Homogeneous assay for measuring low-density ... - CiteSeerX

Clinical Chemistry 44:3 522–531 (1998)

Lipids and Lipoproteins

Homogeneous assay for measuring low-density lipoprotein cholesterol in serum with triblock copolymer and a-cyclodextrin sulfate Hiroyuki Sugiuchi,1* Tetsumi Irie,2 Yoshinori Uji,3 Tomohiro Ueno,1 Toshiko Chaen,1 Kaneto Uekama,2 and Hiroaki Okabe3

A positive correlation between the concentration of total cholesterol in serum and the incidence of coronary heart disease is the most consistent finding to emerge from epidemiological and clinical studies [1]. Because LDLcholesterol (LDL-C) accounts for approximately twothirds of the serum cholesterol and constitutes the primary atherogenic fraction of the circulating cholesterol, monitoring of LDL-C in serum provides the basis for the diagnosis and treatment of hyperlipemia [2].4 A wide variety of methods has been used for determining LDL-C in serum, including a precipitation-based method [3], electrophoresis [4], HPLC [5], sequential and density-gradient ultracentrifugation [6], Friedewald formula [7], and immunoseparation [8]. Of these, the Friedewald formula is the most commonly used approach in clinical laboratories; however, it is limited to specimens from fasting subjects and specimens with triglyceride concentrations ,4.5 mmol/L [9]. Although beta-quantification by ultracentrifugation is the currently accepted reference method for measuring LDL-C in serum [10], it requires relatively large volumes of serum, long turnaround time, and is unsuitable for routine laboratory testing. Therefore, there is a great clinical need to develop a convenient and reliable method for measuring LDL-C in serum without resorting to any fractionation procedures. Interaction of surfactants with lipoproteins has been the focus of many studies [11, 12]. In general, for the determination of total cholesterol in serum, excess amounts of surfactants have been used for solubilizing all the lipoprotein cholesterol in a nonspecific manner and

We have developed a fully automated method for measuring LDL-cholesterol (LDL-C) in human serum without the need for prior separation, using a nonionic surfactant, polyoxyethylene–polyoxypropylene block copolyether (POE-POP), and a sodium salt of sulfated cyclic maltohexaose, a-cyclodextrin sulfate. Of the surfactants tested, POE-POP with a higher molecular mass of the POP block and a greater hydrophobicity reduced the reactivity of cholesterol in lipoprotein fractions; the reactivity in descending order was LDL >> VLDL > chylomicron ' HDL. Gel filtration chromatographic studies revealed that POE-POP removed lipids selectively from the LDL fraction and allowed them to participate in the cholesterol esterase– cholesterol oxidase coupling reaction system. By contrast, a-cyclodextrin sulfate reduced the reactivity of cholesterol, especially in chylomicrons and VLDL. A combination of POE-POP with a-cyclodextrin sulfate provided the required selectivity for the determination of LDL-C in serum in the presence of magnesium ions and a small amount of dextran sulfate without precipitating lipoprotein aggregates. There was a good correlation between the results of LDL-C assayed by the proposed method and the beta-quantification reference method involving 161 sera with triglyceride concentrations ranging from 0.3 to 22.6 mmol/L.

1 Department of Central Laboratory, Kumamoto University Hospital, 1-1-1, Honjo, Kumamoto 860, Japan. 2 Faculty of Pharmaceutical Sciences, Kumamoto University, 5–1, Oehonmachi, Kumamoto 862, Japan. 3 Department of Laboratory Medicine, Kumamoto University, School of Medicine, 1-1-1, Honjo, Kumamoto 860, Japan. *Author for correspondence. Fax (81) 96-362-7540; e-mail sugiuchi@ gpo.kumamoto-u.ac.jp. Received August 28, 1997; revision accepted November 13, 1997.

4 Nonstandard abbreviations: LDL-C, HDL-C, LDL-, HDL-cholesterol; POE-POP, polyoxyethylene–polyoxypropylene block copolyether; CHER, cholesterol esterase; CHOD, cholesterol oxidase; 4-AA, 4-aminoantipyrine; EMSE, N-ethyl-N-(3-methylphenyl)-N9-succinyl ethylenediamine; MOPS, 3-(N-morpholino)propanesulfonic acid; Lp(a), lipoprotein(a); apo, apolipoprotein; and NCEP, National Cholesterol Education Program.

522

Clinical Chemistry 44, No. 3, 1998

allowing it to freely participate in the enzymatic reaction system. By contrast, an approach for the direct measurement of LDL-C in serum has been proposed on the basis of the sequential use of two surfactants that show differential selectivities towards lipoprotein fractions [13]. Recently, we established a direct method for measuring HDL-C in serum, with the combined use of polyethylene glycol-modified enzymes and a-cyclodextrin sulfate [14, 15]. In these studies a-cyclodextrin sulfate, which has a highly concentrated negative charge and possesses heparin-mimicking activity, was used to reduce the reactivity of cholesterol in chylomicron and VLDL fractions in the presence of magnesium ions, without the need for precipitation of these lipoprotein fractions. Furthermore, our preliminary studies have shown that a polyoxyethylene–polyoxypropylene block polyether (POE-POP) reduced the reactivity of cholesterol, especially in HDL, suggesting that a combination of POE-POP with a-cyclodextrin sulfate could provide the needed selectivity for the determination of LDL-C in serum. Building upon these studies, this paper deals with the potential application of POE-POP and a-cyclodextrin sulfate in developing a novel direct assay for LDL-C in serum without prior fractionation.

