388

ASSAY AND REPAIR OF BIOLOGICAL DAMAGE

[40]

[40] G a s C h r o m a t o g r a p h y - M a s s S p e c t r o m e t r y Assays for Lipid Peroxides By FREDERIK J. G. M. VAN KUUK, DAVID W. THOMAS, ROBERT J. STEPHENS, and EOWARD A. DRATZ

Introduction The purpose of this chapter is to summarize recently developed gas chromatography-mass spectrometry (GC-MS) techniques for the measurement of lipid peroxidation in tissues. GC-MS methods for identification of peroxidation products have not been applied to phospholipids with one exception. Hughes et al.l,2 used a GC-MS method to measure phospholipid oxidation products in a study of acute carbon tetrachloride toxicity in mouse liver. Their methodology required an enzymatic incubation step to liberate fatty acids from phospholipids or triglycerides, formation of methyl esters using diazomethane, and several purification steps before analyses. The key to the methods described here is a mild transesterification procedure that allows simple and direct chemical identification and semiquantitative measurement of phospholipid or other glyceroester peroxide products by GC-MS. 3,4The method can be simply extended to full quantitation with the use of isotopically labeled internal standards. 5 The method uses sodium borohydride reduction of the hydroperoxides and trimethylsilyl (TMS) derivatives of hydroxy fatty acid esters. Phospholipid hydroperoxide standards were synthesized and characterized to test the method. 3 The approach presented is applicable to the study of peroxidation of triglyceride storage depots as well as membrane phospholipids.6 The transesterification procedure does not detect free fatty acid peroxides which are usually formed by enzymatic prostanoid metabolism. I H. Hughes, C. V. Smith, J. O. Tsokos-Kuhn, and J. R. Mitchel, Anal. Biochem. 130, 431 (1983). 2 H. Hughes, C. V. Smith, J. O. Tsokos-Kuhn, and J. R. Mitchel, Anal. Biochem. 152, 107 (1986). 3 F. J. G. M. van Kuijk, D. W. Thomas, R. J. Stephens, and E. A. Dratz, J. Free Radicals Biol. Med. 1, 215 (1985). 4 F. J. G. M. van Kuijk, D. W. Thomas, R. J. Stephens, and E. A. Dratz, J. Free Radicals Biol. Med. 1, 387 (1985). 5 F. J. G. M. van Kuijk, D. W. Thomas, R. J. Stephens and E. A. Dratz, this volume [41]. 6 F. J. G. M. van Kuijk, D. W. Thomas, R. J. Stephens, and E. A. Dratz, in "Lipid Peroxidation in Biological Systems" (A. Sevanian, ed.), p. 117. American Oil Chemists' Society, Champaign, Illinois, 1988.

METHODSIN ENZYMOLOGY,VOL. 186

Copyright© 1990by AcademicPress, lnc~ All rightsof reproductionin any formreserved.

[40]

G C - M S ASSAYS FOR LIPID PEROXIDES

389

Methods

Sources of Reagents Methanolic (m-trifluoromethylphenyl)trimethylammoniumhydroxide (0.2 N) was obtained from Applied Sciences Laboratories, Inc. (Bellefonte, PA). N, O-Bis(trimethylsilyl)trifluoroacetamide (BSTFA) plus 1% trimethylchlorosilane (TMCS) was obtained from Regis Chemical Co. (Morton Grove, IL). Pentafluorobenzyl alcohol and sodium tert-pentoxide were purchased from Aldrich Chemical Company (Milwaukee, WI). All solvents were HPLC grade from Fisher Scientific (Fairlawn, N J).

