- Email: [email protected]
A convenient, unified scheme for the differential extraction of conjugated and unconjugated serum C19 steroids on Sep-Pak C18-cartridges
J. steroid Biochem. Vol. 33, No. 2, pp. 289-295, 1989 Printed in Great Britain. All rights reserved
0022-4731/89 $3.00 + 0.00 Copyright 0 1989Maxwell Pergamon Macmillan plc
A CONVENIENT, UNIFIED SCHEME FOR THE DIFFERENTIAL EXTRACTION OF CONJUGATED AND UNCONJUGATED SERUM Cl9 STEROIDS ON SEP-PAK C,,-CARTRIDGES DONNA W. PAYNE,‘***W. DAVID HOLTZCLAW’ and ELI Y. ADASHI’ ‘Department of Obstetrics and Gynecology, University of Maryland School of Medicine, Baltimore, Maryland 21201, and 2Department of Pharmacology and Molecular Sciences, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, U.S.A. (Received 24 September 1988; received for publication 3 March 1989)
Snmmary-In order to conveniently and rapidly isolate by group both conjugated and unconjugated serum androgens, a scheme has been devised for their di~erential extraction from ~rnrne~i~y available, disposable octadecylsilane cartridges (Sep-Pak C,r). Using added radioactive steroid standards and detection of endogenous serum steroids by group-specific enzymatic assays, the quantitative recovery of steroid glucuronides and sulfates in the 47% methanol fraction and of unconjugated steroids in the 100% methanol fraction was observed. Maximum recovery of serum protein-bound steroids (e.g. testosterone) was achieved with serum denatured by urea and heat. In order to separate glucuronides from sulfates, sequential hydrolysis of the conjugated fraction (47% methanol) by enzymatic hydrolysis and then organic solvolysis as well as an additional SepPak cartridge extraction step was required. Groups of extracted steroids may be further separated and assayed by any appropriate method(s). An application is given which employs HPLC and an enzymatic assay for 17&hydroxy- and IFoxo-steroids to provide separate profiles of unconjugated, glucuronidat~, and sulfated androgens in human, male serum.
INTRODUCTION Many methods have been described for the extraction
of steroids from biological materials [l]. A convenient, reliable and efficient method for the extraction of steroids from aqueous solutions employs small disposable cartridges packed with silica gel containing an octadecylsilane bonded phase. Methods have heen described using Sep-Pak C,, or similar commercially available materials for the extraction of conjugated and unconjugated androgens, estrogens, corticoids, progestins and bile acids from urine, milk, tissue, cell culture and serum [2-6]. However, none of these methods is designed to use the same cartridge to extract groups of conjugated and unconjugated steroids into separate pools although this would be of use in studies where both groups must be measured. For example, many recent studies have suggested that the measurement of conjugated as well as unconjugated androgens is important in understanding the etiology of hirsutism in women. Such studies would benefit from a convenient, simultaneous method for the extraction of both conjugated and unconjugated androgens. Therefore a simple and unified scheme was devised for the Sep-Pak cartridge extraction of serum C9 steroids. Using this procedure, groups of *Address all correspondence to: D. W. Payne, Dept. of Obstetrics and Gynecology, University of Maryland School of Medicine, 655 W. Baltimore St, Baltimore, MD 21201, U.S.A. 289
sulfates, glucuronides and unconjugated steroids are obtained by nearly identical means. In addition to the simplicity and relative rapidity of this method, comparisons of conjugated and unconjugated steroids in biological samples should be improved because of the elimination of procedural differences during extraction. EXPERIMENTAL Materials
fl-Glucuronidase (Type VIII from E. cd), glucuronidase-sulfatase (Type H-5, from I% pomatia), dehy~~piandrosterone-3-glucuronide (DHEA-G), androsterone-3-glu~uronide (androsterone-G), and testosterone-17-glucuronide (testosterone-G) were obtained from Sigma (St Louis, MO.). Unlabelled unconjugated steroids were obtained from Steraloids (Wilton, N.H.) and [)H]-labelled steroids from Amersham (Arlington Heights, Ill.) or New England Nuclear (Boston, Mass.). Sera were obtained from normal adult, human males and stored at -20°C until used. Extraction
procedure
Serum or plasma (1 ml), with or without urea/heattreatment (1 ml serum plus 1 ml of 4 M urea, 60°C 30 min), was applied to Sep-Pak CIB cartridges (Waters/Millipore, Milford, Mass.) prepared by washing with 2 ml methanol (HPLC grade) and 5 ml deionized, distilbd H,O. The following fractions
290
D~NNAW.