Materials and Methods materials Cholesterol esterase (EC 3.1.1.13; CHER, 129 kDa from Pseudomonas species) and cholesterol oxidase (EC 1.1.3.6; CHOD, 35 kDa from Pseudomonas species) were supplied by Kyowa Hakko Kogyo Co. 4-Aminoantipyrine (4-AA) and N-ethyl-N-(3-methylphenyl)-N9-succinylethylenediamine (EMSE) were supplied by Kyowa Medex Co. All the surfactants used were commercially available. a-Cyclodextrin was donated by Nihon Shokuhin Kako Co. The sodium salt of a-cyclodextrin sulfate was prepared according to the nonregioselective method described previously [16]. The average degree of substitution of sulfate groups in a-cyclodextrin sulfate was confirmed to be between 12.0 6 0.5 (mean 6 SD of five different lots) by fast-atom bombardment mass spectrometry and elemental analysis. Dextran sulfate (from dextran, with an average molecular mass of 500 kDa) and 3-(N-morpholino)propanesulfonic acid (MOPS) were obtained from Sigma Chemical Co. and Dojin Chemical Co. respectively. Free and conjugated bilirubin dissolved in 30 g/L bovine serum albumin solution (ditaurobilirubin) were obtained from International Reagent Co. Immunoseparation reagent for the determination of LDL-C and latex agglutination reagent for the determination of lipoprotein(a) [Lp(a)] were obtained from Sigma Diagnostics and Daiichi Pure Chemicals Co. respectively. Human sera were obtained from healthy subjects 1 h after lunch, and the four major lipoprotein fractions [chylomicron, hydrated density (d) ,0.950 kg/L; VLDL, d 0.950 –1.006 kg/L; LDL, d 1.006 –1.063 kg/L; and HDL, d 1.063–1.135 kg/L], isolated by ultracentrifugation accord-

523

ing to the method of Hatch and Lees [6], were supplied by Health Care Technology Foundation. All other materials and solvents used were of analytical reagent grade.

gel filtration analysis The separation of lipoprotein fractions was performed by HPLC (CCPA 8000, Tosoh) with a gel filtration column (TSK gel Lipopropak XL, Tosoh) according to the method of Kitamura et al. [17]. The individual lipoprotein fractions or serum samples were eluted with 2.6 mmol/L NaCl, pH 7.0, at a flow rate of 1.0 mL/min. The column effluent was collected at intervals of 30 s with a fraction collector (FRAC100, Pharmacia Biotech.). The concentrations of cholesterol and phospholipids in the fraction separated by HPLC were determined enzymatically (Determiner TC and PL, Kyowa Medex). The concentration of total proteins in the fractions separated by HPLC was determined by the micro BCA protein assay reagent kit (Pierce) and those of apolipoproteins A-I, A-II, B, C-II, C-III, and E by immunoturbidimetry (Daiichi Pure Chemicals Co.). There was no apparent contamination with any other proteins such as serum albumin.

analytical procedure In the final formulation of reagents used for the determination, reagent 1 contained a-cyclodextrin sulfate (0.5 mmol/L), dextran sulfate (0.5 g/L), MgCl2 (2.0 mmol/L), and EMSE (1.4 mmol/L) in MOPS buffer (50 mmol/L, pH 6.75); reagent 2 consisted of CHER (1 kU/L), CHOD (3 kU/L), peroxidase (30 kU/L), 4-AA (2.5 mmol/L), and POE-POP (4 g/L) in MOPS buffer (50 mmol/L, pH 6.75). All the specimens were stored at 4 °C and analyzed within 2 days after blood sampling. Four microliters of the serum sample was added to 300 mL of reagent 1. The mixture was incubated at 37 °C for 5 min before adding 100 mL of reagent 2 and incubated for an additional 5 min. The chromophore formed in a coupled reaction with peroxidase was measured spectrophotometrically at 600 nm; alternatively, dual wavelength measurements [600 nm (main) and 700 nm (subsidiary)] were used, especially when assaying turbid or hemolyzed samples, to avoid any interference with the colorimetric assay. The concentration of LDL-C was calculated by using a serum-based calibrator (LDL-C, 2.77 mmol/L, Precinorm L; Boehringer Mannheim). The instruments used for the determination were a Hitachi 911 automated analyzer and a UV 160A spectrophotometer (Shimadzu). The results of LDL-C in serum assayed by beta-quantification method were obtained by Health Care Technology Foundation, performed according to the Lipid Research Clinic’s protocol [18]. The immunoseparation reagent for the determination of LDL-C was used to remove HDL and VLDL from serum, leaving LDL in the filtrate, and was quantified with a standard cholesterol assay [8]. The concentration of triglycerides in serum was determined enzymatically (Determiner TG, Kyowa Medex) and the concentration of HDL-C in serum was