Extraction of Lipid Peroxides from Tissues Tissue samples or phospholipid hydroperoxides obtained by photooxidation 3 are extracted by the method of Bligh and Dyer, 7 which is modified to use dichloromethane instead of chloroform.3 The solvent ratios used during the extraction are as follows: 1 ml dichloromethane, 1 ml methanol containing 50 ~g/ml butylated hydroxytoluene (BHT), and 0.25 ml aqueous buffer (2 mM EDTA, pH 7.0) are added to 5-50 mg tissue. This single-phase solvent mixture is homogenized in the first step in the extraction. One part dichloromethane (1 ml) is added in the second step, followed by vortex mixing for 60 sec. In the third step 0.5 ml water is added followed by vortex mixing, which creates two phases for extraction. The samples are centrifuged for 2 min at 1000 g, the dichloromethane lower phase is collected, and extraction of the aqueous phase with dichloromethane is repeated. Fractions of dichloromethane containing phospholipid and/or triglyceride hydroperoxides are pooled, dried over sodium sulfate, and evaporated under nitrogen. The aqueous component of the single-phase organic solvent used for lipid peroxide extraction contains EDTA in order to inhibit iron-promoted breakdown of lipid peroxides. BHT is added to the organic component to inhibit formation of additional lipid peroxides during the workup. Desferal (desferroxamine mesylate, desferrioxamine) has been recommended as a substitute for EDTA as a powerful and perhaps even more effective chelating agent. 8 The importance of substitution of dichloromethane for chloroform in the extraction procedure should be emphasized. Chloroform is used in the organic phase for total lipid extraction by most workers, following the widely used methods of Folch et al. 9 o r Bligh and Dyer. 7 However, we 7 E. C. Bligh and W. J. Dyer, Can. J. Biochem. Physiol. 37, 911 (1959). s j. M. C. Gutteridge, R. Richmond, and B. Halliwell, Biochem. J. 184, 469 (1979). 9 j. Folch, M. Lees, and G. H. Sioane Stanley, J. Biol. Chem. 226, 497 (1957).

390

ASSAY AND REPAIR OF BIOLOGICAL DAMAGE

[40]

found that phospholipid hydroperoxides in chloroform tended to undergo decomposition, which may be mediated by formation of the trichloromethyl r a d i c a l , l° We found that dichloromethane greatly stabilized lipid peroxides, and, therefore, the extraction method of Bligh and Dyer7 was modified by substituting dichloromethane for chloroform. The modifications require a substantial change in solvent ratios as described above. Dichloromethane is less toxic to laboratory workers than chloroform and more convenient to remove by evaporation. Colorimetric analysis established that phospholipid hydroperoxides are stable for months in HPLC grade dichloromethane at - 2 0 °.

Derivatization Reactions for GC-MS Analysis of Lipid Peroxides Simplified GC-MS methods were recently published by our laboratory, based on a one-step transesterification reaction at room temperature, to convert fatty acid esters to either their corresponding methyl esters or pentafluorobenzyl (PFB) esters. 3,4 A scheme of the derivatization procedures for analysis of oxidized lipids by GC-MS using PFB esters is shown in Fig. 1.

Transesterification to Form Methyl Esters The lipids of interest are collected in dichloromethane, and the dichloromethane is evaporated with a stream of dry nitrogen. The hydroperoxides are chemically reduced with 10 mg sodium borohydride in 1 ml methanol at 4° for 1 hr. After incubation, 1 ml water is added to decompose the sodium borohydride, and the samples are reextracted into dichloromethane. The extract is dried over sodium sulfate for 1 hr, and the liquid phase is transferred to a clean vial and evaporated under nitrogen to dryness. Subsequently, 50 ~1 of dichloromethane and 20 ~1 of transesterification reagent (0.2 M m-trifluoromethylphenyltrimethylammonium hydroxide in methanol) are added, and samples are shaken on a vortex mixer and incubated at room temperature for 30 min. This procedure provides quantitative conversion of phospholipids and triglycerides to fatty acid methyl esters as shown by TLC. 4,11 After incubation, 150/xl methanol and 200/.d water are added, and the samples are extracted into 1 ml hexane. The methanol-water phase is washed 2 times with 1 ml hexane, and the pooled hexane fractions are evaporated to dryness. Dry l0 R. O. Recknagel, E. A. Clende, and A. M. Hruszkewycz, In " F r e e Radicals in Biology" (W. A. Pryor, ed.), Vol. 3, p. 97. Academic Press, New York, 1977. u F. J. G. M. van Kuijk, D. W. Thomas, J. P. Konopelski, and E. A. Dratz, J. LipidRes. 27, 452 (1986).

