Abstract Introduction - Science Direct

Prostaglandins

45:413-426,

1993

MEASUREMENT OF EICOSANOIDS IN MENSTRUAL FLUID BY THE COMBINED USE OF HIGH PRESSURE CHROMATOGRAPHY AND RADIOIMMUNOASSAY G. Hoferl Ch. Bieglmayer',

B. Kopp3,

H.

Janischl

1Second Department of qbstetrics and Gynecology, University Hospital of Vienna, Institute for Medical and Chemical Laboratory Diagnostics, University Hospital of Vienna, Wahringergurtel 18-20, A 090 Vienna, Austria (Reprints request addrese) and 3PhanUaCOgnOBtiC Institute, University of Vienna

Abstract Methods are described for the quantification of various eicosanoids (cyclooxygenase products: 6-KETO, TXB2, PGE2, PGF2a, DHK; lipoxygenase products: 5, 12-, 15-HETE, LTB4, LTC4, LTD4, LTE4) in menstrual blood collected by tampons. Samples were extracted with acidified ethanol. After purification by SEP-PACK Cl8 columns, the compounds were separated by reversed phase HPLC using a ternary gradient system. The eicosanoid concentrations of the fractionated eluents were measured by radioimmunoassay and corrected for recovery. 12-HETE was the most prominent metabolite of arachidonic acid in menstrual blood (mean: 1174 rig/g blood). With the exception of PGF2a and TXB2 (mean: 343 and 212 rig/g blood, respectively) other eicosanoids were detected in remarkable lower concentrations. Supported b,y grant #P7137 from the Austrian ‘%bnds zur Fiirderung der wissenschaftItchen Forschung ‘:

Introduction Eicosanoids represent a large family of products generated by oxidative metabolism of polyunsaturated C20 fatty acids. The most abundant precursor in humans is the 5,8,11,14-eicosatetraenoic acid (arachidonic acid). Oxidation of arachidonic acid can occur via three major pathways. Prostaglandins (PG) and thromboxanes (TX) are formed by the cyclooxygenase system. Hydroxyeicosatetraenoic acids (HETE), leukotrienes (LT) and lipoxins are products of various lipoxygenases. Additionally epoxyeicosatrienoic acids are produced by cytochrome I’450 epoxygenases. Many of these metabolites have potent biological activities. Besides other effects, they play an important role in reproduction and menstruation. Since Pickles 1965 reported enhanced PC levels in endometrium and men-

Copyright

0 1993 Butterworth-Heinemann

Prostaglandins

414

strual fluid of dysmenorrhoeic women (l), several approaches have been made to measure eicosanoids in menstrual fluid or uterine tissues. Basically three methods have been applied: endomentrial jet wash (2,3), collection of menstrual fluid in diaphragms (4-6) or in tampons and sanitary pads (7-13). These procedures not only differ in the way of sample collection, but also in the methods of quantification. No attempts have been made to analyse lipoxygenase products in the menstrual fluid until now. The aim of the present study was to develop a method for the quantification of various eicosanoids in menstrual fluid collected on tampons, included a solid phase extraction, reversed-phase HPLC separation of cyclooxygenase and lipoxygenase products within one run and RIA measurement of isolated metabolites.

Materials

and methods

Reagents 3H-labeled eicosanoids and steroid hormones were purchased from NEN eicosanoids from Bioscience (USA) and Amersham (UK), unlabeled Products (Switzerland), Cayman Chemicals (USA) and Serva (Germany). Solvents for extraction and HPLC from Merck (Germany) had HPLCgrade quality. Tampons (‘O.B.‘-normal) were a generous gift from Johnson & Johnson (Austria). Radioimmunoassay (RIA) kits for the measurement of eicosanoids were purchased from Advanced Magnetics (USA). Samples The volunteers transferred the used tampons immediately after removal to weighted plastic containers (Art.No.: 5-64882, Nunc, Denmark) which were filled with 30ml of a 8:2 mixture of ethanol and 250mM formic acid. The volume of the alcoholic solution was sufficient to keep the tampons completely submerged during a short storage in a refrigerator and their transport to the laboratory. The volunteers identified the sample by their name, day of birth and the date of collection on a label attached to each container. Furthermore the volunteers noted how long the tampon was in use and estimated their subjective bleeding intensity. Additionally they declared whether they had taken any drugs, especially hormonal contraceptives. Number of samples are given in the tables. In the laboratory we reweighed the containers to determine the menstrual fluid volume, assuming lg equal to lml. The original screw caps were replaced by perforated caps to collect the alcoholic extracts by centrifugation

(l,OOOg at 4°C for 10 minutes,

JG/Beckman,

USA).