524

Sugiuchi et al.: Assay for LDL-cholesterol

determined by a dextran sulfate–phosphotungstate– MgCl2 precipitation method (Daiichi Pure Chemicals Co.) respectively.

data analyses The contribution of various structural and physicochemical factors of surfactants to the selectivity towards LDL-C was statistically examined with multiple regression analysis [19]. Correlations and regression equations were calculated on a PC-9801UF personal computer (NEC), where a stepwise method was used and the validity of regression was judged by the F-test value (.2.0). The data were analyzed statistically by one-way analysis of variance, with Duncan’s multiple comparison test, and P values ,0.01 were viewed as statistically significant.

Results effect of surfactants on reactivity of lipoprotein cholesterol Compositions of lipids and proteins in the four major lipoprotein fractions used for this study are summarized in Table 1. The isolated lipoprotein fractions had the expected compositions, which are in good agreement with those of other compositional studies [20]. A total of 381 surfactants (24 cationic, 77 anionic, 249 nonionic, 6 amphoteric, and 25 other types) were surveyed with regard to the selectivity towards LDL-C. Fig. 1 shows the selec-

Table 1. Lipid composition of isolated lipoprotein fractions (mg/g protein). Lipid

HDL

LDL

VLDL

Chylomicron

Total cholesterol Free cholesterol Triglycerides Phospholipids

211 31 61 369

1052 248 224 757

358 274 1774 832

112 41 935 165

tivity indices of surfactants (2 g/L) for LDL-C (SILDL) in the presence of CHER (1 kU/L) and CHOD (3 kU/L). The SILDL value is defined here as the product of the relative reactivity of cholesterol in the LDL fraction (the first term) and the LDL selectivity (the second term) according to the equation: SI LDL 5 (RLDL/100) 3 [(RLDL 3 3)/ ~RHDL 1 RVLDL 1 RChylomicron)]

(1)

where R is the relative reactivity of cholesterol with the enzymes in each lipoprotein fraction, which is expressed as a percentage of the cholesterol reactivity in the presence of Triton X-100 (2 g/L), and the subscripts refer to each lipoprotein fraction, respectively. The LDL selectivity is defined as the ratio of the RLDL value to the sum of those for the other three lipoprotein fractions, where the RLDL value is multiplied by 3 to maintain the balance between the numerator (one term) and the denominator (three terms). Of the surfactants tested, some nonionic polymers showed greater selectivity towards LDL-C, giving a mean SILDL value of 2.45 (Fig. 1). In particular, a series of POE-POP, as indicated by open circles in Fig. 1, were most superior with respect to LDL selectivity. POE-POP is an A-B-A-type triblock copolyether consisting of hydrophilic polyoxyethylene A blocks attached to a hydrophobic central polyoxypropylene B block. In this study we used 19 POE-POP with an average molecular mass ranging from 1100 to 6500 Da. The relative contribution of structural factors of the POE-POP to the selectivity towards LDL-C was examined by using multiple regression analysis. In an initial multiple regression equation, the variable criterion was the SILDL value, and the explanatory variables were (a) the average molecular mass of POE-POP, (b) the molecular mass of the POP block in the POE-POP molecule (MMPOP), and (c) the

Fig. 1. Selectivity indices of various surfactants (2 g/L) for LDL-C (SILDL) in the presence of CHER (1 kU/L) and CHOD (3 kU/L) in MOPS buffer (pH 7.0) at 37 °C: E, SILDL value for POE-POP; vertical bar represents the mean of the SILDL values in each row.


525

hydrophobicity index for POE-POP as indicated by the fractional molecular mass of the POP block in the POEPOP molecule (FMPOP), respectively. When a stepwise method was used and the validity of regression was judged by the F-test value (.2.0), the first dependent variable was excluded from the equation and consequently the following equation was obtained: SI LDL 5 1.22 3 10 2 3( 6 4.66 3 10 2 4)MMPOP 1 6.43 3 10 2 2 ~ 6 2.54 3 10 2 2 ! FM POP 2 5.10

(2)

~n 5 19, r 5 0.671, F 5 6.56) where the values in parentheses are the estimated standard error of the coefficient. Three-dimensional plots displaying the relation between the SILDL value and the two predominant factors (MMPOP and FMPOP) for POEPOP used are shown in Fig. 2. The partial correlation coefficients for the MMPOP and FMPOP were 0.548 and 0.535, respectively. Because the magnitudes of the two coefficients are not significantly different, the two determinants may contribute equally to the selectivity of POEPOP towards LDL-C. On the basis of the above results, POE-POP with a MMPOP value of 3850 Da and a FMPOP value of 90%, giving a SILDL value of 9.24, was selected as the most preferred reagent for the determination of LDL-C and used in the following studies. Figure 3 shows the relative reactivity of cholesterol in lipoprotein fractions as a function of POE-POP concentrations in the presence of CHER (1 kU/L) and CHOD (3

Fig. 2. Three-dimensional plot displaying the relation between the SILDL value, MMPOP, and FMPOP for POE-POP in the presence of CHER (1 kU/L) and CHOD (3 kU/L) in MOPS buffer (pH 7.0) at 37 °C.

