 ± 10.5 32/28

42 (70) 20 (33) 11 (18) 6 (10)

0 0 0 2

46 (77) 3 (5)

0 (0) 0 (0)

40 (67)

0 (0)

8 (13) 5 (8)

0 (0) 0 (0)

(0) (0) (0) (3)

Note. Data are presented as mean values ± standard deviations or number (%) of subjects.

form a highly stable interaction. Based on this mechanism, the amounts of NBD C6-ceramide absorbed by lipoprotein will increase with the concentration of lipophilic dye at low NBD C6-ceramide concentrations (Fig. 5A). However, the absorption will reach a maximum when the concentration of NBD C6-ceramide is oversaturated (Fig. 5B). To validate this hypothesis, the response peak areas of the lipoproteins increasing lipophilic dye concentration were calculated. Serum of an atherosclerosis patient was incubated with increasing amounts of NBD C6-ceramide and subsequently separated by microchip CE. As shown in Fig. 5C, for the same amount of serum, the peak areas of VLDL, HDL, LDL, and total lipoproteins increased with increasing NBD C6-ceramide concentrations up to 0.4, 0.45, 0.5, and 0.5 mg/ml, respectively. NBD C6-ceramide, therefore, exhibits a saturation point for lipoprotein labeling, and higher amounts of NBD C6-ceramide do not yield an increase in microchip CE peak areas. It should be noted that VLDL and HDL reach saturation before LDL. This phenomenon is likely due to a low concentration of VLDL and HDL in serum samples, which absorb a small amount of NBD C6-ceramide. The amounts of the dye correlate best with the amounts of phospholipids in

Fig.7. (A) Comparison of lipoprotein fraction peak areas obtained from atherosclerosis patients (black bars) and healthy controls (gray bars). (B) Comparison of lipoprotein fraction peak areas obtained from atherosclerosis patients pre-treatment (black bars) and post-treatment (gray bars). AS, atherosclerosis. ⁄P < 0.05; ⁄⁄P < 0.01; ⁄⁄⁄P < 0.001.

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83 Table 2 Lipoprotein fraction peaks as measured by microchip CE according to tertiles of carotid artery intimal media thickness. Pa

Tertiles of CA–IMT

lLDL sdLDL VLDL HDL

Low (<0.65 mm)

Middle (0.65– 0.85 mm)

High (>0.85 mm)

0.45 ± 0.05 0.24 ± 0.04 0.08 ± 0.01 0.29 ± 0.04

0.46 ± 0.05 0.28 ± 0.06 0.10 ± 0.02 0.26 ± 0.02

0.47 ± 0.06 0.35 ± 0.07 0.12 ± 0.02 0.24 ± 0.02

n.s. <0.01 <0.05 <0.05

Note. ns. nonsignificant. a Assessed by an analysis of covariance.

the samples. The amounts of phospholipids differ among the different lipoprotein classes and even among large and small dense LDL. Therefore, the saturation point of the dye correlates with the amount of serum lipoproteins, and the same amount of dye is used in serum having different amounts of the lipoproteins. An NBD C6-ceramide concentration of 0.5 mg/ml, therefore, was selected for the following experiments. All of these results are consistent with the proposal in Fig. 5A and B. Meantime, these results indicate that NBD C6-ceramide is a satisfactory dye for enhancing the sensitivity of lipoprotein detection and can be used for the quantification of serum lipoprotein. Methodology comparison We examined whether microchip CE and the clinical routine method were related. Lipoprotein data obtained by microchip CE

81

analysis were compared with enzymatic methods for HDL and sdLDL quantitation. Serum levels of HDL-C were measured by enzymatic methods. sdLDL (d = 1.044–1.063 g/ml) was separated from serum by sequential flotation in an ultracentrifuge and then measured by enzymatic methods. The correlation of HDL-C and sdLDL-C determined by microchip CE and the enzymatic method is shown in Fig. 6 for a group of 53 samples obtained from patients at our university hospital. We observed a markedly high correlation between these two methods. The correlation coefficients for HDL-C and sdLDL-C are 0.96 and 0.91, respectively, demonstrating good agreement between the microchip CE and clinical routine methods. This finding indicates that microchip CE is an accurate and sensitive method for quantifying lipoprotein classes and subclasses. In addition, microchip CE has the advantage that LDL subclasses can be measured directly in human serum. Application In this study, we used an optimized method to examine the clinical applications of microchip CE. The current study included 60 atherosclerosis patients and 60 control subjects. The characteristics of the patients and control subjects are shown in Table 1. As shown in Fig. 7A, compared with controls, atherosclerosis patients had significantly higher levels of sdLDL (0.33 ± 0.07 vs. 0.24 ± 0.04, P < 0.001) and VLDL (0.11 ± 0.02 vs. 0.08 ± 0.02, P < 0.001). Mean levels of lLDL were increased, whereas HDL levels were decreased, in atherosclerosis patients compared with controls (nonsignificant). We analyzed the lipoprotein levels of a group of atherosclerosis patients at pre- and post-treatment (n = 35). Compared with the pre-treatment samples, post-treatment samples had significantly