The tampons

Prostaglandins

415

were reextracted with 20ml of the ethanol-formic acid mixture and centrifuged. The combined extracts were evaporated to dryness under reduced pressure in a Rotavapor (B&HI, Switzerland). The residues were dissolved in 1.5ml absolute ethanol followed by addition of 8.5ml 60mM formic acid with 0.12% Tween20 (SIGMA, USA). The resulting suspensions were centrifuged (5O.OOOgat 4°C for 30 minutes, Sorvall/DuPont, USA). Pellets were resuspended as described above and centrifuged once more. The combined clear supernatants were loaded to C18-cartridges (SEP-PAK, Waters, USA) which had been primed with ethanol and 50mM formic acid. The cartridges were subsequently washed with 15% ethanol in 50mM formic acid (lOml), 50mM formic acid (10ml) and n-hexane (10ml). Eicosanoids were eluted with 2ml methanol and evaporated to dryness in a vacuum centrifuge (Speedvac, USA). The residues were stored at -70°C until HPLC separation. HPLC Reversed- phase HPLC was carried out using a solvent delivery system with low-pressure mixing valves, a dual-head pump, a variable wavelength monitor and a fraction collector suitable for time window programming from LKB/Pharmacia, Sweden. A column of Spherisorb ODS-2 (125x4.6mm, 3pm particle size) served as stationary phase. The column was equipped with two guard cartridges, one filled with ODS-2/3pm, and the other filled with Lichrospher C18/5pm serving as pre-guard cartridge. The mobile phase consisted of gradients between solvents A (acetonitrile/methanol/formic acid 59:41:0.07), B (water/formic acid 100:0.03) and C (water/formic acid 100:0.03 adjusted to pH 4.1 with ammonium hydroxide): O-10 minutes isocratic 54%A + 46X8, lo-15 minutes gradient to 66%A + 34x8, 15-22 minutes isocratic 66%A + 34%B, 22-22.2 minutes, gradient to 54%A + 46%B, 22.2 to 30 minutes gradient to 68%A+32%C. At the end of each run the column was flushed with 30ml lOO%A and reequilibrated to the initial conditions. The extracted sample residues were reconstituted in 100~1 methanol and solid particles were removed by centrifugation. 20~1 of the supernatants were injected. The flow rate was 1.3ml per minute at an initial pressure of 22MPa. The effluent was monitored simultaneously at 237~1 and 28Onm to detect monohydroxy acids and conjugated trienes, respectively. Effluent fractions corresponding to the reference compounds were collected and stored at -20°C.

Prostaglandins

416

Standardisation,

RIA, recovery

The retention times of all compounds measured by RIA were determined with 3H-labeled standards and additionally with unlabeled standards, except for PG. The retention time of HHT was only verified with an unlabeled standard. The elution behaviour of the reference compounds was checked each day. The concentrations of LTB4 and 12-HETE were measured by an additional external standard calibration and were compared with the results from RIA measurements. Aliquots of the appropriate HPLC fractions were transferred to polystyrene tubes and evaporated to dryness in the vacuum centrifuge. The residues were reconstituted in 100111 of RIA buffer. RIA procedures were according to manufacturers suggestions. The fractions corresponding to LTC4, LTD4 and LTE4 were analysed with the same MA-kit, thus the concentrations of each LT were obtained from cross-reactivity calculations (55% LTC4, 100% LTD4, 51% LTE4). To determine the recovery of the eicosanoids, SH-labeled standards were mixed with different volumes (l-10ml) of pooled peripheral blood and loaded on tampons. These samples were extracted and separated by HPLC as described for the menstrual blood samples. Radioactivity of the fractions was used to calculate the recovery rates. The RIA results of extracted and HPLC-separated volunteers samples were corrected for the overall recovery of the respective compound according to the weight of the collected menstrual blood. The possibility of unknown cross-reactivities was tested with HPLCeluents from menstrual fluid samples and blank tampons (loaded with distilled water) fractionated in intervals of one minute and analysed with RIAs.