Fig. 3. Effects of POE-POP on relative reactivities of cholesterol in various lipoprotein fractions in the presence of CHER (1 kU/L) and CHOD (3 kU/L) in MOPS buffer (pH 7.0) at 37 °C: E, HDL; F, LDL; ‚, VLDL; Œ, chylomicron.

kU/L). The relative reactivity of cholesterol in the lipoprotein fractions decreased with increasing concentrations of POE-POP, with the reactivity decreasing in the order: LDL .. VLDL . chylomicron ' HDL.

effects of poe-pop on gel filtration of lipoprotein fractions Insight into the mechanism by which POE-POP exhibits its selectivity towards LDL-C was gained by gel filtration chromatography. Fig. 4 shows the effect of POE-POP (4 g/L) on the elution pattern for the LDL fraction, as obtained by detecting cholesterol, phospholipids, and proteins in the effluent. In the presence of POE-POP, the first peak at ;30 min contained predominantly apolipoprotein (apo) B, as determined by immunoturbidimetry, and cholesterol and phospholipids eluted later in a separate peak (Fig. 4B). The lipids isolated from the LDL particles would be solubilized in the mixed micelles with POE-POP, making them readily amenable to enzymatic reactions with CHER and CHOD. By contrast, when POE-POP at the same concentration as in Fig. 4 was added to the HDL fraction, it reduced the elution volume of the HDL fraction, with only a small change in the protein/lipid composition (Fig. 5). In fact, the protein fraction shown in Fig. 5B contained apo A-I and apo A-II at a molar ratio of 5:1, which is consistent with the expected protein concentrations for the intact HDL fraction. This suggests that the POE-POP varies the density or surface charge of the HDL particles and (or) induces the aggregation of HDL to macroparticles. If such is the case, the modified HDL particles should be resistant to the enzymatic reactions, eventually leading to reduced reactivity of cholesterol in that lipoprotein fraction. In the case of chylomicron and VLDL, the addition of POE-POP

526


Fig. 4. Elution patterns of the LDL fraction in the absence and presence of POE-POP (4 g/L), as monitored by cholesterol (E), phospholipids (F), and proteins (‚): A, without POE-POP; B, with POE-POP.

to those lipoprotein fractions caused no remarkable separation of proteins and lipids (data not shown). The above results clearly indicate that the POE-POP-mediated selectivity towards LDL-C can be ascribable to selective solubilization of cholesterol in the LDL fraction, thus allowing it to participate in the enzymatic reactions.

combination effects of poe-pop with a-cyclodextrin sulfate The reactivity of cholesterol in the VLDL fraction persisted with the sole use of POE-POP at concentrations up to 4 g/L, as shown in Fig. 3. For accomplishing the direct determination of LDL-C, the reactivities of cholesterol in all the lipoprotein fractions except for LDL must be completely abolished. For this purpose, we used a-cyclodextrin sulfate in combination with POE-POP. Our previous studies have shown that a-cyclodextrin sulfate, having a highly concentrated negative charge and possessing heparin-like activity, reduces the reactivity of cholesterol, especially in chylomicrons and VLDL in the presence of magnesium ions, without precipitating those lipoprotein

Fig. 5. Elution patterns of the HDL fraction in the absence and presence of POE-POP (4 g/L), as monitored by cholesterol (E), phospholipids (F), and proteins (‚): A, without POE-POP; B, with POE-POP.

aggregates. Recent studies have demonstrated that cyclodextrin sulfates exhibit threshold behavior in their heparin-mimicking activities; the introduction of ;2 sulfate groups per glucose unit is adequate [21]. Therefore, in this study we prepared a-cyclodextrin sulfate with a degree of sulfation of ;12, which corresponds to an average substitution of ;2 hydroxyl groups per glucose unit. Fig. 6 shows the relative reactivity of cholesterol in lipoprotein fractions as a function of magnesium ions in the presence of the POE-POP (4 g/L), a-cyclodextrin sulfate (0.5 mmol/L), dextran sulfate (0.5 g/L), CHER (1 kU/L), and CHOD (3 kU/L). The relative reactivity of cholesterol in the lipoprotein fractions decreased with increasing concentrations of POE-POP, with reactivity decreasing in the order: LDL .. VLDL . chylomicron ' HDL. Finally, the residual reactivities with fractions other than LDL were completely abolished by adding magnesium ions at concentrations .2 mmol/L.