Fig.8. Results of serum lipoprotein profiles and the carotid atherosclerotic plaque of a patient pre- and post-treatment. (A) Lipoprotein fractions in the patient before treatment. (C) Lipoprotein fractions in the patient after 3 months of treatment with atorvastatin at 10 mg/day. Compared with pre-treatment (B, 29.7  3.7 mm), patients had significantly lower levels of atherosclerotic plaques (D, 16.1  1.8 mm). Separation conditions were as described in Fig. 1.

82

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83

lower levels of lLDL (0.45 ± 0.05 vs. 0.47 ± 0.07, P < 0.05), sdLDL (0.23 ± 0.04 vs. 0.34 ± 0.07, P < 0.001), and VLDL (0.08 ± 0.02 vs. 0.11 ± 0.02, P < 0.001) as well as significantly higher levels of HDL (0.26 ± 0.04 vs. 0.24 ± 0.02, P < 0.01) (Fig. 7B). These results indicated higher levels of sdLDL and VLDL in atherosclerosis patients than in healthy participants. In addition, the results suggest that lowering VLDL and sdLDL levels might be an effective therapy against atherosclerosis patients after controlling serum LDL cholesterol level, consistent with reports in the literature [24,25]. Clinical and epidemiological studies have consistently shown that HDL is a satisfactory therapeutic target to further reduce the residual cardiovascular risk [26,27]. In addition, clinical trials with new HDL-raising drugs are currently under way to provide definitive evidence that increasing HDL will reduce cardiovascular events [28]. Our results indicated that atherosclerosis patients exhibited significantly higher levels of HDL post-treatment. An atherogenic lipoprotein phenotype has been shown to consist of three elements that may raise cardiovascular risk: a raised VLDL level, a reduced HDL level, and an abundance of sdLDL [29]. Furthermore, we analyzed the serum lipoprotein profiles, CA–IMT, and carotid atherosclerotic plaque levels of a group of carotid atherosclerosis patients at pre- and post-treatment (n = 35). CA–IMT was measured at points 20, 25, and 30 mm proximal to the flow divider on the far wall of the right and left common carotid arteries at the end of the diastolic phase. Using this information, mean CA–IMT was determined for each patient. Table 2 shows that mean levels of sdLDL and VLDL increased and HDL decreased across tertiles of CA–IMT. The results showed that high levels of sdLDL and VLDL in combination with the low HDL level are strongly associated with an increase in the risk of carotid atherosclerosis. A patient was selected to examine the merits of this assay method. The patient was a 62-year-old man with severe carotid atherosclerosis and hypercholesterolemia. Atorvastatin was administered at 10 mg/day for 3 months, and the serum lipoprotein profiles and carotid atherosclerotic plaque were monitored pre- and post-treatment. The serum lipoprotein fractions in the patient after 3 months of treatment with atorvastatin are shown in Fig. 8B, and those before atorvastatin treatment are shown in Fig. 8A. It is obvious that the peak areas of sdLDL and VLDL are significantly reduced with atorvastatin, whereas the HDL fraction areas appeared to be increased after treatment. Fig. 8C and D show the results of the patient with the severe carotid syndrome. After treatment with atorvastatin, carotid atherosclerotic plaques were significantly reduced. Compared with prior treatment, the patient had significantly lower levels of atherosclerotic plaques (29.7  3.7 vs. 16.1  1.8 mm). These results obtained by ultrasonic detection are consistent with the results obtained by microchip CE, indicating that the microchip-based lipoprotein assay gives useful information in the risk assessment and observation of the effects of therapies on atherosclerotic disease. Therefore, the assay demonstrated here has the potential for rapid analysis and highly efficient detection of different clinical relevance lipoprotein forms.