PAYNE
~uREA~HEAT . Tfemd SEdUM I
Ssp-P&,s Dirurd Effluent --I Hz0 Dirard Wash --I 47% hOOH
~HYDAOLYZED
SULFATES[
Fig. 1. Extraction/hydrolysis
were then collected in order (see Fig. 1): serum effluent; 5ml H,O, 3-5 ml 47% methanol (conjugated steroids); 3-5 ml 100% methanol (unconjugated steroids). Serum and all eluents were passed through the cartridge either by syringe or by suction using an &place vacuum manifold (Waters/Millipore, Milford, Mass.). The limiting factor in the time required to perform the extraction(s) is the size of the vacuum manifold since the time required (less than 5 min) to pass all the eluents through the cartridges under vacuum is the same whether one, eight, or more cartridges are present.
procedure (see text for details).
HPLC and steroid ~~y~i~ Dried Sep-Pak fractions containing steroids were dissolved in a small volume of 2-propanol: hexane (5:95) and chromatographed on a Lichrosorb Diol column (EM Reagents, Gibbstown, N.J.) using a concave gradient of 2-propanol : hexane as described earlier [8]. Steroids were detected and quantified using highly specific bacterial hydroxysteroid dehydrogenases [9]. 17@-Hydroxy- plus 17-oxo-steroids were measured by a ~anshydrogenase (TH) assay [lo]. 3a -, 38-, and 17@-Hydroxysteroids were measured by enzymatic oxidation as described previously [8].
Conjugate hydrolysis The 47% methanol fraction was evaporated to dryness (Speed-Vat, Savant Instr., Hicksville, N.Y.), and the residue reconstituted in 1 ml 75 mM potassium phosphate buffer (pH 7). The glucuronides were hydrolyzed at 37°C overnight with I?. coli /?glucuronidase (5 Units) and extracted in the 100% methanol fraction of a second Sep-Pak cartridge (Fig. 1). The 47% methanol fraction of the second cartridge (Fig. 1) contained unhydrolyzed sulfates which were dried, reconstituted in 3 ml ethyl acetate containing 2.5 ~1 concentrated HCl and incubated for 1.5 h at 37°C [7]. In some ex~~rnen~ (not part of standard protocol) where it was convenient to simultaneously hydrolyze sulfates and glucuronides, the conjugates were reconstituted in 1 ml 25 mM sodium acetate buffer (PH 4.5) containing glucuronidase/ sulfatase and hydrolyzed at 37°C for 24 h. Preliminary experiments were performed to determine the optimal conditions of hydrolysis which are described above.
RESULTS AND DISCUSSION
Comparison ofpretreatment md extraction procedures Initially, extraction of endogenous serum and plasma steroids by Sep-Pak cartridges was compared to extraction by a variety of organic solvent extraction procedures. The enzymatic assay used (TH Assay) can detect any steroid containing a 178hydroxy or 17-0~0 function (& steroids), and thus proved very useful in the reliable evaluation of the maximum quantities of endogenous serum steroids which could be extracted by the various methods. Whereas extraction by Sep-Pak cartridges was as good as, or better than, organic solvent extraction for most unconjugated steroids (not shown), this was not so for testosterone and some other androgens. 17/?Hydroxysteroids strongly interact with serum sexbinding globulin and many steroids bind (less stroigly) to serum albumin [ll]. Presumably this prevents their quantitative extraction from Sep-Pak
Extraction of serum steroids
291
Fig. 2. Effect of pretreatment with urea and/or heat on the extraction of endogenous plasma steroids. Four 1 ml aliquots of normal human male plasma were diluted with either 1 ml of phosphate buffered saline pH 7 (C, 60”) or with 1 ml 4 M urea (U, 60” + U) and incubated at either room temperature of 60°C for 30 min as noted. Each sample was then extracted and conjugates hydrolyzed as described in Fig. 1. The total pool of hydrolyzed glucuronides and hydrolyzed sulfates were each assayed in duplicate for 171-hydroxy and 17-oxo-steroid content by the TH assay. The pool of unconjugated steroids was subjected to HPLC and TH assay in duplicate to individually quantify androstenedione, testosterone, 3a-dial, and DHEA. Therefore, a variety of pretreatments (urea, heat, acid, base, methanol) used alone or in combination were tested and the amounts of endogenous conjugated and unconjugated steroids extracted were compared. Urea/heat treatment (Fig. 2) proved best, giving results similar to ethyl acetate
cartridges.
extraction for unconjugated steroids (not shown). Urea (2 M final concentration) plus heat (60°C 30 min) in combination, significantly improved the Sep-Pak cartridge extraction of plasma testosterone, androstenedione, and Sa-androstane-3a,l7fl-dial (3cr-diol) compared to untreated plasma or either
60_
-
DHEA 7
40-
20-
DHEA-G
60.