527

Fig. 6. Effects of MgCl2 on relative reactivities of cholesterol in various lipoprotein fractions in the presence of POE-POP (4 g/L), a-cyclodextrin sulfate (0.5 mmol/L), dextran sulfate (0.5 g/L), CHER (1 kU/L), and CHOD (3 kU/L) in MOPS buffer (pH 7.0) at 37 °C: E, HDL; F, LDL; ‚, VLDL; Œ, chylomicron.

time course of enzymatic reaction Figure 7 shows changes with time in absorbance at 550 nm during the reaction of LDL-C, as determined at intervals of 20 s for 10 min, with the four lipoprotein fractions and the serum samples (diluted five times serially with 155 mmol/L NaCl solution) from a patient with hyperlipemia (total cholesterol 11.12 mmol/L, triglycerides 2.82 mmol/L, HDL-C 0.78 mmol/L). Only LDL-C from both the lipoprotein fractions and the serum sample exhibited reactivity, reaching a plateau within ;5 min after the addition of reagent 2.

calibration curve and linearity When calibration for the LDL-C assay was performed with control serum Precinorm L (LDL-C, 2.77 mmol/L, Boehringer Mannheim) for lipid determination, a straight line passing through the origin was obtained, and the calibration was stable during the 1-month period, with absorbance readings varying within 1%. The LDL fraction (LDL-C, 15.5 mmol/L) isolated by ultracentrifugation was diluted with 155 mmol/L NaCl solution, and linearity of the method was investigated. As shown in Fig. 8, good linearity was obtained with the serially diluted fractions of LDL-C solutions up to 15.5 mmol/L. The minimum detectable concentration was 0.005 mmol/L. When serum samples from patients with hyperlipemia (triglycerides 2.3–18.1 mmol/L, total cholesterol 5.2–15.5 mmol/L) were serially diluted with 155 mmol/L NaCl solution and used for assessment of the linearity, good linearity was obtained, suggesting that the influence of cholesterol derived from VLDL and chylomicrons in sera of patients with hyperlipemia was negligible or nonexistent.

Fig. 7. Changes with time in absorbance at 550 nm during the reaction of cholesterol in various lipoprotein fractions (A) and in the serum samples (diluted five times serially with 155 mmol/L NaCl solution) from a patient with hyperlipemia (total cholesterol 11.1 mmol/L, triglycerides 2.8 mmol/L, HDL-C 0.8 mmol/L) (B): (A) E, HDL; F, LDL; ‚, VLDL; Œ, chylomicron; M, 155 mmol/L NaCl solution; (B) E, 1/5; F, 2/5; ‚, 3/5; Œ, 4/5; M, 5/5.

imprecision and recovery tests Repeatability and reproducibility of the method with three different serum samples are shown in Table 2. For run-to-run precision studies the specimens were stored at 280 °C and the proposed method indeed worked with the frozen sera. In all three cases, the values obtained were within 4% of the target values as recommended by the National Cholesterol Education Program (NCEP) [22]. For the recovery test, 100 mL of the LDL fraction (LDL-C, 5.82 mmol/L) isolated by ultracentrifugation was added to 400 mL of the pooled serum (LDL-C, 1.34, 2.72, and 3.54 mmol/L) and the assay was performed with the reagent. The mean 6 SD recovery was 101.1% 6 3.8% with nine specimens, and ranged between 97% and 105%.

interfering substances Addition of conjugated bilirubin at final concentrations up to 0.68 mmol/L (0.14, 0.27, 0.41, 0.55, and 0.68

528


Clinic’s Program) with serum samples from healthy volunteers (n 5 86), patients with hyperlipemia (triglycerides ,4.5 mmol/L, n 5 41), and from patients with hyperlipemia (triglycerides $4.5 mmol/L, n 5 34). For serum samples from these three populations, there was a high degree of correlation between the results of the proposed method and those of the beta-quantification method, with r 5 0.989 or better. As shown in Fig. 9, the mean percentage bias of standard deviation for LDL-C determined by the proposed method in comparison with the beta-quantification method was 0.73% 6 5.78% (0.03 6 0.16 mmol/L) with serum samples from healthy volunteers, showing 1.01% 6 5.27% (0.05 6 0.18 mmol/L), and those from patients with hyperlipemia, showing no significant difference (P ,0.5). There was no increase in the percentage bias, even when the concentrations of triglycerides and HDL-C were high. Correlations between the proposed method and other methods (Friedewald formula, immunoseparation, and HPLC method) are sumFig. 8. Linearity of direct method for assaying LDL-C by using LDL fraction and serum samples (diluted five times serially with 155 mmol/L NaCl solution) from patients with hyperlipemia (triglycerides 2.3–18.1 mmol/L, total cholesterol 5.2–15.5 mmol/L): E, LDL fraction (triglycerides 1.4 mmol/L); F, triglycerides 4.4 mmol/L and total cholesterol 14.2 mmol/L; ‚, triglycerides 9.6 mmol/L and total cholesterol 7.2 mmol/L; Œ, triglycerides 17.6 mmol/L and total cholesterol 6.7 mmol/L; M, triglycerides 13.1 mmol/L and total cholesterol 4.7 mmol/L.

