BIOCHEMICAL
MEDICINE
33, 291-296
(1985)
Quantitation of Malondialdehyde (MDA) in Plasma, by Ion-Pairing Reverse Phase High Performance Liquid Chromatography C. R. WADE, P. G. JACKSON, AND A. M.
VAN
RIJ
Department of Surgery. University qf Otago. Duneditl, New Zealand Received
September
29, 1983
Malondialdehyde (MDA) is a volatile @scission product formed from the free radical-mediated peroxidation of certain polyunsaturated fatty acids (PUFAs) that occur largely in cell membranes (1). Measurements of MDA levels in biological fluids have been used extensively as a measure of lipid peroxidation and an indirect measure of prostaglandin production (2, 3). Elevated levels of MDA have allegedly been found in numerous diseases with associated lipid peroxidative changes (2, 4-6). MDA has been most commonly measured by the 2-thiobarbituric acid (TBA) method. This method involves the complexing of MDA with two molecules of TBA in an acid solution to form a red chromogen that is measured either by spectrophotometry at 532 nm (7) or by fluorimetry (8). However, quantitation of MDA in complex biological fluids is limited by its lack of specificity, as pointed out by several workers (5, 9, 10). Furthermore characterization of the TBA reaction has shown that the red chromogen is primarily associated with nonvolatile compounds that decompose during the acid heating step to form MDA (I, 11). These compounds have been shown to be cyclic and noncyclic hydroperoxides and p-unsaturated aldehydes (12). Other nonlipid compounds have also been shown to react with TBA, giving absorption spectra that interfere with measurements at 532 nm (5, 9, 13). The fluorometric determination of MDA is also subject to interference by nonlipid compounds (9). A method is described using HPLC for the quantitative measurement of MDA in plasma which separates the effect of inte~e~ng TBA-reacting substances. MATERIALS
AND METHODS
Chemicals and reagents. Analytical grade chemicals were used throughout and were purchased from Sigma Chemical Company (St. Louis, MO.). Solvents of analytical grade were obtained from J. T. Baker Chemical Company (Phillipsburg, NJ.) except for methanol, which was of HPLC grade (Waters). Deionized distilled water was used throughout. ~rep~r~ti~~ of reagents and standards. ~iobarbitu~c acid reagent (TBA reagent) was prepared as a 0.67% (w/v) solution in distilled water/glacial acetic acid (501 291 0006-2944/85
$3 .OO
Copyright Q 1985: by Academic Press, Inc. AD rights of reproduction in any form reserved.
292
WADE.
JACKSON.
AND
VAh-
KIJ
50) (8). MDA standards were prepared by acid hydrolysis of I, 1.3,3-tetramethylpropane in TBA reagent. Reaction of TBA reagent with plasma. Platelet-poor plasma (PPP) was prepared by centrifugation from blood collected into heparinized tubes from healthy volunteers. The TBA reaction was carried out by adding 1 ml of the acidic TBA reagent to 1 ml of plasma and heating for 20 min in a boiling water bath. After cooling the reacted sample was either further prepared for HPLC or measured directly at 532 nm for “TBA-reactive” substance. DEAE-cellulose separation. The reacted sample was applied directly after cooling to a 0.5-ml column of DEAE-cellulose (equilibrated with 4.5 M acetic acid), in a Pasteur pipet. Then it was washed with 2 ml of distilled water, followed by 4 ml of 0.04 M NaOH to wash off the interfering chromogens. The MDA complex was then eluted with 5 M NaOH and immediately neutralized with 5 M HCl. The neutralized solution was subsequently extracted into a final volume ot 0.5 ml of butanol for HPLC injection. HPLC analysis. HPLC of the standards and plasma samples was performed on a Waters ALC/GPC 204 liquid chromatograph with a spectrophotometric detector at 532 nm using a 0.39 x 30-cm PBondapak C,, column of 3-pm particle size. A range of methanol/water concentrations was used and the pH range was adjusted with 0.1 M phosphoric acid. Cetyltrimethylammonium bromide (Cetrimide) was employed as the ion-pairing reagent at a range of concentrations between 0.5 and 5% (w/v), to determine the optimal separating concentration. The various peaks eluting from the HPLC were collected separately and scanned in the spectrophotometer from 400 to 650 nm and the spectra were compared with the authentic MDA standard. Recovery of the MDA-TBA complex from the DEAE-cellulose columns was observed by standard additions of the MDA to the plasma before the TBA reactions. Spectrophotometry. Absorbance measurements were made on a Pye-Unicam scanning spectrophotometer either in the scanning mode or at the 532-nm maxima observed for the TBA reaction with the MDA standard. Comparison of methods. Duplicate l-ml samples of plasma from volunteer healthy subjects were analyzed for comparison by both the direct spectrophotometric method and HPLC. RESULTS
HPLC Conditions for Standards and Plasma MDA
Initial experiments on alkylphenol and PBondapak Cl8 columns using various concentrations of methanol in water were unsuccessful in separating the MDATBA chromogen from other TBA-reactive substances. However, using Cetrimide as an ion-pairing agent at a 1% (w/v) concentration in a mobile phase consisting of methanol/water (45155 v/v), pH 6.5, a good separation of the MDA-TBA chromogen from the other interfering chromogens was achieved (Fig. 1). The spectral profile of the plasma MDA-TBA fraction collected from the HPLC was identical to that of the authentic standard (Fig. 2). There was a linear response for the MDA-TBA chromogen standards (Fig. 3).