%-DHEA-s
4020-
20
30
40
50
60
70
60
% METHANOL
Fig, 3. Recoveries of steroids from Sep-Pak cartridges eluted with increasing methanol concentration. DHEA (12.5 nmol), DHEA-G (5 nmol), or [‘HIDHEA-S (5 pmol) were added to 1 ml aliquots of phosphate buffered saline, applied to a cartridge, and eluted with successive 4 ml fractions of 20% to 80% methanol. DHEA was assayed in duplicate by enzymatic oxidation with 3/?-hydroxysteroid dehydrogenase after hydrolysis if necessary (DHEA-G) or by scintillation counting (DHEA-S).
DONNAW. PAYNE
292
treatment alone. Extraction of other conjugated and unconjugated steroids was relatively unaffected (Fig. 2). Differential elution of conjugated and unconjugated steroids from Sep-Pak
Conjugated and unconjugated steroids are retained by Sep-Pak Cl8 cartridges and may be quantitatively eluted with methanol [2,4,5]. Based on the widely different polarities of most conjugated as compared to unconjugated steroids, differential elution of the two groups from the Sep-Pak cartridges was expected. Figure 3 demonstrates the validity of this concept, using selected steroid standards. Total recovery was greater than 90% in each case. On the basis of preliminary experiments (not shown) 47% methanol was chosen to elute conjugated steroids followed by 100% methanol to elute unconjugated steroids. Sample pH affected the elution of DHEA-S. [3H]DHEA-S was extracted by Sep-Pak cartridges from buffered solutions of pH4.7, 7.0, or 8.5 (not shown). Only 22% of recovered radioactivity eluted in 47% methanol (78% in 100% methanol) when the sample pH was 4.7. In contrast, 97% and 99% of [3H]DHEA-S was recovered in the 47% methanol fraction at pH 7.0 and 8.5, respectively. This suggests the importance of adjusting the pH to 7 or greater,
if necessary, in order to elute conjugates in the 47% methanol fraction. Using the procedure shown in Fig. 1, the quantitative extraction of endogenous and added steroids from serum was examined. Figure 4 shows the recoveries and distribution of radioactive steroids, with wide ranging polarities, added to serum and recovered in the four Sep-Pak fractions. Recoveries ranged from 82 to 97% in the expected fraction with all the unconjugated steroids (Nos 1-l 1) eluting primarily in 100% methanol, and DHEA-S (No. 12) in 47% methanol. A small amount (7-12%) of the most polar unconjugated steroids tested (estradiol and cortisol) eluted in the 47% fraction. Other experiments (not shown) with micromolar quantities of unlabelled A4-3-0~0 steroids and androsterone-, DHEA- and testosterone-glucuronides gave similar results. Since recoveries of both tracer (i.e. radiolabelled) and micromolar quantities of steroid were excellent, it seems probable that quantitative extraction is achieved throughout the usable range of steroid concentration. Although one Sep-Pak is adequate for the quantitative extraction of serum steroids, use of two Sep-Pak cartridges (joined by a short piece of tubing) as used in Fig. 4, reduces somewhat the loss of steroid in the initial breakthrough volume (“serum effluent”).
H,O
Serum Effluent
Wash
20--
I23456769lOlll2
I234s6769101112
1 8
47%
Methanol
loo
SO
T
100%
Methanol _.u
z
t
i
12345676910
12346671
STEROID Fig. 4. Recoveries of steroids added to serum and extracted by Rep-Pak cartridges. One of the following radiolabelled steroids (13.000-16.000 cnm) was added to identical I ml-aliquots of pooled normal human, male serum: (1) Progesterone, (2) androstenedione, (3) pregnenolone, (4j androsterone, (5) DHEA, (6) ZOa-dihydroprogesterone, (7) estrone, (8) testosterone, (9) 5a-androstane-3/3,17/I-diol, (10) 17/3-estradiol, (11) cortisol, (12) DHEA-S. The sera were incubated at 37°C for 15 min to equilibrate steroids, treated with urea and heat, and eluted throuh two joined Rep-Pak cartridges as described in Fig. 1. Radioactivity in duplicate 100~1 aliquots was determined by scintillation counting for the initial serum et&rent and subsequent 5 ml fractions of H,O, 47% methanol, and 100% methanol. The results are represented as a percent of radioactivity recovered in each fraction to the total radioactivity recovered in all fractions. The latter value was 93.9% f 6.4 (mean f SD) for the twelve samples shown.
