The susceptibility of lipoproteins to oxidation is thought to be a critical step in a complex process culminating in arterial lipid deposition; however, the role of the oxidizability of lipoproteins in atherosclerosis has not been demonstrated in any large-scale clinical trials (1)(2)(3)(4). Although several measurements of the oxidizability of LDL have been developed, the most commonly used method is the measurement of the rate of formation of conjugated hydroperoxides of polyunsaturated fatty acids in response to exposure to copper ions, which leads to the formation of conjugated dienes (CDs) (5). Three major parameters are used to describe the oxidizability of LDL by the CD assay: (a) the lag time (LT) during which lipoprotein-associated antioxidants are consumed, thereby sparing the LDL lipids from oxidation; (b) the maximum rate of CD formation (Max V); and (c) the total amount of CD formed (ΔCD).
In practice, the CD assay requires the sequential measurement of absorbance at 234 nm in a spectrophotometer over a prolonged period of time, often >3 h. Because the automated cell carrier of most spectrophotometers holds a maximum of only six cuvettes, the analysis of large numbers of specimens, such as for large-scale clinical trials, is difficult and time-consuming. In addition, the analysis of replicate specimens or the inclusion of quality-control materials generally is precluded by the small number of specimens that can be analyzed simultaneously. We have developed methods that allow the isolation of LDL from a large number of specimens and the analysis of the oxidizability of this isolated LDL. Several factors were necessary for these goals to be attained: (a) the use of an ultracentrifuge rotor with capacity for a large number of specimens (Type 25; Beckman Instruments); (b) the recent availability of a 96-well microtiter plate reader that reads absorbance at 234 nm (PowerWave 200 reader; Bio-Tek Instruments); and (c) the development of a disposable 96-well microtiter plate that is transparent at 234 nm (Costar UV-Plate; Corning Costar).
Blood was collected into Vacutainer Tubes containing liquid EDTA (1.5 g/L blood final concentration) from healthy volunteers after a 12- to 14-h fast, and the plasma was removed immediately after separation of cells by centrifugation at 1700g at 4 °C. LDL was either prepared immediately from fresh plasma or 1-mL aliquots of plasma were stored in cryovials with no other added chemicals at −80 °C by a two-step ultracentrifugation procedure. Specimens were protected at all times from direct light, and all solutions were kept cold and contained 10 μmol/L EDTA. Frozen specimens were thawed at room temperature on a rotator, shielded from light by aluminum foil. To a 1.0-mL aliquot of plasma, 298 μL of a solution of KBr (ρ = 1.225 kg/L) was added. An aliquot (650 μL) of this plasma was transferred to a Beckman 8 × 51 mm polycarbonate ultracentrifugation tube and overlaid with 350 μL of a solution of KBr (ρ = 1.063 kg/L). The tubes were centrifuged in the outer ring of the Beckman Type 25 rotor at 85 925gavg for 18 h at 10 °C in a Beckman L8–55 ultracentrifuge. If >44 tubes were centrifuged simultaneously, the middle ring was used to accommodate the extra tubes and the time of centrifugation was extended to 20 h. After centrifugation, a Beckman CentriTube tube slicer was used to separate the top layer containing VLDL and LDL from the bottom layer containing HDL and other plasma proteins. The top layer (200 μL) was transferred to another 8 × 51 mm polycarbonate centrifuge tube, overlaid with 800 μL of NaCl (ρ = 1.006 kg/L), and centrifuged as above. After centrifugation, the LDL was concentrated in an orange-colored layer at the bottom of the tube. The tube was sliced with the CentriTube slicer such that the top 800 μL was removed and discarded. The concentrated LDL was transferred to a clean cryovial, brought to a final volume of 500 μL with 8.77 g/L NaCl, and applied to a prepacked PD-10 Sephadex G-25 M column (Pharmacia Biotech) to remove residual salts and water-soluble antioxidants such as bilirubin, uric acid, and ascorbic acid (6). The desalted LDL was eluted with 10 mmol/L phosphate-buffered saline (PBS), pH 7.4. To avoid potential errors caused by contamination by albumin and the measurement of protein by the Lowry or dye-binding methods, the apolipoprotein (apo) B concentration of the LDL preparations was measured directly by nephelometry (Behring Diagnostics) after column filtration. The purity of the LDL was assessed by lipoprotein electrophoresis. On the final day of LDL isolation, the total time from the slicing of the tubes to the application of the LDL to the microtiter plate for analysis was ∼5 h.