[40]

G C - M S ASSAYS FOR LIPID PEROXIDES /

cH,

391

--

o o H

o

NaBH4

CH~

/

--

C-O-( ' ~ o II

o H

V

HO- C H z - ~ F F F F

OH3~

F

C-O-CHz%F H

F F BSTFA ,I-

1% TMCS F v

.I

o

.j

v

F

'?--r

..

o

TMS

F

F

FIG. 1. Derivatization procedure for analysis of oxidized phospholipids by GC-MS. The top structure shows the fatty acid hydroperoxide esterified in a lipid. First, a reduction step with sodium borohydride is employed to form a hydroxy derivative (still esterified to the lipid). Second, transesteritication is performed in the presence of pentafluorobenzyl alcohol to yield PFB esters. Methyl esters are produced if methanol is used instead of pentafluorobcnzyl alcohol. The third step shows the conversion of the hydroxyl groups to TMS ethers. pyridine (25 ~1) and B S T F A containing 1% T M C S (25/zl) are added to c o n v e r t the alcohols to the corresponding T M S derivatives.

Transesterification to Form Pentafluorobenzyl Esters Phospholipid a n d / o r triglyceride hydroperoxides are extracted and reduced to alcohols as described a b o v e and shown in Fig. 1. Transesterification to f o r m P F B esters is c a r d e d out with a reagent containing 20% (v/v) pentafluorobenzyl alcohol and 3% (w/v) sodium tert-pentoxide in dic h l o r o m e t h a n e TM (Fig. I). A 20-/.,1 aliquot of this solution is added to a

392

ASSAY AND REPAIR OF BIOLOGICAL DAMAGE

[40]

sample of phospholipids or triglycerides in 50/zl dichloromethane. The vial is closed with a Teflon-lined screw cap, shaken on a vortex mixer, and incubated for 30 min at room temperature (phospholipids) or at 60° (triglycerides). After incubation dichloromethane is first evaporated, and 1 ml hexane, 200/.d water, and 200/~1 methanol are added in this order. Further derivatization is carried out as described for methyl esters. It is recommended that the completeness of the transesterification reaction be monitored by TLC on representative samples. 4,11 Sodium tertpentoxide is hygroscopic and may be inactivated to the corresponding alcohol by traces of water, which leads to incomplete transesterification. This problem can be avoided by preparing the transesterification reagent immediately before use. Furthermore, traces of methanol react rapidly to form fatty acid methyl esters instead of PFB esters and therefore should be carefully excluded. Methyl esters are not formed if hexane and water are added first after transesterification.

Gas Chromatography-Mass Spectrometry G C - M S analysis is carried out using a Ribermag R10-10 C G C - M S system. Chromatography is carried out on a 5-m DB-5 capillary column (J + W Scientific, Rancho Cordova, CA) with a temperature program from 150 to 250° at 10°/min for fatty acid methyl esters or from 180 to 280° at the same rate for PFB esters. A Ros glass falling needle injector is used at 270° . Mass spectra are obtained by either electron ionization (El) at 70 eV for methyl esters or by negative ion chemical ionization (NICI) at 60 eV using ammonia reagent gas for PFB esters. Results and Discussion Methyl ester derivatives with EI detection proved to be useful for obtaining structural information on lipid peroxides. EI allowed detection of the different positional isomers from photooxidized unsaturated fatty acid chains. 3 It was found that the nonconjugated isomers, which are specific for singlet oxygen-mediated photooxidation reactions, could bc easily separated and distinguished from the conjugated isomers. 3,4 The conjugated isomers are formed both by autoxidation and by photooxidation. The pentafluorobenzyl (PFB) esters only yield one ion (M - H) in the mass spectrometer if detected by NICI. 4,12The lack of fragmentation and the higher thermal stability of the PFB esters allow detection at the picogram level. 12 R. J. Strife and R. C. Murphy, Prostaglandins, Leukotrienea Med. 13, I (1984).