Results The retention times of several eicosanoids and steroid hormones are shown in table 1 and figure 1A. Fractions were chosen for automated collection of effluents for further RIA-determinations. The PGs were collected in one fraction from 1:00 to 4:00 (minutes:seconds). LTB4 was collected from 6:30 to 8:30. HHT was collected from 10:00 to 12:00, the HETEs between 17:30 to 21:30, LTC4 from 24:40 to 26:40, LTE4 from 26:40 to 28:40 and LTD4 from 29:00 to 31:O0.

Prostaglandins

417

Table 1, Retention times of eicosanoids and steroid hormones a) not measured by photometry due to an interference of the eluent solution Numbers in table correspond to numbers in figure 1

-

NO

Eicosanoid

1 -

20-COOH-LT6.t

2 3 4 5 6 7 8 9 -10 11 12 13 14 15 16 17 18 19 20 -21 22 23 -24

mm:ss I:30

6-KETO TXBz PGE2

I:40

2:oo 2:20 2:30

PGFza DHK Estradiol

PGB2 8.15DiHETE 5;15-DiHETE 6-Vans-LTB4 6-Vans-12-epi-LTB, 4 LTB4 ’ HHT Progesterone

S(S),G(R)-DiHETE 5(S),G(S)-DiHETE

I

ISHETE 11 -HETE IS-HETE 5-HETE LTC4 LTE; LTD4

c

L

3:oo 3:30 4:45 530 6:30 6:45 7:oo 7:15

A 280nm 192nm (a) (a) (a) (a)

*192nm a192nm *192nm ’192nm

(a) 280nm

280nm 280nm 237nm 280nm 280nm 280nm

IO:15

237nm

II:20 14:30 15:40 18:20 19:lO 19:40 21 :oo 25:30 27:15 29:45

237nm

280nm 280nm 237nm

237nm 237nm 237nm 280nm 280nm 280nm

Extracts of blank tampons contained a variety of compounds which we were unable to characterise. These substances neither derived from the plastic containers nor solvents used for extraction. Apparently they derived either from fatty acids bound to the tampon material or from artefacts formed during the production of the cellulose and could not be removed by the precleaning procedures (figure 1B). Chromatograms obtained with peripheral blood were almost identical with those from blank tampon extracts (data not shown). In chromatograms of menstrual fluid extracts (figure 1C) some additional peaks were detected. Distinct peaks of LTB4, HHT and 12-HETE were visible in all samples.

Prostaglandins

418

standard

blank

0

5

10

IS

20

Figure 1. A) Reversed phase HPLC of standard mixture, C) menstrual blood extract. AU: arbitrary absorbance compound numbers

tampon

25

mixture

exlracl

30 mmutss

B) blank tampon extract and units; Compare table 1 for

419

Prostaglandins

To test whether tampon impurities or cross reactants may influence the results, one minute fractions were collected over the whole run and measured by RIAs. The 15-HETE-RIA showed negligible immunoreactivity with blank tampon extracts in fractions corresponding to the retention time of the 15-HETE standard (figure 2).

T : 0

15-HETE

I

:

:

: 5

I_

ml”“k?, , 10

Figure 2. Distribution of 15HETE extracts fractionated by HPLC.

15

20

immunoreactivity

25

30

35

in menstrual blood and blank tampon

The LTCeRIA-kit exhibited some cross-reactivity with blank tampon extracts as well as with several other compounds present in menstrual fluid (figure 3).

