Biochimica et Biophysics Acta, 963 (1988) Sol-508

Elsevier

501

BBA 52991

Prostaglandins E, and FSn are excreted in human urine

after ingestion of n - 3 polyunsaturated fatty acids

Sven Fischer a, Clemens von Schacky a and Horst Schweer ’

’ Medizinische Klinik Innenstadt der LJnroersifZir Miinchen, Miinchen and ’ Kinderklinrk der Universrttit Heidelberg,

Heidelberg (F. R. G.)

(Received 16 May 1988)

(Revised manuscript received 31 August 1988)

Key words: Prostaglandin; Polyunsaturated fatty acid; (Human)

Prostaglandms E, and F3a, presumably of renal origin, were characterized for the first time in urine of volunteers after ingestion of n - 3 polyunsaturated fatty acids by combined gas chromatography-mass spectrometry. Quantitation of prostaglandins E,, E,, F& and FZo: using deuterated internal standards showed low levels of the 3 series prostaglandins in the control period. Levels of prostaglandins E, and Fjcr rose about lo-fold by the 12th week of the dietary trial and were still elevated 4-fold after a wash-out period of 20 weeks. Excretion of prostaglandins E, and FZa tended to be depressed in the 12th week of the dietary trial and rose again to control values after the wash-out period. Our data indicate that n - 3 polyunsaturated fatty acids are incorporated into the human kidney and are retained there for a long time. Prostaglandins E, and FSa may contribute to the observed favorable effects of marine oils rich in n - 3 polyunsaturated fatty acids on certain renal diseases.

Introduction

Prostaglandin (PG) E, and F,, in urine are considered to be of renal origin [l] and may be of relevance in the physiology and pathology of blood flow and excretory function of the kidney [2,3]. The main urinary metabolite, PGE-M of PGE,/ E,, however, is believed to reflect non-renal pro- duction [4]. Animal models [5-81 and also a report in humans [9] have shown beneficial effects of

Abbreviations: PG, prostaglandin: PGE-M, 7a-hydroxy-5,11-

diketotetranorprostane-l,lh-dioic acid; PGE,-M, 7a-hydroxy-

5,11-diketotetranorprost-13-en-1,16-dioic acid; PUFAs, poly- unsaturated fatty acids; SIM, selected ion monitoring; BSTFA,

N,O-bis(trimethylsilyl)trifluoroacetamide.

Correspondence: S. Fischer, Medizinische Klinik Innenstadt

der Universitlt Mtinchen, Ziemssenstrasse 1, 8000 Mtinchen 2,

F.R.G.

dietary n - 3 polyunsaturated fatty acids (PUFAs) on certain renal diseases. As mediators, prosta- noids of the 3 series might be of crucial impor- tance. We therefore analyzed urine of volunteers who had ingested cod liver oil for evidence of

formation of PGE, and F,,. We also searched for PGE,-M, which might reflect non-renal PGE,, in analogy to PGE-M. For analysis of prostaglan- dins, we used combined capillary gas chromatog- raphy-tandem mass spectrometry in the triple and single stage modes, respectively, which are consid- ered to be the most selective and sensitive meth- ods for characterization and quantitation of this class of compound [lo].

Materials and Methods

Study protocol

Urines were obtained from a study that has been described in detail previously [II]. In short,

OOOS-2760/88/$03.50 0 1988 Elsevier Science Publishers B.V. (Biomedical Division)

502

in addition to an otherwise unchanged western diet, six healthy male volunteers ingested cod liver oil (9.4% eicosapentaenoic acid, 13.8% docosahexaenoic acid) in increasing amounts of

10, 20 and 40 ml per day for 4 weeks each without a pause, and then 20 ml daily for 8 weeks. The duration of the dietary study was 20 weeks. For

analysis of PGE,/E,, F3a/F2a and E,-M/E-M, 24 h urine samples were collected before the trial (control values), after ingestion of 40 ml cod liver

oil daily for 4 weeks (12th week of the trial) and 20 weeks after the end of the study. Thus, forma- tion of trienoic prostaglandins at the time point of highest intake of n - 3 PUFAs and also the reten- tion time of these fatty acids in the kidney after the end of the trial were monitored. Urines were kept frozen at - 20 o C and urine samples of four volunteers were analyzed within 6 months after

the end of the trial.