HPLC QUANTITATION
OF PLASMA
MALONDIALDEHYDE
293
TM hL 5
10
15
time I mm1
FIG. 1. HPLC protiles: (A) TBA-reacted plasma applied directly to the HPLC column. (B) Reacted plasma after absorption onto DEAE-cellulose and washing with water. (C) Reacted plasma after absorption and washing with water and 0.05 M NaOH, followed by elution with 5 M NaOH. (D) Separation of MDA-TBA chromogen (M) formed from hydrolysis of 1. I ,3,3-tetramethylpropane from the peak due to TBA itself (T).
A
400
450
500
550
6
wavelength lnml
FIG. 2. Absorption spectra: (A) TBA-reacted plasma before ion-exchange chromatography and HPLC (-). MDA-TBA chromogen eluted from the DEAE-cellulose column and collected after HPLC (---). (B) Absorption spectra for MDA standard formed from hydrolysis of 1,1,3,3-tetramethylpropane in TBA reagent.
294
WADE, JACKSON,
AND VAN RIJ
]
MDAl~gm/ml~
FIG. 3. HPLC response curves for MDA standards (0) and TBA-reacted plasma samples after the addition of MDA standards (0). Each point represents the mean value of triplicate samples and SD 6 0.02 pg/ml.
DEAE-Cellulose
Chromatography
for Plasma
Chromogen
Sample Yreparution
Unfortunately direct application of the plasma following the TBA reaction on the HPLC column is unsatisfactory. Several chromogens are formed and identified at 532 nm, and an undesirable overlap occurred in some plasma samples (Fig. 1). Absorption onto DEAE-cellulose, followed by elution with water and 0.04 M NaOH, effectively removed the problem of interfering chromogens (Fig. 1). The MDA-TBA complex bound strongly to the DEAE-cellulose but was eluted with 5 M NaOH, which after neutralization and butanol extraction was analyzed by HPLC. Complete recovery of MDA-TBA after NaOH elution was confirmed by the method of standard additions (Fig. 3). The visible spectra of the TBA-reacted plasma is complex and has absorption maxima at wavelengths other than at 532 nm. The MDA-TBA separated by ion exchange and HPLC, however, had a profile identical to the MDA standard (Fig. 2). A comparison of plasma MDA, measured as TBA-reactive substance spectrophotometrically at 532 nm and as MDA-TBA following HPLC separation, is presented in Table 1. Complete recovery in the HPLC system was confirmed by rechromatographing the collected MDA-TBA fraction. In plasma MDA-TBA was less, by the new method, than TBA-reactive substance and showed no other consistent relationship between samples from different subjects. DISCUSSION
The TBA reaction has been employed increasingly during the last 10 years to detect peroxidative changes occurring from enzymatic and nonenzymatic free radical-mediated reactions (2, 3). In certain diseases increased levels of MDA in plasma, synovial fluids, and tissue exudates have been demonstrated by spectrophotometric and fluorometric techniques (2, 4-6). However, the specificity of the TBA reaction and its measurement by these techniques is questioned (5, 9, 10). More appropriately the products of the TBA reaction are referred to as
HPLC QUANTiTATiON
OF PLASMA
MALONRIALDEHY~E
295
TABLE 1 Comparison of Spectrophotomet~c and HPLC Values for MDA Levels in the Plasma(nmole/m~~ of Normal Subjects” HPLC 5.0 4.8 4.75 4.5 4.3 4.1
--
Spectrophotometric _~10.4 10.2 8.1 8.4 7.9 8.0
u Mean values for triplicate samples. SEM was less than 6% of the mean for every value.