[40]

GC-MS

ASSAYS FOR LIPID PEROXIDES

393

TABLE I NICI FRAGMENTS OF PENTAFLUOROBENZYL ESTERS FROM NONOXIDIZED FATTY ACIDS AND SINGLE- AND

DOUBLE-PHOTOOXIDIZEDFATTYACIDS, AFTER REDUCTION OF HYDROPEROXIDES AND FORMATION OF

O-TMS DERIVATIVES Oxidized fatty acid (M - C7FsH2)a Fatty acid

M - CTFsH2a

Single

16 : 1 16 : 0 17 : 0 18 : 3 18 : 2 18:1 18 : 0 20 : 5 20 : 4 20:0 22 : 6 22 : 5 22 : 4

253 255 269 277 279 281 283 301 303 311 327 329 332

341

Double

(365)b 367 369 (389)b 391 415 417 419

479 503

a On chemical ionization in the mass spectrometer, the C7FsH2 group is cleaved from the PFB esters and the carboxylate anion is generated. b The oxidized fatty acids and mass values of carboxylic ions shown in parentheses were not detected during this study. T a b l e I s h o w s t h e m/e v a l u e s f o r t h e major n e g a t i v e i o n s o f e a c h o f t h e P F B e s t e r s f r o m t h e m o s t c o m m o n n a t u r a l l y o c c u r r i n g f a t t y a c i d s in b i o l o g i c a l s y s t e m s , a n d v a l u e s f o r t h e p r o d u c t s o f single o x i d a t i o n s a n d double oxidations of these unsaturated fatty acids. The double-oxidized s p e c i e s a r e o b s e r v e d in s m a l l a m o u n t s in t h e s e e x p e r i m e n t s w i t h t h e c h a r a c t e r i s t i c m a s s e s s h o w n in T a b l e I a n d a r e d i s c u s s e d b e l o w . A s c a n b e s e e n in T a b l e I, all t h e s e s p e c i e s c a n b e d i s t i n g u i s h e d b y t h e u n i q u e m a s s e s o f t h e i r n e g a t i v e c a r b o x y l a t e i o n f o r m e d f r o m t h e P F B e s t e r s in the mass spectrometer.

Photooxidized Rat Retinal Lipids A NICI total ion chromatogram of total rat retina lipids, photooxidized

in vitro a n d a n a l y z e d a s P F B e s t e r s , is s h o w n in Fig. 2a. A b o u t 2 ng o f a

394

ASSAY AND REPAIR OF BIOLOGICAL DAMAGE

16:0

I

I

I

L

18:0

Jl

18:l~

II

18:2R

I

1

I

t

I

[40]

o.Total 22 6

I

I

I

I i

Oxidized 22:6 M'CrFsH2

I

I

I

I

I

i i

I

I

I

I

b.415

,r l i

I

I

I

I

I

I

I

I

1

,~-~,,

I

-~l

I

I

u

I

I

,I

Oxidized 20:4 M-C7 F5 H 2

I

I

I

c.591

J

I I I I I I I I I I

I

I

I

I

I

I

I

1

50

I

I

I

90

I

I

I

I

d, 369

Oxidized 18:1 M-C7 F5 H,~

Ii

I

I

I

150

I

J I

I

I

t

i

I

I

I

t

I

I

I

I

f

I

170 0 250 290 530 M/S Scan Number FIG. 2. Negative ion chemical ionization GC-MS of photooxidized rat retinal lipids, after derivatization according to the PFB ester method. The mass spectrometer performed 1 scan per 1.5 sec. (a) Total ion chromatogram. (b) Specific ion monitoring for the carboxylate anion of 22 : 6-O-TMS pentafluorobenzyl esters minus CTFsH2 at m/e 415. (c) Specific ion monitoring for the carboxylate anion of 20:4-O-TMS pentafluorobenzyl esters minus CTFsH2 at m/e 391. (d) Specific ion monitoring for the carboxylate anions of 18 : I-O-TMS pentaliuorobenzyl esters minus C7FsH2 at m/e 369.

[40]