Figure 3. Distribution of peptido-LT immunoreactivity pon extracts fractionated by HPLC.

in menstrual blood and blank tam-

420

Prostaglandins

The LTB+RIA-kit revealed a remarkable sensitivity for 12-HETE in menstrual blood samples without any immunoreactivity to blank tampon extracts (figure 4). Other RIA-kits showed no cross-reactions (not shown).

0

10

5

Figure 4. Distribution fractionated by HPLC.

of LTB4

25

20

15

immunoreactivity

35

30

in an extract

of menstrual

blood

The peak areas of LTB4 and 12-HETE in chromatograms of menstrual fluid samples revealed a high correlation to the RIA results (r=0.97 and r=0.72, respectively, figure 5 A,B). 1200 1

.

I

1ow

.

.

r=0,72

.

l

.

. 600

.

2 E

l

6W

. *

l

l

P

l

.

.’

.

. .*

4oil

* .+ _e* . ml

. .

(A) LTB4 : +-_____ 0

10

i

+,

20 30 40 nglml HPLC

/

50

*r

.

. (B)

.

IZHETE

i

0 60

Figure 5. Comparison of HPLC quantification

0

2w

‘ml nghl

ml

ml

,000

HPLC

versus RIA, (A) LTB4, (B) 18HETE

,200

421

Prostaglandins

The concentrations of 12-HETE measured with the RIA-Kit were 1.4 fold higher than those calculated from the peak areas in the chromatograms. On the other hand the 12-HETE-RIA showed an increased immunoreactivity to a purchased 12-HETE standard separated by HPLC, than to its own internal kit-standard. Because of the uncertain standardisation we used only the RIA data for further evaluations. The intraassay and the interassay variation coefficients of the RIA-kits were determined by repeated measurement of several selected samples serving as precision controls (table 2). Table 2. Intraassay- and interassay- coefficients of variation (CV) of the eicosanoid RlAs

Metabolite

lntraassa -CV lnterassa -CV

~~ -I.

I

DHK 5HETE HETE 12-HETE 15HETE LTl34 LT ( LTCdD4/E4 1

WI”

14% 3% 14% 10% 14% 12%

\‘.

w,

(n=4) (n=2) (n=4) (n=4) (n=8) (n=12)

,

I”

13% 12% 16% 10% 25%

\.’

I

,

(n=6) (n=6) (n=6) (n=7) (n=4)

Storage of samples in the HPLC eluents at -20°C and repeated thawing and freezing did not destroy the metabolites over a period of 6 month. The recovery rates of the metabolites after extraction and HPLC separation depended inversely on the amount of blood loaded on the tampons (table 3). The correlation was linear between lml and 8ml blood volume (not shown). The metabolites of related structure showed similar recovery rates. All MA-results were corrected by individual recovery-factors, which were calculated from the menstrual fluid volume of the respective samples. Concerning the concentration of eicosanoids in menstrual blood we report only on samples from women, who did not complain about menstrual pain and did not use hormonal contraceptives (table 4). 12-HETE was the predominate metabolite of arachidonic acid in menstrual blood. PGF2o and TX82 also appeared in high concentrations.

Prostaglandins

422

Table 3. Recovery rates of eicosanoids from peripheral blood sucked into tampons

) PG

I Table 4. Average concentrations afler HPLC-separation.

I

1

77%-54%

LTE4 1

26%-13%

of eicosanoids

.._1_1__I.I_

1

I

I

6-KETOI --_

PGFwI

LT

1 I-8ml blood

Metabolite 6-KETO

in menstrual blood measured

-

-_--

281 -~-

,__,_\

by RIA

O,,“,

’

--

21 t

LTl34 LTC4 LTD4

22 15 17

2513 991

66% 63%

LTE4

15

16,0

65%

Discussion During the last years several attempts have been made to elucidate the role of PGs in menstruation. The results of some studies indicate, that PGs may be important mediators in menstrual disorders such as menorrhagia and dysmenorrhoea. However there is no information about the concentrations of lipoxygenase products in menstrual fluid. Since Rees has de-