“MDA-like” or “TBA-reactive” compounds. Gutteridge (9) has shown, using a thin-layer chromato~aphic separation technique followed by visible and fluorescent spectrophotometry, that five chromogens are present in TBA-reacted plasma. By the use of the ion-pairing HPLC method described in this paper, we have also been able to demonstrate that several chromogens contribute substant~auy to the total absorbance at 532 nm. However, under the conditions described, the chromogen of specific interest MDA-TBA can be separated and accurately quantified in plasma. MDA, in samples of 0.25-0.5 ml of plasma, is accurately measured using a visible-wavelen~h detector at 532 nm. Levels in smaller samples could probably be measured with the use of a fluorescence detector. The use of DEAE-cellulose columns serves to concentrate the TBA reactants, and separate off some of the interfering chromogens prior to HPLC injection, Gutteridge (14) has shown that using acetic acid in the acid-heating stage of the TBA reaction can result in the formation of a TBA-reactive adduct with absorbance at 532 nm. We have found that heating the TBA reagent by itself can result in a 532-nm absorbing HPLC peak (peak T in Fig. 1D). This peak can be separated from the authentic MDA-TBA chromogen (peak M in Fig. ID) using the HPLC technique described. This coelutes in the DEAE-cellulose wash with 5 M NaOH because of its polar nature, and interferes with any direct spectrophotometry at this stage. To overcome this interference it is necessary to carry out the further separation by the use of the ion-prong HPLC method described here. The nature of the other interfering TBA reactants is not clear but their ~ont~bution to the measured absorption may be substantial. TBA reactions have been observed with a variety of substances, including biliverdin, ribose, 2-aminopyrimidines, and sialic acid (9, 13). Unfortunately changes in these TBA reactants may vary independently of changes in MDA-TBA when measured in patients with different diseases. Widely divergent values for the MDA-TBA chromogen and TBAreactive substances may occur when the chromogen change is not due to the MDA-TBA chromogen but due to the other interfering TBA reactants. Our preliminary observations confirm this possibility. This finding casts some doubt
296
WADE, JACKSON,
AND VAN RfJ
on the significance of the numerous observations published of elevated TBAreactive substances in a variety of clinical disorders. Difficulty in comparison of these observations has also been contributed to by the many different. assay conditions used, including, for example, various pH. acids. and wavelength conditions, for absorption measurement (3, 5). Each of these assay conditions may have a corresponding group of interfering chromogens. The method described not only measures bound MDA in plasma but also measures that MDA known to be released from lipid peroxides during the TBA reaction (I, 11). This amplifies its usefulness in the measurement of lipid peroxidation. The availability of a method to measure the MDA-TBA chromogen alone may lead to some clarification of the clinical significance of plasma MDA and lipid peroxidation.
An ion-pairing high performance liquid chromatography method is described for the separation and quantitation of malondialdehyde in plasma. The MDA is determined as the thiobarbiturate chromogen formed by reaction of the plasma with 2-thiobarbituric acid under acid and heating conditions. However. under these conditions other interfering chromogens can also be formed. Using DEAEcetlutose chromatography followed by ion-pairing HPLC, we have been able to separate and quantitate the levels of MDA-TBA chromogen formed in plasma from other interfering chromogens. Measurements of MDA levels in the plasma of six normal individuals by HPLC gives a mean value of 4.57 2 0.33 nmolei ml, whereas the spectrophotometric determined value is 8.83 + I, 15 nmoleiml. These data suggest that some reevaluation of the numerous papers published on MDA levels in plasma using spectrophotometric methods may be necessary. ACKNOWLEDGMENT This work was supported by the New Zealand Medical Research Council.
REFERENCES 1. Dahle, L. K.. Hill, E. G.. and Homan. R. T., Au&. Biocheru. Bioph,v.s. Y8, 212 (I%?). 2. Yagi, K., in “Lipid Peroxides in Biology and Medicine” (K. Yagi. Ed.). p. 223. Academic Press. New York, 1982. 3. Smith, J. B., Ingerman. C. M.. and Silver. M. J., .1. Lrrh. <‘fizz. Med. 88, 167 (1976). 4. Muus, P., Bonta, I. L.. and der Oudsten. S. A., Prosr. Med. 2. 63 (19791. 5. Satou, K., Cfin. Chem. Actu 90, 37 11978). 6. Matkovics, B., Laszio, A., and Szabo, L., C/in. <:/rem. Acfrr 118, 289 (198’1). 7. Kohn. H. I., and Liversedge, M., f. Phurm~col. FJ.up.Titer. 82, 292 (1944). 8. Yagi, K.. Biochem. Med. 15, 212 (1976). 9. Gutteridge. J. M. C.. Anal. Biochem. 91, 250 (1978). IO. Gutteridge, J. M. C.. Inz. /. Biochem. 14, 649 (1982). 1 I. Sinnhuber, R. 0.. and Yu, T., J. Food. Res. 23, 626 (1962). 12. Pryor, W. A., Stanley, J. P., and Blair, E., Lipids 11, 370 ( 1976). 13. Waravdekor, V. S., and Saslow, L. D.. J. Biol. Chem. 234, 1942 (1959). 14. Gutteridge, J. M. C.. Anuf. Biochetn. 69, 518 (1975).