G C - M S ASSAYS FOR LIPID PEROXIDES

395

mixture of pentafluorobenzyl ester derivatives was injected on the G C - M S system. Figure 2b-d shows chromatograms that were obtained by specific ion monitoring for the carboxylate anions of the products of single oxidations of docosahexaenoic acid, arachidonic acid, and oleic acid (molecular weights are listed in Table I). The oxidation was limited to about 10-15% of the total double bonds by limiting the amount of oxygen available. 3 Under these conditions it was possible to detect only small amounts of a double-oxidized arachidonic acid di-TMS derivative (m/e 479) near scan number 225 in Fig. 2 and double-oxidized docosahexaenoic acid di-TMS derivative (m/e 503) near scan number 275 (data not shown). These double-oxidized derivatives were about one-tenth as abundant as the single-oxidized products. The nonconjugated and conjugated derivatives from oxidized arachidonic acid are separated (Fig. 2c), but separation was not obtained between conjugated and nonconjugated isomers from oxidized docosahexaenoic acid (Fig. 2b). The unresolved hump at longer retention times than the sharp, large oxidized docosahexaenoic acid peak monitored at mass 415 contained predominantly conjugated species with some double-oxidized species on the trailing edge. We purposely used short capillary columns with modest chromatographic resolution to maximize the speed of the assays and the recovery of the products. 3,4,6 No common impurities such as the plasticizer dioctyl phthalate were observed in the NICI total ion chromatograms. 4 Owing to the enhanced NICI sensitivity obtained on analysis of PFB esters, the samples can often be diluted to such an extent that impurities are not detected. The traces in Figures 2b-d show that, with specific ion monitoring, small amounts of oxidation products can easily be detected in the presence of larger amounts of nonoxidized fatty acids.

Evaluation of Pentafluorobenzyl Ester Method NICIMass Spectrometry. In the NICI system used in this study, more than 90% of the total ionization was found in the carboxylate anion. 4 Under optimal EI conditions ionization is spread over many more fragments, and there is typically at most 0.5% ionization in any structurally significant ion. Strife and Murphy ~2calculated an advantage factor of 20100 using NICI conditions with PFB esters relative to optimal EI conditions with PFB esters. When methyl esters are used, there is little difference in sensitivity between EI and NICI. Transesterification. Methods previously available for forming PFB esters required free fatty acids.lZ Therefore, a lipase or saponification step would be required to use these methods for analysis of phospholipids or

396

ASSAY AND REPAIR OF BIOLOGICAL DAMAGE

[40]

triglycerides. Both the lipase and the saponification conditions are relatively vigorous and can lead to formation and/or decomposition of peroxidized polyunsaturated phospholipids or glycerides depending on the detailed conditions. Furthermore, the manipulation required impairs the recovery of small amounts of oxidized products. Direct transesterification is much simpler and, with the use of a sterically hindered quaternary ammonium hydroxide catalyst in methanol to form methyl esters, appears to be significantly milder than the enzymatic approaches mentioned above. McCreary et al. 13showed that this transesterification method produced derivatization of triglycerides to fatty acid methyl esters at room temperature. Their method was applied successfully for derivatization of phospholipid peroxides that were first reduced to the corresponding hydroxy derivatives) The one-step transesterification reaction to form PFB esters, rather than using enzymatic release and derivatization of free fatty acids in two steps, was desirable. Analysis of several authentic synthetic phospholipids such as (16 : 0)(18 : 2)-phosphatidylcholine and (16 : 0)(22 : 6)-phosphatidylcholine by the PFB ester method gave the expected ratio of fatty acids. These results indicate that quantitative conversion of the fatty acid esters in phospholipids to PFB esters was obtained with saturated as well as with highly polyunsaturated fatty acids. Analysis of the total fatty acid composition of the whole rat retina by flame ionization detection using PFB ester derivatives produced a fatty acid distribution identical to that reported by Farnsworth et al., ~4 who used conventional methanol-BF3-catalyzed transesterification to form methyl esters. Derivatization on photooxidized model phospholipids, after reduction of the hydroperoxides to the corresponding hydroxy derivatives, showed that PFB esters were also efficiently produced from hydroxy fatty acids. This indicates that these functional groups are preserved during the transesterification step with the base catalyst employed. Ultrasensitioe D e t e c t i o n . The PFB esters are clearly superior to the methyl esters for analysis of small amounts of lipid peroxidation products by GC-MS. The combined advantages of thermal stability in the gas chromatograph and the enhanced NICI sensitivity in the mass spectrometer provide an at least I000 times lower detection level for the PFB ester derivatives. As little as 1-10 pg (2.5-25 fmol) of oxidation product can be detected. The greater stability of the PFB ester derivatives of the phospholipid oxidation products on the GC column also allowed detection of oxidation products with two (or perhaps more) oxidation sites. When 13 D. K. McCreary, W. C. Kossa, S. Ramachandran, and R. R. Kurtz, J. Chrornatogr. Sci. 16, 329 (1978). 14 C. C. Farnsworth, W. L. Stone, and E. A. Dratz, Biochim. Biophys. Acta 552, 281 (1979).