Prostaglandins

423

tected lipoxygenase activity in human endometrium and myometrium, it must be assumed that these metabolites may also serve an important function in menstruation (14). The aim to measure PG as well as HETE and LT from the same sample made it necessary to develop a new extraction procedure for the menstrual fluid samples. Pure ethanol as storage medium for used tampons had little effect on the recovery of PG, but reduced the recovery of the HETE and the peptido-LT to less than a half and to 5-10% of the actual values, respectively. Reducing the ethanol concentration below 80% was not advisable as the alcohol was used not only for extraction but also to inhibit extracorporal enzymatic conversion of the arachidonic acid metabolites. To obtain satisfactory recoveries of all metabolites, acidifying of all solvents used in the extraction process was essential. Resuspension of the evaporated extracts was another critical step. The addition of Tween 20 significantly reduced the amount of precipitated material, which tended to adsorb the eicosanoids. To remove non-polar lipids from the SEP-PAK cartridges n-hexane behaved superior over petrol ether (15), because it did not influence the HETE-recovery and was useful to remove all water from the columns. Therefore the evaporation time after the elution with methanol could be reduced. The HPLC-system based on an idea of Borgeat (16). However we used only volatile solvents in regard to the following RIAs. Due to the gradient profile and the particular ratio of acetonitrile and methanol, the PG eluted right behind the front peak, and the LTB4-isomers as well as the HETEs were well separated. The peptido-LT were eluted by rising the pH of the mobile phase. The baseline drift at 237nm caused by the increasing concentration of methanol in the solvents over the first 20 minutes was eliminated by a concomitant reduction of formic acid. This modification allowed an UV control of the eluted compounds at high sensitivity. HPLC separation of eicosanoid fractions was necessary since the LTC4RIA exhibited marked cross-reactions with other compounds, especially with LTD4 and LTE4. Similarly the kits for LTB4 and to a less extend for 15-HETE showed some cross talk with unrelated fractions. Blank tampon extracts showed several distinct peaks at 237 and 280~1, which were eluted after 17-20 minutes. Mostly the areas of these peaks decreased in relation to the loaded blood volume. Despite of several attempts we failed to get rid of these possible interferences. LTB4, 12-HETE and additionally HHT (unpublished observations) could be quantified by peak area measurements. The discrepancies to 12-HETE

Prostaglandins

424

RIA results may be due to an incorrect concentration of the purchased 12HETE standard. For experimental reasons the recovery rates had to be elucidated with 3Hlabeled compounds dissolved in peripheral blood. Recovery rates did not change significantly when tampons were incubated for one to six hours at 37°C suggesting a sufficient stability of metabolites during tampon application. Similarly no significant degradation occurred when tampons were store in acidic ethanol at room temperature for one day or in a refrigerator for one week. The negative correlation of recovery to the blood volume could not be explained by a limited penetration of the acidic ethanol used for extraction, since the tampons were squeezed by centrifugation and reextracted with fresh solvent. By using a dye we observed that penetration of ethanol even into used tampons was a rapid process. Most likely adsorption effects of denatured proteins reduced the recovery. A reabsorption of metabolites cannot be excluded during tampon application. However outward diffusion may be restricted by the flow of liquid into the tampon due to the sucking behaviour of the compressed cellulose. Eicosanoid concentrations in menstrual blood were scattered in a wide range. The most abundant metabolite was the lipoxygenase product 12HETE. Concentrations of PC metabolites were similar to previous findings (5,6,10). Preliminary observations from menstrual fluid of dysmenorrhoeic patients showed an increased metabolism of arachidonic acid not only by the cyclooxygenase but also by the lipoxygenase pathway. Oral contraceptives reduced PG production and the synthesis of 5- and 12-HETE and LTB4.

Acknowledgements The study was supported by grant #P7137 from the Austrian Fiirderung der wissenschaftlichen Forschung”.

“Fonds zur

Abbreviations DiHETE

(dihydroxy-eicosatetraenoic

acid),

DHK

(13,1-dihydro-15-keto-

prostaglandin F~M), HETE (monohydroxy-eicosatetraenoic acid), HHT (12-hydroxy-5,8,10-heptadecatrienoic acid), 6-KETO (6-keto-prostaglandin FIN), LT (leukotriene),

PG (prostaglandin),

TX (thromboxane).