mmol/L) to the pooled serum produced a slightly negative error of up to 10% in the assay result. No adverse effect was observed when free bilirubin (up to 0.68 mmol/L), ascorbic acid (up to 2.84 mmol/L), hemoglobin (up to 5 g/L), Intralipos (up to 10 g/L, The Green Cross Co.), EDTA-2Na (up to 0.27 mmol/L), or citrate (up to 0.34 mmol/L) was included during the measurement of LDL-C with this assay. However, when citrate at a final concentration of 12.9 mmol/L was used as an anticoagulant for whole blood, it produced a slightly negative error of 7.8% in the assay result.

comparison with other methods The proposed method was compared with ultracentrifugation (beta-quantification method of the Lipid Research Table 2. Precision of the proposed method for measuring LDL-C in serum. Serum

Mean, mmol/L

SD, mmol/L

CV, %

I II III

0.91 2.97 10.50

0.013 0.023 0.026

1.43 0.78 0.25

I II III

0.91 3.00 10.58

0.021 0.034 0.093

2.29 1.12 0.88

Within-run precisiona

Run-to-run precisionb

a

Twenty aliquots of each serum were assayed simultaneously. The three sera were assayed in duplicate for 10 days with a single reagent lot. b

Fig. 9. Biases between direct method and beta-quantification method for assaying LDL-C with serum samples from healthy volunteers (n 5 86), patients with hyperlipemia (triglycerides ,4.5 mmol/L, n 5 41), and from patients with hyperlipemia (triglycerides $4.5 mmol/L, n 5 34), as associated with LDL-C (beta-quantification method), triglycerides, and HDL-C concentrations: F, sera from healthy volunteers and patients with hyperlipemia (triglycerides ,4.5 mmol/L); E, sera from patients with hyperlipemia (triglycerides $4.5 mmol/L).

529


marized in Table 3. In the serum samples from healthy volunteers, the results obtained by the proposed method were well correlated with other methods (r 5 0.98 – 0.99), but in those from patients with hyperlipemia (triglycerides $4.5 mmol/L), the correlation coefficient increased in the order: Friedewald formula ,, immunoseparation , HPLC , beta-quantification method. The LDL fraction used in this study is heterogeneous, and contains remnant particles of IDL (d 1.006 –1.019 kg/L) and Lp(a) (d 1.050 –1.080 kg/L). The proposed method for measuring LDL-C would actually include the contribution of IDL and Lp(a), since there was a good correlation between the results of LDL-C assayed by the proposed method and the beta-quantification method. This estimate is supported by the fact that 77.6% 6 3.3% of total cholesterol (0.39 6 0.13 mmol/L) in the IDL fraction isolated by ultracentrifugation responded to the proposed method. On the other hand, no direct evidence is available to clarify whether the proposed method would include particles of Lp(a), because of the difficulty in isolating the Lp(a) from the LDL and HDL fractions by ultracentrifugation. Nevertheless, there was no increase in the percentage bias for LDL-C determined by the proposed method in comparison with the beta-quantification method as the concentration of Lp(a) increased. This indicates that the proposed method would also include Lp(a).

Discussion Many of the current techniques used for the determination of LDL-C in serum are cumbersome, time consuming, and require specialized instrumentation, which limits their use in the clinical laboratory. The NCEP Working Group on Lipoprotein Measurement has been developing recommendations for LDL-C measurement capable of

quantifying LDL-C directly; it should not be based on calculation of difference between two or more measured values [22]. We established a direct method for determining the concentration of LDL-C in serum in a convenient format with the combined use of POE-POP and a-cyclodextrin sulfate. Only a very small sample volume (4 mL) is needed for this method, without the need for isolation of LDL. LDL-C can be determined in a short time (10 min) by this homogeneous method, which can be easily automated. The strategy used for direct measurement of LDL-C in serum is based on the cooperative actions of POE-POP as a quencher for HDL-C and a-cyclodextrin sulfate as a quencher for chylomicrons and VLDL-C. This is distinctly different from the approach used for the homogeneous assay for the direct measurement of LDL-C in serum described previously [13], in which one surfactant selectively hydrolyzes chylomicrons, VLDL, and HDL in a non-color-forming reaction, and then another surfactant hydrolyzes only LDL for color development. Previously, poloxamers including POE-POP have been successfully utilized as surface modifiers for improving the stability of latex particles [23] and as vehicles for transdermal drug delivery [24]. In this study, POE-POP showed limited specificity towards LDL-C, solubilizing it into mixed micelles and thus allowing it to participate in the enzymatic reaction. Some surfactants can be used to isolate apo B and lipids from the LDL fraction [25, 26], but show no selectivity towards LDL-C under the present conditions. Although the mechanism that confers such LDL selectivity to POE-POP is not clear, it is possible that the surfactant may be able to recognize differences in hydrated density, net charge, or size of the various lipoprotein fractions. As shown in Fig. 3, the higher the molecular mass of the POP block and hydrophobicity, the

