Ro!Na@mniins Lcukotrieneo and Essential Fatty Acids (1990) 39,131-134 fj?J Loqman Group UK Ud 1990

Extraction with Methyl TertrButyl Ether Overcomes Erratic Elution Patterns of bketo-Prostaglandin F1, on High Pressure Liquid Chromatography

W. A. Noort, F. A. de Zwart and M. J. N. C. Keirse

Department of Obketrics and Gynecology, Leiden University Hospital, PO Box 9600, 2300 RC Leiden, the Netherlands (Reprint requests to MJNCK)

ABSTRACT, Solvent extraction of 6-keto-PGF1, from aqueous solutions with ethyl acetate was found to result in variable and it-reproducible elutiou patterns, when the extracts were subjected to high pressure liquid chromatography. These problems could not be resolved satisfactorily by using ethyl acetate from dii- ferent suppliers, nor by changing acids or pH for acidification. After a number of unsuccessful attempts to resolve this problem, we found that variable and irreproducible elution patterns could be avoided by using methyl t-butyl ether as extraction solvent.

INTRODUCTION

Radioimmunoassay (RIA), particularly when com- bined with high pressure liquid chromatography (HPLC), has become the method of choice for the measurement of minute quantities of prostanoids present in biological fluids (1). The application of HPLC not only enhances specificity of the assay and removal of interfering substances, but also permits measurement of several prostanoids from one and the same sample. Obviously, this requires reliable extraction from biological material before HPLC separation. Ethyl acetate is commonly used be- cause it efficiently extracts even the most polar prostanoids.

However, several investigators (2-6) have reported that 6-keto-PGFr, decomposes during ex- traction with ethyl acetate to form several compounds with different chromatographic prop- erties on thin layer chromatography. Flower and Sun (3) reported that all of these compounds, found in a variable number (3-5) of different spots on thin layer chromatography, produce mass-spectra that are identical to that of 6-keto-PGFr, itself when sub- jected to gas chromatography-mass spectrometry (GC-MS).

During our attempts to measure &keto-PGFi, levels in pregnancy we observed that multiple peak formation during chromatography was not confined to thin layer chromatography but was also observed on HPLC of ethyl acetate extracts, irrespective of whether these were obtained from amniotic fluid, urine or even water. This led to both irreproducible elution patterns of 6-keto-PGFt, and to co-elution of 6-keto- PGFt, with other prostanoids.

These problems and the way in which they were

131

resolved are reported here, since that information may be of help to others in search of a reliable quantitation of 6-keto-PGFr, and other prostanoids in biological material.

EXPERIMENTAL PROCEDURES AND RESULi’S

Original extraction and assay procedure

Endogenous prostanoid levels were measured in several samples of amniotic fluid (n=35), urine (n=7), and other aqueous solutions (n=lO). Before extraction, samples were spiked with 3,000 cpm of [3H]-PGE2, [3H]-PGF2,, [3H]-PGD_,, 13H]-13,14- dihydro-15-keto-PGFzo, ([3H]-PGFM) and/or [3H]-TXB2 and [3H]-6-keto-PGF1, (New England Nuclear, Boston, Mass.) for recovery determina- tions. Samples were acidified with citric acid (pH 3) and extracted with 10 ml petroleum ether which was discarded. The aqueous phase was then extracted twice with 10 ml ethyl acetate (Merck, spectroscopic grade). These ethyl acetate phases were pooled and evaporated under nitrogen at 40 “C. The dry residue was dissolved in 250 Dl water, containing 0.75% acetic acid (v/v) and 10% acetonitrile (v/v) for high pressure liquid chromatography (HPLC). The HPLC system used has been described before (7). Of the 1 ml fractions eluted from HPLC, 100 Dl was used for peak detection by scintillation counting and the remainder was used for radioim- munoassay.

Influence of solvent extraction

Ethyl acetate. The HPLC elution Patterns of [3H]-6-

132 Prostaglandins Leukotrienes and Essential Fatty Acids

Fig. 1 HPLC elution patterns of [%f]-6-keto-PGF,, from ethyl acetate extracts of amniotic fluid (left) and urine (right).

keto-PGFla from ethyl acetate extracts derived from amniotic fluid, urine and water were all vari- able and irreproducible. The majority of the samples, extracted with ethyl acetate, eluted without a clear [3H]-6-keto-PGF1, peak on HPLC. Some samples showed most, other samples only a minority, of the [3H]-6-keto-PGF1, in the ap- propriate fractions (fractions 9-10; Fig. 1). The remainder was spread over several peaks throughout the gradient and the elution pattern was not reproducible even when the same extracts were run in duplicate.