Extraction of serum steroids
293
Because of the unavailability of many steroid conjugated standards, the Sep-Pak cartridge elution distribution of endogenous serum steroid conjugates as well as two available added standards was studied (Fig. 5). Serum containing endogenous steroids as well as added testosterone-G and androsterone-G was applied to a Sep-Pak cartridge and the conjugates eluted in successive l-ml fractions of 47% methanol. These steroids were then hydrolyzed by glucuronidase/sulfatase so that both steroid glucuronides and sulfates would be detected in the subsequent assay. Figure 5 shows that none of the tested serum steroid conjugates were eluted in any appreciable quantity in the unconjugated (100% methanol) fraction. Since enzyme assays capable of detecting any 17/I-, 3a-, or 3fl-hydroxysteroid present were used, it was shown that: Top panel (Fig. 5)-most endogenous 17/I-hydroxysteroid glucuronides and sulfates, as well as added testosterone-G are eluted in the first 2 ml of 47% methanol.
Middle panel. (Fig. 5)-most 3a-hydroxysteroid glucuronides and sulfates, as well as added androsterone-G are eluted with 5 ml of 47% methanol. The approximately 4 nmol/ml of endogenous khydroxysteroid conjugates detected in excess of the 3 nmol/ml of androsterone-G added is much greater than the quantity of Cl9 steroid glucuronides likely to be present in serum (c$ Fig. 2; also Ref. 12). The 3a-hydroxysteroids detected in the 47% methanol fraction are most likely the highly polar bile acids and bile acid conjugates which are 3cr-hydroxy compounds that are detected by this enzyme assay and which are known to be quantitatively extractable by Sep-Pak cartridges [3, 131. Lower panel (Fig. 5F most 3/I-hydroxysteroid conjugates, of which the most abundant in serum is DHEA-S [12], are eluted in the first 2 ml of 47% methanol. In summary, Fig. 5 demonstrates that most endogenous steroid glucuronides and sulfates, as well as bile acids, are eluted in 47% methanol with negligible quantities (O-4%) appearing in the 100% “unconjugated” fraction.
I?@HYDROXYSTEROID
Application of the extraction procedure to the detection of serum C,p steroids
lnmal
TESTOS7ERONE*
nmol
total
G
I
n___l,
I
ii ii
ADOED (3.9
2 E
CONJUGATES
plu*
3
3a-HYDROXYSTEROID
i
1
CONJUGATES
plu* lnmol
% 3 .1E 5 2 .!Z p I ..
ADDED
ANDROSTERONE
t 7nmol
-
total
1
nmol
total)
l
G
8 8
ID-ttYOROXYS7EROIO
:
3
‘.
8 N
2
‘.