Conclusions Lipoproteins are macromolecular complexes of lipids and proteins held together by hydrophobic and electrostatic forces. Lipoprotein analysis can be very useful for clinical diagnosis of atherosclerosis and related diseases. Current methods for analysis of the atherogenic lipoprotein profile levels are labor-intensive, time-consuming, and costly. Microfluidic devices appear to be a promising platform to perform such analyses rapidly and efficiently. Recently, we demonstrated that a quartz microchip can be used for serum lipoprotein analysis. In contrast to previous studies, the current article reported that the sdLDL peak undergoes

a focusing effect and exhibited an apparent efficiency of 3.1  104 plates (N). Moreover, we demonstrated the utility of NBD C6-ceramide as a lipoprotein-specific stain for microchip CE. To evaluate the ability of microchip-based lipoprotein analysis, we examined the linearity of the relationship between levels of lipoprotein fractions and lipoprotein cholesterol. Furthermore, we determined that the assay of serum atherogenic lipoprotein profile provides useful information for risk assessment and observation of the effect of therapies on atherosclerotic disease. This work clearly indicates that microchip CE may be a valuable tool for a fast, reliable, and automatable analysis of lipoprotein classes and subclasses in the clinical laboratory.

Acknowledgments This work was supported by grants from the Shanghai Municipal Health Bureau new ‘‘one hundred talents’’ program (XBR2011044), Shanghai shen-kang hospital development center at the municipal level hospital Clinical auxiliary departments ability construction project (SHDC22014001), the Development Program of Clinical Medical College of Yangzhou University (yzucms201041), the Outstanding Young Talent Plan of Shanghai (XYQ2013095), and the Society Development Foundation of Jiangsu Province (BE 2010679). We thank the NPG language service for editing the article.

References [1] F.K. Swirski, M. Nahrendorf, Leukocyte behavior in atherosclerosis, myocardial infarction, and heart failure, Science 339 (2013) 161–166. [2] M. Rizzo, K. Berneis, Lipid triad or atherogenic lipoprotein phenotype: a role in cardiovascular prevention?, J Atheroscler. Thromb. 12 (2005) 237–239. [3] S.M. Grundy, T. Bazzarre, J. Cleeman, R.B. D’Agostino Sr., M. Hill, N. HoustonMiller, W.B. Kannel, R. Krauss, H.M. Krumholz, R.M. Lauer, I.S. Ockene, R.C. Pasternak, T. Pearson, P.M. Ridker, D. Wood, Prevention conference V: beyond secondary prevention: identifying the high-risk patient for primary prevention: medical office assessment: writing group I, Circulation 101 (2000) E3–E11. [4] K.K. Berneis, R.M. Krauss, Metabolic origins and clinical significance of LDL heterogeneity, J. Lipid Res. 43 (2002) 1363–1369. [5] R. Carmena, P. Duriez, J.C. Fruchart, Atherogenic lipoprotein particles in atherosclerosis, Circulation 109 (2004) 1112–1117. [6] P.A. VanderLaan, C.A. Reardon, R.A. Thisted, G.S. Getz, VLDL best predicts aortic root atherosclerosis in LDL receptor deficient mice, J. Lipid Res. 50 (2009) 376– 385. [7] D.P. Barr, E.M. Russ, H.A. Eder, Protein–lipid relationships in human plasma. II. In atherosclerosis and related conditions, Am. J. Med. 11 (1951) 480–493. [8] P.T. Williams, X.Q. Zhao, S.M. Marcovina, J.D. Otvos, B.G. Brown, R.M. Krauss, Comparison of four methods of analysis of lipoprotein particle subfractions for their association with angiographic progression of coronary artery disease, Atherosclerosis 233 (2014) 713–720. [9] A. Sawle, M.K. Higgins, M.P. Olivant, J.A. Higgins, A rapid single-step centrifugation method for determination of HDL, LDL, and VLDL cholesterol, and TG, and identification of predominant LDL subclass, J. Lipid Res. 43 (2002) 335–343. [10] X. Niu, F. Pereira, J.B. Edel, A.J. Mello, Droplet-interfaced microchip and capillary electrophoretic separations, Anal. Chem. 85 (2013) 8654–8660. [11] M.Y. Liu, C.J. McNeal, R.D. Macfarlane, Charge density profiling of circulating human low-density lipoprotein particles by capillary zone electrophoresis, Electrophoresis 25 (2004) 2985–2995. [12] G. Schmitz, C. Mollers, V. Richter, Analytical capillary isotachophoresis of human serum lipoproteins, Electrophoresis 18 (1997) 1807–1813. [13] B.H. Weiller, L. Ceriotti, T. Shibata, D. Rein, M.A. Roberts, J. Lichtenberg, J.B. German, N.F. de Rooij, E. Verpoorte, Analysis of lipoproteins by capillary zone electrophoresis in microfluidic devices: assay development and surface roughness measurements, Anal. Chem. 74 (2002) 1702–1711. [14] L. Ceriotti, T. Shibata, B. Folmer, B.H. Weiller, M.A. Roberts, N.F. de Rooij, E. Verpoorte, Low-density lipoprotein analysis in microchip capillary electrophoresis systems, Electrophoresis 23 (2002) 3615–3622. [15] G. Ping, B. Zhu, M. Jabasini, F. Xu, H. Oka, H. Sugihara, Y. Baba, Analysis of lipoproteins by microchip electrophoresis with high speed and high reproducibility, Anal. Chem. 77 (2005) 7282–7287. [16] H. Wang, H.M. Wang, Q.H. Jin, H. Cong, G.S. Zhuang, J.L. Zhao, C.L. Sun, H.W. Song, W. Wang, Microchip-based small, dense low-density lipoproteins assay for coronary heart disease risk assessment, Electrophoresis 29 (2008) 1932– 1941.