The same applied to endogenous levels of 6-keto- PGFr, present in biological samples. On occasions, no 6-keto-PGF,, could be detected by radio- immunoassay of the appropriate HPLC fractions, while duplicates would show endogenous levels of 6-keto-PGF$, to be present.

This phenomenon was not observed for any ofthe other prostanoids, all of which eluted in the same way as pure standards. However, the variable amount of [3H]-6-ketoPGFr, that co-eluted with each of the other prostanoids resulted also in un- reliable and variable recovery determinations for these prostanoids (Fig. 2).

CPM

0 10 20 30 40 +B fractions

Fig. 2 HPLC elution pattern of [%I]-prostanoids from ethyl acetate extracts 4 and methyl t-butyl ether m derived from amniotic fluid.

Several approaches were used to overcome these problems. Ethyl acetate obtained from different suppliers (Baker, Merck, Ratburn) showed the same erratic patterns irrespective of the purity or source of the solvent. After extraction at pH 2-4 from 25% to 46% (n=6) of the [3H]-6-keto-PGFI, was recovered in fractions other than the [3H]6- keto-PGFI, fraction (fractions 9-10). As suggested by Flower and Sun (3), we substituted HCl for citric acid in the acidification step. Extracts obtained at pH 4 provided a satisfactory elution pattern on HPLC (92-95% of injected [3H]-6-keto-PGFr, in the appropriate fractions), but this pH was rather critical. For example, after extraction at pH 2 be- tween 29% and 63% of the [3H]-6-keto-PGQ, eluted again in other prostanoid fractions. Washing the ethyl acetate extract with water (twice) prior to evaporation, as advised by Salmon and Flower (2), reduced the problem in that 79 + 0.3% (mean f sd, n=3) of injected [3H]-6-keto-PGFI, eluted in the appropriate fractions, but erroneous peaks remained present.

Replacing solvent by solid-phase extraction (8) with ethyl acetate (citric acid, pH 2) resulted in an elution pattern with 80 + 1.3% (mean f sd, n=3) of the injected [3H]-6-keto-PGF, in the appropriate fractions.

Methyl t-butyl ether. Substituting methyl t-butyl ether (Fisons) for ethyl acetate for extraction consistently resulted in one clear 6-keto-PGFr, peak on HPLC (Fig. 3). All prostanoids eluted efficiently in frac- tions 9-10; m 85% when citric acid and m 94% when HCl was used for acidification.

A further advantage of methyl t-butyl ether is that

70

a60 il.. 0

4- B 50 1” Ii

j 40

%

30

0 5 to 1s 20 is 30 e FRACTIONS

Fig. 3 HPLC elution pattern of [%I]-6-keto-PGFla from methyl t-butyl ether extracts of amniotic fluid.

Methyl Tert-Butyl Ether Extraction of dketo-PGF,, 133

it is easier to evaporate, being considerably more volatile than ethyl acetate. Extraction should be performed below pH 4, however. At pH 4 or more 6-keto-PGE, is not extracted well from aqueous solutions with methyl t-butyl ether (acidified with citric acid or hydrochloric acid respectively 45% and 22% at pH 4, 88% and 71% at pH 3.5, and 96% and 96%at pH 3.0).

Solid-phase rather than solvent extraction with methyl t-butyl ether was found to provide similar results in terms of reliable elution patterns on sub- sequent HPLC.