[40]

G C - M S ASSAYS FOR LIPID PEROXIDES

397

TABLE II COMPARISON OF METHYL ESTER AND PFB ESTER METHODSa Parameter Substrate specificity Detection limit MS detection Main ion Application GC detection

Methyl ester

PFB ester

Fatty acid hydroperoxide esters Hydroxy fatty acid esters 10 lag E1 Fragments Adipose FID

1-10 pg NICI M - H All tissues ECD

a El, Electron ionization; NICI, negative ion chemical ionization; FID, flame ionization detection; and ECD, electron capture detection.

oxidation products were analyzed as methyl esters, 3only the single oxidation products were detected, probably because of lower stability of the products. The differences between the methyl ester and PFB ester methods are summarized in Table II. Quantitation. Analysis as PFB esters by GC-MS under NICI conditions provides a relatively simple approach to detect lipid peroxidation products with as yet unsurpassed sensitivity and selectivity. However, this method yields qualitative information only, since recoveries may vary considerably between different tissues and different analyses. Hence, it would be highly desirable if absolute levels could be determined by including suitable internal standards. Hughes et al.2 reported a method for quantitation of lipid peroxidation products in total hepatic lipid based on the use of methyl 15-hydroxyarachidonate as internal standard. Three isomers, 11-, 12-, and 15-hydroxy fatty acids of arachidonate, were quantitated as the methyl ester O-TMS ether derivatives. A disadvantage of this approach for quantitation results from differences in retention time for the internal standard and the compounds of interest. The conditions in the mass spectrometer may change rapidly in time, especially with the very sensitive NICI method; therefore, it is much more desirable to use stable isotope internal standards. We attempted to use a rare saturated fatty acid (20 : 0) as an internal standard during the present study and could not obtain reliable results with NICI methods. Quantitative GC-MS analyses are typically carded out, using compounds labeled with deuterium (D), carbon-13, or oxygen-18. 2,12,15,16Since 15 W. K. Rohwedder, Prog. Lipid Res. 24, 1 (1985). 16 H. Frank, M. Wiegand, M. Strecker, and T. Dietmar, Lipids 22, 689 (1987).

3~8

ASSAY AND REPAIR OF BIOLOGICAL DAMAGE

[40]

these types of internal standards elute very close to (D) or coelute (~3C and ~sO) with the authentic compounds, errors arising from changes in mass spectrometer conditions are minimized or eliminated. Also, peak identification is more facile in the presence of the higher backgrounds usually encountered when tissue extracts are analyzed. An example of this approach is presented elsewhere in this volume, where a quantitative assay of cleavage products of lipid peroxides is developed. In this assay deuterium-labeled internal standards are used which nearly coelute with the compounds measured. Quantification of the methods developed here for analysis of hydroxy fatty acids awaits future application of stable isotope internal standards. An example of such an assay was recently reported by Frank et al.,~6 who employed lSO-labeled standards. Conclusions Transcstcrification to form P F B esters and NICI G C - M S provide several advantages compared to analyses with methyl esters. Lipid peroxides arc detected not only in vitamin E deficiency, but also in normal nutritional states. The P F B NICI method can detect and identify lipid peroxides in animals raised on lab chow diets. The El G C - M S methods allow identification of the contribution of singlet oxygen-mediated reactions during damage to tissues. Future work requires quantitation with stable isotope internal standards 5 in order to determine the differences in lipid peroxide contents between groups of animals raised on different diets, animals exposed to different conditions that produce pathology, and in investigation of human pathology. The methods developed can also bc used in a simple G C system without a mass spectrometer. In case of P F B derivatives, very sensitive detection (I pg) can bc obtained with electron capturc detection (ECD). A major difference in the G C assay is the requirement for only one internal standard which must bc chromatographically separated from commonly occurring oxidized fatty acids. A commcrciaUy available synthetic (16 :0)(16 : l)-phospholipid can serve as an internal standard for G C after photooxidation, since the palmitoleic acid content of most biological tissues is relatively low.