Methods Extraction and purification of PGE,/ E2, F,,/

F,, and E,-M/E-M from urine. Extraction and purification of urinary PGE,/E,, F&FZa and Es-M/E-M for mass spectrometric analysis were

performed according to a method described previ- ously [12]. In short, 50 ng of each 2H,-PGE, and ‘H,-PGF,, were added to 50 ml of urine as inter- nal standards. 3H-PGE2 and 3H-PGF,, were fur- ther added as tracers for straight-phase HPLC. Urine samples were acidified to pH 3.5 (20% HCOOH) and lipids were extracted with a Cl8 reverse-phase cartridge (Sep-Pak, Waters Assoc.) which was washed with water and then eluted with chloroform. After methylation with ethereal CH,N, the sample was dissolved in chloroform and extracted with a straight-phase cartridge (Sep- Pak). The cartridge was washed with chloroform and the sample was eluted with 2.3% methanol in chloroform. 3H-13,14-dihydro-15-keto-PGF2a-Me was added as a tracer for PGE,-M/E-M-Me. Straight-phase HPLC of the methylated pros- taglandins was performed with chloroform (20 min) and then a linear gradient of 2% methanol in chloroform was employed for another 40 mm [12]. Flow rate was 1 ml/mm. 2 and 3 series PGE, F and E-M-Me were collected together relative to 3H-PGE2-Me, 3H-PGF2,-Me and 3H-13,14-dihy- dro-15-keto-PGF,,-Me. Retention times of

authentic standards of PGE,-Me and PGF,,-Me relative to 3H-PGE2-Me and 3H-PGF2,-Me on straight-phase HPLC had been determined by off-line mass spectrometric detection of character- istic fragments. Fractions containing methyl esters

of PGF,,/F,, were derivatized to Me,Si ethers with BSTFA (2 h, 60” C), fractions containing methyl esters of PGE,/E, were derivatized to MO-Me,Si ethers with methoxyamine hydrochlo- ride (12 h, 20 o C) and BSTFA (2 h, 60 ‘C). PGE,- M/E-M-Me was quantified after methoximation via the deuterated methoxime ‘H,-PGE-M-Me-

MO added after straight-phase HPLC as internal standard [13]. Recovery of PGE-M-Me was calcu- lated via 3H-13,14-dihydro-15-ketoprostaglandin F2,-Me. The Me,Si ether was prepared with

BSTFA (2 h, 60°C). Mass spectrometric characterization and anulysis

of urinary PGE,/E?, F,,/F,, and E,-M/E-M.

Mass spectra of PGE,/E,-Me-MO-Me,Si and

PGF,,/F,,-Me-Me,Si were achieved by capillary GC-MS in the positive chemical ionization mode

(NH,). For quantitation molecular ions (M +

NH,)’ 600, 602 and 606 of PGF,,, PGF2, and 2H,-PGF2, were monitored. For measurement of PGE,/E, via deuterated PGE, in the triple stage mode, daughter ions m/z 275, 277 and 281 result- ing from the cleavage of parent ions (M + H)+ 538, 540 and 544 were employed (M + H - 173 -

90)’ corresponding to (M + H - (C+Z,,) - Me,SiOH)‘. For analysis of PGE,-M/E-M-Me- MO-Me,Si in the positive chemical ionization mode (NH,), molecular ions (M + H)+ 485, 487 and 493 of PGE,-M, PGE-M and the deuterated internal standard were monitored. The GC-MS- system was a Finnigan TSQ 45 run in the single quadrupole or triple-stage mode. A fused silica

DB 1701 capillary column was used (J + W Scien- tific).

Materials Cod liver oil was from Miiller, Oslo, Norway,

PGE, and FZa and deuterated 3,3’,4,4’-*H,-PGE, and 3,3’,4,4’-‘H,-PGF,, were a kind gift from Dr. J. Pike, The Upjohn Company, Kalamazoo, MI, U.S.A. Tritiated PGE, (100-200 Ci/mmol) and 3H-PGF2, (150-180 Ci/mmol) were from NEN Research Products, Dreieich, F.R.G. PGE, and F3a were purchased from Cayman Chemical, Ann

503

Arbor, MI, U.S.A. Methoxyamine hydrochloride and BSTFA were purchased from Pierce Chemical Co., Rockford, IL, U.S.A. Deuterated methoxy- amine hydrochloride was prepared according to Ref. 4. All solvents were of analytical grade (Merck, Darmstadt, F.R.G.).