DISCUSSION

Problems with the chromatography of 6-keto-PGF1, after solvent extraction with ethyl acetate have been reported by several authors (2-6). Mitchell and -his associates (5) experienced problems not only with ethyl acetate as extraction solvent, but also with cyclohexane and diethyl ether. They changed parameters, such as pH and type of acid for acidification, but reported no solution for the prob- lem. Forder and Carey (4) also experienced problems with methanol as extraction solvent and finally shifted to methyl formate. Flower and Sun (3) reported that the problem could be resolved by using hydrochloric rather than citric acid in ethyl acetate extractions, but we found the pH to be too critical in order to be entirely reliable. The sugges- tion of Salmon and Flower (2) to wash the ethyl acetate with water prior to evaporation reduced the problem, but did not resolve it entirely. Substituting chloroform for ethyl acetate (2) is not an attractive solution, because this solvent is less polar and 6- keto-PGF,, is not extracted well from the sample

(1). The problems apply to both added and en-

dogenous 6-keto-PGFi, present in the sample. They are not confined to a particular biological fluid. Am- niotic fluid, urine, and water extracts all show this variable and unreproducible elution pattern (Fig. 1). For quantitative measurements by radio- immunoassay, this phenomenon results in a number of problems. A variable amount of [3H]-6-keto- PGF,, co-elutes with other prostanoids (Fig. 3), whose recovery is therefore overestimated.. Al- though one may use specific antibodies for the other prostanoids this does not necessarily exclude the possibility of cross-reactions with erroneously elut- ing 6-keto-PGFi,. Indeed, the immunogenic properties of 6-keto-PGFi, eluting with other pros- tanoids are not necessarily identical to those of the 6-keto-PGE, that elutes in its appropriate fraction

(9). Flower and Sun (3) found that the decomposition

products of 6-keto-PGF1, after extraction and TLC,

Fig. 4 6-keto-PGE, and the hemiketal isomer.

produced mass-spectra on GC-MS identical to 6- keto-PGF,, itself. The suggestion was that 6-keto-PGfi, is probably present as ketal or hemiketal isomers (4, 6, 8) (Fig. 4) and their tautomers. Barrow and colleagues (6) reported that the possibility of ketal formation was avoided in GC-MS analyses because the ketone group was trapped by derivatization with methyl oximating reagent. They suggested that derivatization could better be performed before rather than after purification. This would be an elegant solution for RIAs based on methyl oxime derivates (10).

A few possibilities are now available to overcome the erratic elution pattern of 6-keto-PGF,, on HPLC. Washing ethyl acetate extracts before evaporation reduces the problem, and the use of HCl may nearly eliminate it provided that the pH is kept at a critical level. The use of methyl t-butyl ether for extraction and acidification with hydrochloric acid consistently resolves the problem. It also results in an extract that is more easily evaporated than an ethyl acetate extract.

Acknowledgement

The research was supported by grant 281118 from the Praeven- tiefonds, the Netherlands.

References

1. Wu-Wang C Y, Neu J, Gillespie T A, Yost R A. Purification and quantification of arachidonate metabolites: Critical evaluation of methods and interpretation of results. Semin Perinatol 40: 255, 1986.

2. Salmon J A, Flower R J. Extraction and thin-layer chromatography of arachidonic acid metabolites. Methods Enzymol 86: 477, 1986.

3. Flower R J, Sun F F. Discussion. p. 130 in Prostacyclin ( J R Vane and S Bergstrom eds). Raven Press, New York, 1979.

4. Forder R A, Carey F. Measurement of human venous plasma prostacyclin and metabolites by radioimmunoassay: a reappraisal. Prostaglandins Leukotrienes Med 12: 323, 1983.

5. Mitchell M D, Brunt D J, Webb R. Instability of 6-keto-PGQ, when subjected to normal extraction procedures. Prostaglandins Med 6: 437, 1981.

6. Barrow S E, Waddell K A, Ennis M, Dollery C T, Blair I A. Analysis of picomolar concentrations of 6-oxo-prostaglandin Fjar in biological fluids. J Chromatogr 239: 71, 1982.

7. Noort W A, de Zwart F A, Keirse M J N C. Changes in urinary 6-keto-PGFj, excretion during pregnancy and labor. Prostaglandins 35: 573, 1988.

134 Prostaelandins Leukotrienes and Essential Fattv Acids

8. Powell, W S. Rapid extraction of arachidonic acid metabolites from biological samples using

circulating concentrations. Prostaglandin Med 1: 13, 1978.

octadecylsilyl silica. Methods Enzymol 86: 467, 1982.

9.. Mitchell M D. A sensitive radioimmunoassay for 6-keto-PGF,,: Preliminary observation on

10. Kelly R W, Deam S, Cameron M J, Seamark R F. Measurement by radioimmunoassay of prostaglandins as their methyl oximes. Prostaglandins Leukotrienes Med 24: 1, 1986.