Table 3. Correlations between the proposed LDL assay and other methods with sera from healthy volunteers (I), patients with hyperlipemia (triglycerides 4.5 mmol/L) (III).a Method

Proposed method

Friedewald formula

Immunoseparation

HPLC

Beta quantification

a b

Group

I II III I II III I II III I II III I II III

Slope (a)

0.973 0.933 0.395 0.936 0.981 0.988 0.995 1.023 0.932 1.001 1.008 1.010

In the form y 5 ax 1 b; y 5 the proposed method; x 5 other method. P ,0.01 vs group I, c P ,0.01 vs the proposed method.

y-intercept (b)

0.137 0.468 2.338 0.140 0.047 20.044 0.003 20.202 0.059 0.010 20.065 20.052

r

n

Mean 6 SD, mmol/L

0.987 0.990 0.552 0.972 0.980 0.974 0.987 0.989 0.980 0.989 0.992 0.992

86 41 34 86 41 34 86 41 34 86 41 34 86 41 34

2.752 6 0.954 5.131 6 2.333b 2.979 6 1.378 2.715 6 0.975 4.994 6 2.483b 2.020 6 1.771b,c 2.718 6 0.918 5.076 6 2.312b 2.907 6 1.384 2.741 6 0.962 5.048 6 2.397b 2.834 6 1.309 2.764 6 0.965 5.164 6 2.361b 2.964 6 1.381

530


greater the selectivity towards LDL-C. Apo A-I and apo A-II in the HDL particle are known to be water soluble, whereas apo B in the LDL fraction has an extremely large molecular mass of 540 kDa and is highly hydrophobic in nature [27]. Apo B and its aggregates isolated from the LDL particle by POE-POP may be solubilized in the complex form with the surfactant and have an apparent molecular mass of 2000 –3000 kDa, as judged by the elution volume in the gel filtration. In such a complex, the peripheral hydrophilic POE blocks should point freely in solution, whereas the central hydrophobic POP block anchors the surfactant to the hydrophobic surface of the apolipoprotein. On the other hand, the inability of POEPOP to remove apo A-I and apo A-II from the HDL particle may be explained in part by the hydrophilicity of those apolipoproteins and the physical size of the POEPOP micelles, which may not allow them to penetrate the interior of the HDL particle and thus may not solubilize the lipids. The heparin-like activity of a-cyclodextrin sulfate was utilized to reduce the reactivity of cholesterol in chylomicrons and VLDL in a manner similar to that reported previously [14, 15]. a-Cyclodextrin sulfate has an average molecular mass of 2194 Da, which is ;100 times smaller than those of the polyanions used in precipitation-based methods. Furthermore, the charge density of a-cyclodextrin sulfate is higher than those of the polyanions because of spatial constraints imposed on the sulfates by the hexasaccharide ring structure. In the presence of magnesium ions, a-cyclodextrin sulfate may form water-soluble and submicron-sized complexes with those lipoproteins with a low protein/lipid ratio, which are resistant to the enzymatic reactions. In fact, there was no noticeable visible precipitation during the measurement of LDL-C with this assay. By contrast, when dextran sulfate was included instead of a-cyclodextrin sulfate, it increased the turbidity of the reaction mixture to such an extent that it interfered with the determination of LDL-C in serum. The proposed method for measuring LDL-C would include the contribution of IDL and Lp(a), as the Friedewald equation or beta-quantification method does. Because all of the particles of this wide-density LDL population are atherogenic, the proposed method might be a more sensitive indicator of risk for premature coronary artery diseases than a method that has the limited specificity for the narrow-density LDL population. In conclusion, the homogeneous assay described here is simple and reliable for measuring LDL-C in serum without prior fractionation and should prove to be quite useful in the routine clinical laboratory.

We express our thanks to Masao Umemoto of the Health Care Technology Foundation, Tokyo, Japan, for supplying the samples of the four lipoprotein classes used in this study.