For a standard oxidation measurement, purified LDL was adjusted to 55 mg/L apo B with PBS containing 10 μmol/L EDTA. A stock solution of CuSO4 · 5 H2O (11 mmol/L) was diluted to 275 μmol/L with PBS without EDTA. Aliquots of LDL (250 μL, 12.5 μg) and CuSO4 (25 μL) were applied to a Costar UV-Plate; the final concentrations were 50 mg/L LDL apo B, 25 μmol/L Cu2+, and 9.1 μmol/L EDTA. The plate was inserted into the temperature-controlled PowerWave 200 reader and prewarmed to 37 °C, and readings were taken at 234 nm every 3 min for 3 h. The plate was shaken automatically for 15 s before each reading. For routine use, specimens were analyzed in duplicate, allowing the analysis of 46 specimens and 2 quality-control materials in each assay. The total time for analysis of 48 specimens from thawing to reporting of results was <2.5 days, with only ∼12 h of technician time.
The KC4 program (revision 12) provided with the PowerWave 200 reader calculated minimum and maximum absorbance values and the slope (ΔAbsorbance/min) and the time of the inflection point for the propagation phase for each kinetic curve. These values and the absorbance readings for each well were exported to a QuattroPro (Corel) spreadsheet that calculated the LT (min), Max V (μmol CD/min-g apo B), and ΔCD (μmol CD/g apo B). LT was defined as the time at the intercept of the propagation phase with the baseline (5). The amount of CD formed was based on the molar absorptivity of CD, ε234 = 29 500 L/mol-cm (5). Details of the spreadsheet calculations can be obtained upon request.
The inclusion of EDTA throughout the isolation procedure, the protection from light, and the constant refrigeration of the specimens minimized the possibility of premature oxidation of LDL during the various isolation steps. When these factors were not carefully controlled, LT decreased significantly, although Max V and ΔCD were less affected.
The use of gel filtration columns for the final purification of the isolated LDL removed the relatively high concentration of residual KBr from the ultracentrifugation steps and also removed low-molecular weight, water-soluble antioxidants, such as bilirubin, ascorbic acid, and uric acid, which have been shown to be potent antioxidants (5)(6)(7)(8). The LTs for isolated LDL that had not been processed by gel filtration were approximately twofold longer than the LTs for the same specimens after column filtration. These findings suggest that the nonfiltered specimens contained antioxidants that were removed by column filtration and were in agreement with the observations of and Brussaard et al. (9) and Scheek et al. (10) (by dialysis), and of McDowell et al. (7) (by gel filtration). The columns also provided a rapid (∼10 min) method for the purification of LDL that did not lead to a loss of protein or lipophilic antioxidants (α-tocopherol, lycopene, and β-carotene), which has been shown to occur as a result of dialysis (10).
After preliminary validation experiments, we chose a Cu2+ concentration of 25 μmol/L (∼16 μmol/L free Cu2+) for use in the standard oxidation assay. At this concentration, all oxidation parameters were stable and did not show the concentration-dependent variability observed at lower Cu2+ concentrations. Although the “proper” ratio of Cu2+ to LDL for meaningful interpretation of the effects of physiologic, pharmacologic, or dietary perturbations on oxidation has been discussed, no consensus has been reached (6)(11)(12). The inclusion of 10 μmol/L EDTA in all solutions including the oxidation reaction mixture preserved the LDL and prevented oxidation during the relatively long isolation process required for the treatment of large numbers of specimens.