Results

HPLC of PGE,/E,, PGF,,/F,, and PGE,-M/

E-M as methyl esters Authentic PGE,-Me eluted five fractions of 1

ml/mm before 3H-PGE,-Me used as tracer for

PGE,-Me from straight-phase HPLC, when a gradient of 2% CH,OH in CHCl, was employed. The elution time of authentic PGF,,-Me was 6-7

fractions before 3H-PGF,,-Me (Fig. 1). For simultaneous mass spectrometric analysis of PGE,/E, in the urinary samples of the dietary

trial, ten HPLC fractions (fractions 30-39) and for analysis of PGF,,/F,,, 12 fractions (fractions 46-57) were pooled. For analysis of PGE,-M/E- M, six fractions before and four fractions behind 3H-13,14-dihydro-15-keto-PGF,,-Me used as tracer were pooled (fractions 7-17) (Fig. 1). Re- coveries of 3H-PGE,-Me and 3H-PGF2,-Me after HPLC were between 38 and 55%.

Mass spectrometric identification of PGE, and

PGF,, but not of PGE,-M in urine after dietary n - 3 PUFAs

Positive chemical ionization mode (NH,). In Fig. 2A, a mass spectrum of PGE, extracted from urine of a volunteer after ingestion of cod liver oil for 12 weeks is presented. For comparison, a mass spectrum of PGE, from the same urine sample was obtained. A mass spectrum of PGF3a and, for comparison, of PGF,, was also obtained from the

same urine sample (Fig. 2B). PGF,,-Me-Me,Si: m/z 600: (M + NH,)+; m/z 510: (M + NH, - Me,SiOH)+; m/z 493: (M + H - Me,SiOH)+; m/z 403: (M + H - 2Me,SiOH)+; m/z 313: (M

+ H - 3Me,SiOH)+. PGE,-M could not be de- tected in urine as the Me-MO-Me,Si-derivative after dietary n - 3 PUFAs.

Electron impact mode. Mass spectra of endoge- nous PGE, and PGF,, were also obtained in the electron impact mode. PGE,-MO-Me-Me,Si: m/z 522: (M- CH,)+; m/z 506: (M - OCH,)+; m/z 447: (M - Me,SiOH)+; m/z 436: (M -

(Cl,-C,,) - CH,OH)+; m/z 416: (M -

Me,SiOH - OCH,)+; m/z 378: (M - Me,SiOH

- (CKC,,))‘; m/z 288: (M - 2Me,SiOH -

(C,,-C,,))‘; m/z 213 (base peak). PGF,,-Me-

PGE -M-Me PGE3.Me PGF3,- Me

9 3H-13.1L-DH-lSK-PGF2a-Me

98%C/.CHCl~ Z%CH3OH

Fig. 1. Localisation of authentic PGE,-Me and PGF,,-Me on straight-phase HPLC by mass spectrometry monitoring fragment ions

m/z 522 (M -CH,)+ of PGE,-Me-MO-Me,% and m/r 567 (M -CH3)+ of PGF,,-Me-Me,Si in the electron impact mode. 3H-PGE,-Me and ‘H-PGF,,-Me were used as tracers for the methyl esters for both PGE,/E, and PGF,,/F,,. 3H-13.14-dihydro-

15-keto-PGFz,-Me was used as a tracer for PGE-M-Me [12]. The gradient of the elution system is also indicated.

504

100.0

50.0

A

50.0 i

Fig. 2. Mass spectra obtained in the positive chemical ionization mode (NHI) of endogenous PGE, ((A), MH*: m/z 538,

Me-MO-Me,Si derivative) and of endogenous PGF,, ((B), (1%4 + NH,)‘: m/z 600. Me-Me,% derivative). Prostaglandins were

extracted from urine after ingestion of n - 3 PUFAs.