Microchip for atherogenic lipoprotein profile analysis / H. Wang et al. / Anal. Biochem. 467 (2014) 75–83 [17] H. Wang, D.X. Wang, J.C. Wang, H.M. Wang, J. Gu, C.X. Han, Q.H. Jin, B.J. Xu, H. Chao, L. Cao, Y. Wang, J.L. Zhao, Application of poly(dimethylsiloxane)/glass microchip for fast electrophoretic separation of serum small, dense lowdensity lipoprotein, J. Chromatogr. A 1216 (2009) 6343–6347. [18] H. Wang, C.X. Han, H.M. Wang, Q.H. Jin, D.X. Wang, L. Cao, G.Z. Wang, Simultaneous determination of high-density lipoprotein, very low-density lipoprotein, and low-density lipoprotein subclass in human serum by microchip, Chromatographia 74 (2011) 799–805. [19] J.F. Chen, Q.H. Jin, J.L. Zhao, Y.S. Xu, A signal process method for DNA segments separation in micro-channel electrophoresis, Biosens. Bioelectron. 17 (2002) 619–623. [20] R.J. Havel, H.A. Eder, J.H. Bragdon, The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum, J. Clin. Invest. 34 (1955) 1345–1353. [21] M.J. Shiddiky, Y.B. Shim, Trace analysis of DNA: preconcentration, separation, and electrochemical detection in microchip electrophoresis using Au nanoparticles, Anal. Chem. 79 (2007) 3724–3733. [22] C.J. Yu, C.L. Su, W.L. Tseng, Separation of acidic and basic proteins by nanoparticle-filled capillary electrophoresis, Anal. Chem. 78 (2006) 8004– 8010.

83

[23] E.J. de Melo, W. de Souza, Pathway of C6-NBD-ceramide on the host cell infected with Toxoplasma gondii, Cell Struct. Funct. 21 (1996) 47–52. [24] A.A. Rahsepar, M.S. Alavi, M. Ghayour-Mobarhan, G. Ferns, Reply to ‘‘small dense low-density lipoprotein could be used as a therapeutic marker for treatment in patients with acute coronary syndrome’’, Angiology 64 (2013) 645. [25] E. Kanda, M. Ai, M. Okazaki, Y. Maeda, S. Sasaki, M. Yoshida, Clinical utility of gray scale renal ultrasound in acute kidney injury, BMC Nephrol. 14 (2013) 188, http://dx.doi.org/10.1186/1471-2369-14-188. [26] H. Deguchi, N.M. Pecheniuk, D.J. Elias, P.M. Averell, J.H. Griffin, High-density lipoprotein deficiency and dyslipoproteinemia associated with venous thrombosis in men, Circulation 112 (2005) 893–899. [27] D. Duffy, D.J. Rader, Update on strategies to increase HDL quantity and function, Nat. Rev. Cardiol. 6 (2009) 455–463. [28] H.B. Brewer Jr., Clinical review: the evolving role of HDL in the treatment of high-risk patients with cardiovascular disease, J. Clin. Endocrinol. Metab. 96 (2011) 1246–1257. [29] M.A. Austin, M.C. King, K.M. Vranizan, R.M. Krauss, Atherogenic lipoprotein phenotype: a proposed genetic marker for coronary heart disease risk, Circulation 82 (1990) 495–506.