References 1. Kannel WB, Castelli WP, Gordon T. Cholesterol in the prediction of atherosclerotic disease: new perspectives based on the Framingham Study. Ann Intern Med 1979;90:85–91. 2. Gordon T, Kannel WB, Castelli WP, Dawber TR. Lipoproteins, cardiovascular disease and death; the Framingham study. Arch Intern Med 1981;141:1128 –31. 3. Wieland H, Seidel D. A simple method for precipitation of low density lipoproteins. J Lipid Res 1983;24:904 –9. 4. Nauck M, Winkler K, Ma¨rz W, Wieland H. Quantitative determination of high-, low-, and very-low-density lipoproteins and lipoprotein(a) by agarose gel electrophoresis and enzymatic cholesterol staining. Clin Chem 1995;41:1761–7. 5. Carrol RM, Rudel LL. Lipoprotein separation and low density lipoprotein molecular weight determination using high performance gel-filtration chromatography. J Lipid Res 1983;24:200 –7. 6. Hatch FT, Lees RS. Practical methods for plasma lipoprotein analysis. Adv Lipid Res 1968;6:1– 68. 7. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18:499 – 502. 8. Pisani T, Gebski CP, Leary ET, Warnick GR, Ollington JF. Accurate direct determination of low-density lipoprotein cholesterol using an immunoseparation reagent and enzymatic cholesterol assay. Arch Pathol Lab Med 1995;119:1127–35. 9. Warnick GR, Knopp RH, Fitzpatrick V, Bronson L. Estimating low-density lipoprotein cholesterol by the Friedewald equation is adequate for classifying patients on the basis of nationally recommended cutpoints. Clin Chem 1990;36:15–9. 10. Belcher JD, McNamara JR, Grinstead GF, Rifai N, Warnick GR, Bachorik P, Frantz I. Measurement of low density lipoprotein cholesterol concentration. In: Rifai N, Warnick GR, eds. Methods for clinical laboratory measurement of lipid and lipoprotein risk factors. Washington, DC: AACC, 1991:75– 86. 11. Ikai A. Extraction of apo B cluster from human low density lipoprotein with Tween 80. J Biochem 1980;88:1349 –57. 12. Hino K, Nakayama M, Nakanishi K, Manabe M. A new method for the homogeneous assay of serum HDL-cholesterol [Abstract]. Clin Chem 1996;42:299. 13. Halloran P, Pisani T, Long S. A homogeneous assay for the direct measurement of LDL-cholesterol in serum [Abstract]. Clin Chem 1987;43:254. 14. Sugiuchi H, Uji Y, Okabe H, Irie T, Uekama K. Differential effects of modified cyclodextrins on serum lipoprotein. Application to direct measurement of high-density lipoprotein in serum. In: Osa T, ed. Proceedings of the 7th International Symposium on Cyclodextrins. Tokyo: Komiyama Printing Co., 1994:532–5. 15. Sugiuchi H, Uji Y, Okabe H, Irie T, Uekama K, Kayahara N, Miyauchi K. Direct measurement of high-density lipoprotein in serum with polyethylene glycol-modified enzymes and sulfated a-cycrodextrin. Clin Chem 1995;41:717–23. 16. Folkman J, Weisz PB, Joullie MM, Li WW, Ewing WR. Control of angiogenesis with synthetic heparin substitutes. Science 1989; 243:1490 –3. 17. Kitamura T, Ito S, Moriyama H, Kato Y, Sasamoto K, Okazaki M. Quantitative analysis of serum lipoproteins (CM, VLDL, LDL and HDL) by high-performance gel filtration chromatography. Chromatography 1996;17:33–7. 18. Manual of laboratory operation: Lipid Research Clinics Program. Vol. 1, DHEW publication no. (NIH) 75– 628. Washington, DC: Department of Health, Education, and Welfare, 1974. 19. Tanaka U, Tarumi T, Wakimoto K. Handbook of statistical analysis by personal computer II: multivariate analysis. Tokyo: Kyoritsu, 1984:16 –37.


20. Havel RI, Kane JP. Introduction: structure and metabolism of plasma lipoproteins. In: Scriver CR, Beaudet AL, Sly WS, et al., eds. The metabolic and molecular bases of inherited disease, 7th ed . New York: McGraw–Hill, 1995:1841–51. 21. Burke CJ, Volkin DB, Mach H, Middaugh CR. Effect of polyanions on the unofolding of acidic fibrobrast growth factor. Biochemistry 1993;32:6419 –26. 22. Bachorik PS, Ross JW. National Cholesterol Education Program recommendations for measurement of low density lipoprotein cholesterol: executive summary. Clin Chem 1995;41:1414 –20. 23. Juha´sz J, Lenaerts V, Bellemare M, Pimienta C, Ong H. Stability and adsorption to polymeric surfaces of rat ANF in poloxamer 407 solutions. Int J Pharm 1991;77:309 –13.

531

24. Cappel MJ, Kreuter J. Effect of nonionic surfactants on transdermal drug delivery: II. Poloxamer and poloxamine surfactants. Int J Pharm 1991;69:155– 67. 25. Helenius A, Simons K. Removal of lipids from human plasma low-density lipoprotein by detergents. Biochemistry 1971;10: 2542–7. 26. Patterson BW, Kilgore LL, Chun PW, Fisher WR. Structural studies on apolipoprotein B: controllable heterogeneity of the complex formed with the surfactant, Triton X-100. J Lipid Res 1984;25: 763–9. 27. Chan L. Apolipoprotein B, the major protein component of triglyceride-rich and low density lipoproteins. J Biol Chem 1992;267: 25621– 4.