Me,Si: m/z 582: (M)+; m/z 567: (M - CH,)+); m/z 492: (M - Me,SiOH)+; m/z 481: (M - (C,,-CZO) - CH,OH)+; m/z 423: (M - Me,SiOH - (C,,-C,,))‘; m/z 402: (M - 2Me,SiOH)+; m/z 391: (M - (C,,-CzO) - CH,OH - Me,SiOH)+; m/z 333: (A4 - ZMe,SiOH - (C,,-C,,))‘; m/z 312: (M - 3Me,SiOH)+; m/z 301: (M - (ci6-c2a) - CH,OH - 2Me,SiOH)+; m/z 243: (M - 3Me,SiOH - (C,6-C20))+; m/z 217; m/z 191 (base peak).

Mass ~pectro~letric ~u~~t~tat~on of urinary PGE,/

E2, F,,/F2, and E-M after ingestion of n - 3 PUFAs in the diet

SIM profiles of the molecular ions of urinary PGE,/E,-Me-MO-Me,Si obtained in the single quadrupole mode by positive chemical ionization exhibited interfering compounds, especially with PGE,, and were not used for quantitation. SIM profiles of daughter ions obtained in the triple stage mode for simultaneous quantification of PGE,/E, from urine of volunteers after ingestion

505

PCE3

RIG,

449. e 1035

1 F’CE2 A

365.9- 1832

M-FciE2

RICl

1’ 1 r I I, \

I I I u I 960 988 lW30 1820 1840 1060 1080 llea saw

Fig. 3. SIM profiles obtained in the triple stage mode of the daughter ions m/z 275, 277 and 281 of parent ions MH+ 538, 540 and

544 of PGE,, PGE, and internal standard ‘H,-PGE, extracted from urine of volunteer B after ingestion of n -3 PUFAs.

Me-MO-Me,Si derivatives were measured (A). SIM profiles (volunteer A) of urinary PGF,,, PGF,, and the internal standard

‘H,-PGF,, as their Me-Me,Si derivatives in the single quadrupole model (B). Molecular ions (M +NH,)+ 600, 602 and 606 were

monitored. Intensities of tracings and scan numbers are indicated in A and B.

of cod liver oil were free of interfering compounds analyzed in the single quadrupole mode and were (Fig. 3A). A standard mixture of PGEJE, mea- generally free of interfering biological back- sured in the single and triple stage mode exhibited ground, especially with PGFj, (Fig. 3B). For the same ratios under both mass spectrometric quantitative estimation of PGF,,, a corrective fac- conditions. The SIM profiles of the molecular ions tor of 0.43, derived from mass spectrometric anal- of urinary prostaglandins FJu/Fla used for quanti- ysis of equal amounts of PGF,, and PGF,,, was tation by positive chemical ionization were employed as the molecular ion (M + NH4)+ of

506

TABLE I

TIME-COURSE OF URINARY EXCRETION OF PGE,/E, (UPPER PANEL) AND PGFs,/F,, (LOWER PANEL) AFTER

INGESTION OF COD LIVER OIL BY THE FOUR VOLUNTEERS A-D

Control values, values after 12 weeks of the dietary regimen and after a wash-out period of 20 weeks (40th week) are indicated.

Values are expressed in rig/g creatinine (+ S.D.). n.d. = below detection limit; * P ( 0.025; * * P = 0.10 as compared to the controls.

Control 12th week 40th week Control 12th week 40th week

PGE, PGE, A n.d. 565 n.d. 132 3237 241 B 77 95 75 756 410 2459 C 46 259 46 6215 1866 4.570 D 26 386 203 2512 1465 819

22rtI9 326f199 * 81+ 87 240452733 1744+ 1169 2022+1941

PGFsa PG&, A 110 1248 188 1536 1269 1439 B 16 304 67 371 436 978 c 32 612 87 899 791 1430 D 80 196 366 2309 365 1219

6Ort43 5905473 ** 177 * 137 1279& 835 715f 413 1267+ 218

PGF,, had a different percentage of the total ion current as compared to PGF*,. Excretion of PGE- M was 6934 +_ 7243 (control) and 2780 + 821 (12th

week) rig/g creatinine (rt S.D.; n = 3) whereas PGE,-M was not detectable. The detection limit of PGE,-M was about 3% of that of PGE-M.