425

Prostaglandins

References 1. Pickles,V.R., Hall, W.J., Best, F.A. and Smith, G.N. Prostaglandins in endometrium and menstrual fluid from normal and dysmenorrhoeic subjects, JObstet.Gynaecol. Br.Common. 72,185-192,1965 2. Demers, L.M., Halbert, D.R., Jones, D.E.D. and Fontana, J. Prostaglandin F levels in endometrial jet wash specimens during the normal human menstrual cycle. Prostaghdins, 10, 1057-1065, 1975 3. Halbert, D.R., Demers, L.M., Fontana, J. and Jones, D.E.D. Prostaglandin levels in endometrial jet wash specimens in patients with dysmenorrhoea before and after indomethacin therapy. Prostaglandins, l!I, 1047-1056,1975 4. Pulkkinen, M.O. blood prostaglandin 137-142,1979

and Csapo, A.I. Effect levels in dysmenorrhoeic

of ibuprofen on menstrual women. Prostaglandins, 18,

5. Lumsden, M.A., Kelly, R.W. and Baird, D.T. Primary dysmenorrhoea: the importance of both prostaglandins E2 and F2a. Br.J.Obstet.Gynaecol., 90,1135-1140,1983 6. Rees, M.C.P., Anderson, A.B.M., Demers, L.M. and Turnbull, AC. Prostaglandins in menstrual fluid in menorrhagia and dysmenorrhoea. Br.J.Obstet.Gynaecol., 91, 673-680, 1984 7. Chan, W.Y. and Hill, J. Determination of menstrual prostaglandin levels in non-dysmenorrhoeic and dysmenorrhoeic subjects. Prostaglandins, 15, 365-375,1978 8. Chan, W.Y., Dawood, M.Y. and Fuchs, F. Prostaglandins in primary dysmenorrhoea. Comparison of prophylactic and nonprophylactic treatment with ibuprofen and use of oral contraceptives. AM.JMed., ZL!,535541,198l 9. Chan, W.Y., Fuchs, F. and Powell, A.M. Effects of naproxen sodium on menstrual prostaglandins and primary dysmenorrhoea. Obstet.Gynecol., E&285-291,1983

Prostaglandins

426

10. Powell, A. M., Chan, W.Y., Alvin, P. and Litt, T.F. Menstrual PGF2a, PGE2 and TXAZ in normal and dysmenorrhoeic women and their temporal relationship to dysmenorrhoea. Prostaglandins, 29,273-289,1985 11. Zahradnik, HP., Stengele, E., Kraut, E., Scharpf, G. and Breckwoldt, M. Neue Aspekte zur Pathogenese und Therapie der Dysmenorrhoe. Dtsch.med. Wschr,, lQ3,1270-1273,1978

12. Zahradnik, HP. Wirkungsmechanismus von Dydrogesteron bei Dysmenorrhoe. Prostaglandinspiegel im Menstrualblut. FortschrMed., lO2, 439-442,1984

13. Zahradnik, H.P. and Breckwoldt, M. Contribution to the pathogenesis of dysmenorrhoea. Arch.Gynecol,, 2X,99-108,1984 14. Rees, M.C.P., Di Marzo, V., Tippins, J.R., Morris, H.R. and Turnbull, A.C. Leukotriene release by endometrium and myometrium throughout the menstrual cycle in dysmenorrhoea and menorrhagia. J.Endocr., 113, 291-295,1987

15. Powell, W.S. Rapid extraction of arachidonic acid metabolites from biological samples using octadecylsilyl silica. Methods in Enzymology, sh, 467-477,1982

16. Borgeat, P., de Laclos, B.F., Rabinovitch, H., Picard, S., Braquet, P., Hebert, J. and Laviolette, M. Generation and structures of the lipoxygenase products. Eosinophil-rich human polymorphonuclear leukocyte preparations characteristically release leukotriene C4 on ionophore A23187 challenge. ~Allergy.Cli~~.Immrmol., Z&310-315,1984 Editor:

E.

Granstrom

Received:

1-22-93

Accepted:

3-11-93