( 2.7
CONJUGATES
0% 111
I
2nd
3rd
4th
SUCLSSIVE Iml FRACTIONS 47K MIon
5th OF
2rnl loon YIOH
Fig. 5. Recoveries and elution order of endogenous and added serum steroid conjugates from Sep-Pak cartridges. Normal human, male serum (1 ml) was incubated with 3 nmol each of testosterone-G and androsterone-G and eluted through one Sep-Pak cartridge. Successive l-ml fractions of 47% methanol were dried and the steroids therein hydrolyzed with ghtcuronidase/sulfatase from H. pornaria. The individual groups of 178, 3cr-, and 38hydroxysteroids were sequentially assayed by enzymatic oxidation using highly specific bacterial hydroxysteroid dehydrogenases as previously described [8]. SB 33,2-J
An example of the results obtained using the extraction and analysis procedure shown in Fig. 1 is provided in Fig. 6. Normal adult human, male serum contains significant picomolar quantities of at least four unconjugated and four glucuronidated Cl9 steroids, and nanomolar quantities of at least two sulfated C,,, steroids. Probable identities of these steroids are indicated (arrows) by coelution with steroid standards. The numbers inside the peaks indicate the quantity of steroid present. These results are consistent with those of others using entirely different methodologies [12, 14, 151. In order to check the reproducibility of the extraction scheme for serum, four identical 1 ml aliquots of adult male serum were extracted on Sep-Pak cartridges (not shown) as described above, and the hydrolyzed sulfates, hydrolyzed glucuronides and unconjugated steroids subjected to HPLC. For each group, the DHEA peak was collected and assayed by the TH assay. The values obtained (mean k SD) for DHEA, DHEA-G and DHEA-S were 24.5 f 2.1, 11.5 + 1.3, and 7110 + 300 pmol/ml (coefficient of variation = 8.6, 11.3, 4.2%) respectively. The method described herein extends earlier work on the Sep-Pak cartridge exraction of unconjugated and conjugated steroids [2,4,5] by providing a means to obtain maximum extraction of protein-bound steroids and by introducing the differential and sequential extraction of steroid conjugates and unconjugated steroids. Mainly the extraction of serum steroids was studied. However, since Sep-Pak cartridges have been used to extract steroids from tissues [ 161, milk [6], cell culture [17] and urine [2] it is likely that the differential extraction of conjugated and unconjugated steroids from other biological materials
294
DONNA
W.
PAYNE
19998 (E.Y.A., D.W.P.) from the National Institutes of Health. E.Y.A. is the recipient of N.I.H. Research Career Development Award l-KO4-HD-00697. A preliminary account has been published: D. W. Payne, W. D. Holtzclaw and E. Y. Adashi: Endocrinology Suppl. (1988) 122, 207. REFERENCES 1.
Sjiivall J. and Axelson M.: Newer approaches to the isolation, identification, and quantitation of steroids in biological materials. In Vitamins and Hormones (Edited by P. L. Munson, E. Diczfalusy, J. Glover and R. E. yll$. Academic Press, New York, Vol. 39 (1982) pp.
2. Shackleton C. H. L. and Whitney J. 0.: Use of Sep-Pak cartridges for urinary steroid extraction: evaluation of the method for use prior to gas chromatographic analysis. Clin. chim. Acta 107 (1980) 231-243. 3. Whitney J. 0. and Thaler M. M.: A simple liquid chromatographic method for quantitative extraction of hydrophobic compounds for aqueous solutions. J. liquid Chrotnat. 3 (1980) 545-556.
I
SULFATES
4. Heikkinen R., Fotsis T. and Adlercreutz H.: Reversedphase C,, cartridge for extraction of estrogens from urine and plasma. Clin. Chem. 27 (1981) 1188-1189. 5. Axelson M. and Sahlberg B. L.: Group separation and gas chromatography-mass spectrometry of conjugated steroids in plasma. J. steroid Biochem. 18 (1983) 313-321. 6. Sahlberg B. L. and Axelson M.: Identification and quantitation of free and conjugated steroids in milk from lactating women. J. steroid Biochem. 25 (1986)
I I I
t
379-391. 5
IO
I5
20
25
30
35
40
45
50
ELUTIONVOLUME(mL) Fig. 6. Profiles of unconjugated, glucuronidated, and sulfated C,, steroids in normal male serum. Normal adult human, male serum (1 ml) was urea/heat treated, extracted on Sep-Pak cartridges, hydrolyzed and subjected to HPLC (concave gradient of from 2 to 6% 2-propanol in hexane) as described in Fig. 1. Each 1 ml HPLC fraction was assayed for l’lr!I-hydroxy- or 17-oxosteroid content by TH assay. The arrows indicate elution volumes of standard steroids. The numbers in the peaks represent the quantity of steroid (pmol per ml serum) based on a testosterone standard and calculated by taking into consideration the differential reactivity of some of the steroids compared to testosterone (androsterone, 3cc-diol: 49%; DHEA: 59%; see Ref. 10) in the TH assay.
7 Aso T., Aedo determination terone and immunoassay
A. R. and Cekan S. Z.: Simultaneous of the sulphates of dehydroepiandrospregnenolone in plasma by radiofollowing a rapid solvolysis. J. steroid
Biochem. 8 (1977) 110>1108.