Kinetics of the urinary excretion of 3 and 2 series

prostaglandins E and F during long-term dietary

n - 3 PUFAs (Table I)

Levels of PGE, and F3a were low in the control period and rose about LO-fold in the 12th week of the dietary trial. Levels after a 20-week wash-out period still were elevated by a factor of 3-4 as compared to the controls. Excretion of PGE, and F,, tended to be depressed in the 12th week of the dietary trial and rose again to control values after the wash-out period. Levels of PGE,/E, and

FXa/FZoI showed a considerable variation between individuals and also between days of collection. However, high values in excretion of PGE,/F,, corresponded on the same collection day to high levels of PGE,/F,, in the same person.

Discussion

In this study, excretion of PGE, and Fsa in urine was demonstrated for the first time in hu-

mans after ingestion of n - 3 PUFAs in the diet. These primary prostaglandins of the 3 series are most probably of renal origin like their 2 series analogues, PGE, and FZa [l]. Non-renal or in- fused PGE, and F;, are known to be quickly deactivated in situ or by a lung passage [14,15] and are excreted in urine as oxidized and chain- shortened inactive metabolites with a skeleton of 16 carbon atoms [4,16]. However, we were not able to detect PGE,-M with a chemical structure analogous to PGE-M in urine of the volunteers. This finding confirmed in humans a different

metabolic pathway for trienoic prostaglandins, as has likewise been described in the rat [17].

Structural proof of urinary prostaglandins of the 3 series was achieved by mass spectra in the positive chemical ionization mode showing the same fragmentation pattern as the commercially

available standards of PGE, and F;3n or standards obtained from biochemically synthesized PGH, [lS]. Structures of urinary PGE, and Fja were also proven by electron impact mass spectra showing intense fragmentation in the upper region and the characteristic allylic cleavage of the methyl side- chain from C,, to C,, (m/z 69) due to the delta- Ill-double bond of 3 series prostanoids [20].

For simultaneous quantitation of PGE,/E, and F3JFZa selected ion monitoring in the electron

impact mode is not suitable, because common fragments of prostaglandins of the 2 and 3 series carry different percentages of the total ion current, as has been shown for thromboxanes [19] and

prostacyclins [20]. Positive chemical ionization with ammonia as reactant gas gave rise to only the MH+ ion for PGE,- and PGE,-Me-MO-Me,% carrying nearly 100% of the total ion current. However, SIM profiles of MH+ of the PGE, derivative exhibited interfering compounds which

were eliminated when measuring corresponding daughter ions in the triple stage mode, as had been shown previously for other urinary pros-

taglandins [lo]. The analysis of the daughter ion was a further proof of the structure and showed that excretion of PGE, was 20% that of PGE,. Since the ratio of a standard mixture of PGE,/E, remained the same, whether obtained either in the single stage mode from the MH+ ions or in the triple stage mode from the daughter ions, the cleavage pattern of PGE of the 3 and 2 series should be identical, and should grant a reliable

quantitation of PGE,. However, the PGF,,/F,, derivative exhibited a

more complex cleavage pattern in the positive chemical ionization mode than the PGE,/E, de- rivative, SIM profiles of equal amounts of PGF,,/F,, showed different values for PGF,,; therefore, a corrective factor for quantitation was introduced. The absolute values for PGF,, should be considered as semi-quantitative, reflecting,

however, a reliable time-course of the formation of PGF,,. Whereas excretion of the main urinary metabolite PGE-M of non-renal PGE,/E, was quantified via the deuterated methoxime, the

structural analogue PGE,-M with an additional delta-17-double bond could not be detected by HPLC, confirming similar results in the rat [17]. Presumably, the delta-17-double bond of the metabolite prevents w-oxidation of the methyl group of the P-side chain to the carboxyl group, as has been equally shown for PGF,, in the rat [21].

Excretion of PGE,/F,, was low in the control period and rose about lo-fold after 12 weeks of the dietary trial with n - 3 PUFAs. 20 weeks after the end of the trial, values of PGE,/F,, were still elevated 3-4-fold. Thus, a very slow turnover of fatty acids in the human kidney is likely. Values of PGE,/E, and F,,/F,, showed high biological

507

variation between the individuals and the different times of urine sampling. However, in general, high levels of PGE,/F,, corresponded to high levels of PGE,/F,, on identical days of collection. The

same pattern of excretion was found for low levels of PGE,/F,,. This fact might be due to con- tamination of the urine in some cases with seminal fluid rich, especially, in PGE, [22]. Therefore, in

the urines with very high values of PGE,, only a minor part of PGE,/E, might be of renal origin.