The development of a disposable plastic 96-well microtiter plate with minimal absorbance at 234 nm has made possible the simultaneous analysis of a large number of specimens. To evaluate the well-to-well variability between measurements, all wells of the Costar UV-Plates were filled with either deionized water or a routinely processed LDL specimen, and the absorbance at 234 nm was read. The mean absorbance was 0.099 (CV = 4.8%; n = 96) for water and 0.406 (CV = 2.7%; n = 96) for the LDL specimen, which is similar to the absorbance reported for specimens in quartz cuvettes in a standard spectrophotometer (13). To evaluate the within-run reproducibility of the assay on the microtiter plate, a large amount of LDL was prepared and applied to each well of a microtiter plate, and the oxidation assay performed. Table 1⇓ describes the well-to-well reproducibility of the measurement of the three oxidation parameters. The interwell reproducibility was excellent, with a CV <3% for each of the oxidation parameters. To evaluate the reproducibility of the LDL preparation method as well as of the oxidation assay, 19 replicate preparations of LDL were made by the standard method from the plasma of one subject. The standard oxidation assay was performed on each LDL in quadruplicate on a single microtiter plate. The CV of the LT was 2.3%; however, the Max V and ΔCD were more variable, with CVs of 5.9% and 4.8%, respectively. Between-run imprecision was assessed using two frozen quality-control specimens that were analyzed in each assay with clinical trial specimens (Fig. 1⇓ ). The between-run CV for LT was 5.9% to 9.0% over 18 or 20 assays performed over 37 weeks. The CV for Max V and ΔCD was 4.0% to 5.8%. Values for within-run and between-run imprecision are difficult to find in the literature because of the limited number of specimens that can be analyzed in a single assay in most systems. Princen and co-workers (14)(15)(16) reported within-run and between-run CVs similar to ours, but using a different assay format that used a 10-position automated spectrophotometer. Kleinveld et al. (12) also reported CVs similar to ours.
The collection of blood in Vacutainer Tubes containing liquid EDTA, followed by rapid separation of plasma from cells by centrifugation and freezing at −80 °C in plastic cryovials, appears to be sufficient for the maintenance of the oxidizability of the specimen. Antioxidants such as α-tocopherol, β-carotene, and lycopene have been found to be stable in frozen plasma specimens prepared by this method (17). To be practical for use in clinical trials, specimens must be unaffected by freezing. To evaluate the effect of freezing on the oxidizability of LDL, we collected EDTA plasma from four healthy laboratory employees after a 12-h fast. Aliquots of plasma were either used immediately or frozen in cryovials at −80 °C. LDL was prepared by the standard method on the freshly collected plasma or after 1, 2, and 4 weeks of storage, and the oxidizability of the isolated LDL was determined. None of the oxidation parameters deviated from the fresh specimen by >5% over 4 weeks. Fig. 1⇓ shows the effects of freezing for up to 37 weeks on plasma. No trends were observed in any of the oxidation parameters for either pool, indicating a negligible degradation of the specimens over at least 37 weeks of storage at −80 °C. Van der Vijver et al. (18) have established the stability of EDTA plasma stored for 18 months at −80 °C without harmful effects on the oxidizability of LDL.
In summary, we have optimized the performance of the CD assay to determine the oxidizability of LDL for use in clinical trials. The Cu2+-catalyzed oxidation of LDL has been used previously by many investigators as an indication of the amount of antioxidants associated with LDL. In our method, plasma is collected easily and stored without extraordinary efforts, and purified LDL is prepared from large numbers of specimens. Up to 96 kinetic measurements of the formation of CDs are made simultaneously in a temperature-controlled environment. In this format, multiple replicate measurements can be made, increasing the precision of the assay and allowing the use of quality-control materials in each assay. Although the CD assay may never become a routine clinical assay for the assessment of the oxidizability of LDL, it is useful for research and can now be applied to specimens from large clinical trials and epidemiologic surveys.
Within- and between-run precision.
Analysis of quality-control materials stored at −80 °C.
Two quality-control materials were prepared, and aliquots were frozen at −80 °C. For each assay, a vial of each material was thawed, the LDL was prepared, and the oxidation assay was performed in duplicate. The horizontal lines in each panel indicate the mean value of all assays over 37 weeks. No consistent trend was obvious for any of the oxidation parameters, which would indicate degradation of the quality-control materials. •, QC1; ——–, mean of 20 assays; ○, QC2; – – – –, mean of 18 assays.
Acknowledgments
We thank Mark Witham of Bio-Tek Instruments, Inc. (Winooski, VT) and Ken Brown of Bio-Resources, Inc. (St. Louis, MO) for providing access to preproduction versions of the PowerWave 200 instrument.
Footnotes
Core Laboratory for Clinical Studies, Department of Medicine, Washington University School of Medicine, Box 8046, 660 S. Euclid Ave., St. Louis, MO 63110
- © 1999 The American Association for Clinical Chemistry