Endogenous PGE, has been found previously in seminal fluid of volunteers on a western diet [23]. PGE, possesses smooth muscle-stimulating activ-

ity on isolated rabbit duodenum by a factor of 10 smaller than that of PGE, [23]. In vivo formation of PGF,, in man has not yet been demonstrated, to our knowledge.

The mechanisms by which kidney diseases in animal models and also in humans can be allevia-

ted by dietary n - 3 PUFAs are unknown. Struct- ural alterations of biomembranes but also func-

tional effects of 3 series prostanoids may be in- volved. Further studies are needed to elucidate the

biological activities of PGE, and F3,, their possi- ble roles in the kidney in states of health and disease and their contribution to the beneficial effects of n - 3 PUFAs on this organ.

Acknowledgement

The authors are supported by the Deutsche Forschungsgemeinschaft.

References

Frolich, J.C., Wilson, T.W., Sweetman, B.J., Smigel, M.,

Nies, AS., Carr, K., Watson, J.T. and Oates, J.A. (1975) J.

Clin. Invest. 55, 163-770.

Stark, J.E., Rahman, M.A. and Dunn, M.J. (1986) Am. J. Med. 80, 34-45.

Schlondorff, D. and Ardaillou, R. (1986) Kidney Int. 29,

108-119.

Hamberg, M. and Samuelsson, B. (1971) J. Biol. Chem. 246, 22, 6713-6721.

Prickett, J.D., Robinson, D.R. and Steinberg, A.D. (1981) J. Clin. Invest. 68, 556-559.

Prickett, J.D., Robinson, D.R. and Steinberg, A.D. (1983) Arthritis Rheum. 26, 133-139.

Kelley, V.E., Ferretti, A., Izui, S. and Strom, T.B. (1985) J. Immunol. 134, 1914-1919.

Robinson, D.R., Prickett, J.D., Polisson, R., Steinberg,

A.D. and Levine, L. (1985) Prostaglandins 30, 51-75.

508

9 Hamazaki, T., Tateno, S. and Shishido, H. (1984) Lancet ii,

1017-1018.

10 Schweer, H., Seyberth, H.W. and Schubert, R. (1986) Bio-

med. Environ. Mass Spectrom. 13, 611-619.

11 V. Schacky, C., Fischer, S. and Weber, P.C. (1985) J. Clin.

Invest. 76, 1626-1631.

12 Mtiller, H., Mrongovius, R. and Seyberth, H.W. (1981) J.

Chromatogr. 226, 450-454.

13 Hamberg, M. (1972) B&hem. Biophys. Res. Commun. 49,

720-726.

14 Anggard, E., Bohman, S.O., Griffin, J.E., Larsson, C. and

Maunsback, A.B. (1972) Acta Physiol. Stand. 84, 231-235.

15 Piper, P.J., Vane, J.R. and Wyllie, H.H. (1970) Nature 225,

600-604.

16 Grandstrom, E. and Samuelsson, B. (1971) J. Biol. Chem.

246. 5254-5263.

17 Diczfalusy, U. and Hamberg, M. (1986) Biochim. Biophys.

Acta 878, 387-393.

18 Smith, D.R., Weatherly, B.C., Salmon, J.A., Ubatuba, F.B.,

Gryglewski, R.J. and Moncada, S. (1979) Prostaglandins

18, 423-438.

19 Fischer, S. and Weber, P.C. (1983) Biochem. Biophys. Res.

Commun. 116, 3,1091-1099.

20 Fischer, S. and Weber, P.C. (1984) Nature 307, 165-168.

21 Dimov, V. and Green, K. (1973) Biochim. Biophys. Acta

306, 257-269.

22 Templeton, A.A., Cooper, I. and Kelly, R.W. (1978) J.

Reprod. Fertil. 52, 147-150.

23 Samuelsson, B. (1963) J. Biol. Chem. 238, 3229-3235.