8. Payne D. W., Holtzclaw W. D. and Adashi E. Y.: The steroidogenic characteristics of primary testicular cell cultures from adult hypophysectomized rats: enhanced formation of C,, steroids. Biol. Reprod. 39 (1988) 581-591. 9 Payne D. W. and Talalay P.: Isolation of novel microbial 3a-3P- and 17/I-hydroxysteroid dehydrogenases. J. biof. Chem. 257 (1982) 1301-1308. 10. Payne D. W. and Talalay P.: A one-step enzymatic assay for the measurement of 17p-hydroxy- and 17-0x0steroid profiles in biological samples. J. steroid Biochem. 25 (1986) 403-410.
11. Burton R. M. and Westphal U.: Steroid hormonebinding proteins in blood plasma. Metubolism 21 (1972) also be achieved provided a neutral pH is maintained during elution (see above). The extraction scheme (Fig. 1) provides groups of steroids which have been processed in a similar
could
manner. This should be an advantage in directly comparing steroid profiles. Although the steroids may be assayed by any techniques following extraction, the extraction/hydrolysis scheme (Fig. 1) in
combination with the separating power of HPLC and the specific enzyme assay for Cn steroids provides unique and exhaustive steroid profiles of biological samples. Acknowledgemerrrs-The authors express their gratitude to Dr Paul Talalav (Johns Hopkins University) for his contin-
ued support a& hisccming~advice,and to Theresa Yeshnik for preparation of the manuscript. This work was supported in part by Grants AM074222 (D.W.P., W.D.H.) and HD
253-276.
12. Barden C. W. and Santen R. J.: Androgen assay in blood and other biological fluids. In Steroid Hormones (Edited by R. I. Dorfman). American Elsevier, New York (1975) pp. 6&65. 13. Setchell K. D. R. and Worthington J.: A rapid method _.___ for the quantitative extraction-of bile acid‘s and their conjugates from serum using commercially available reverse-phase octadecylsilane bonded silica cartridges. Clin. chim. Acra 125 (1982) 135-144. 14. Belanger A., Brochu M. and Cliche J.: Plasma levels of steroid alucuronides in ~repubertal, adult and eldertv men. J. ;teroid Biochem: 24 (1986) 1069-1072. ----’ 15. Matteri R. K.. Stanczvk F. Z.. Kaufman F. R., Vii4 A. G., Anderson R. E: and Lobo R. A.: Measu$s and comparison of circulating C9 sulfates and glucuronides in normal and hirsute women and men. Society for Gynecologic Investigation Annual Mtg. (1988) Abstract No. 221, p. 170. 16. Anderson S. H. G. and Sjijvall J.: A method combining solvent and gel extraction for isolation and preliminary
Extraction of serum steroids purification of steroids in tissues. Analyt. Biochem. 134 (1983) 309-312. 17. Ramirez L. C., Millot C. and Maume B. F.: Sample purification using a &-bonded reversed-phase car-
295
tridge for the quantitative analysis of corticosteroids in adrenal cell cultures by high performance liquid chromatography or gas chromatography mass spectrometry. J. Chromat. 229 (1982) 267-281.
0022-4731/89 $3.00 + 0.00 Copyright 0 1989Maxwell Pergamon Macmillan plc
A CONVENIENT, UNIFIED SCHEME FOR THE DIFFERENTIAL EXTRACTION OF CONJUGATED AND UNCONJUGATED SERUM Cl9 STEROIDS ON SEP-PAK C,,-CARTRIDGES DONNA W. PAYNE,‘***W. DAVID HOLTZCLAW’ and ELI Y. ADASHI’ ‘Department of Obstetrics and Gynecology, University of Maryland School of Medicine, Baltimore, Maryland 21201, and 2Department of Pharmacology and Molecular Sciences, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, U.S.A. (Received 24 September 1988; received for publication 3 March 1989)
Snmmary-In order to conveniently and rapidly isolate by group both conjugated and unconjugated serum androgens, a scheme has been devised for their di~erential extraction from ~rnrne~i~y available, disposable octadecylsilane cartridges (Sep-Pak C,r). Using added radioactive steroid standards and detection of endogenous serum steroids by group-specific enzymatic assays, the quantitative recovery of steroid glucuronides and sulfates in the 47% methanol fraction and of unconjugated steroids in the 100% methanol fraction was observed. Maximum recovery of serum protein-bound steroids (e.g. testosterone) was achieved with serum denatured by urea and heat. In order to separate glucuronides from sulfates, sequential hydrolysis of the conjugated fraction (47% methanol) by enzymatic hydrolysis and then organic solvolysis as well as an additional SepPak cartridge extraction step was required. Groups of extracted steroids may be further separated and assayed by any appropriate method(s). An application is given which employs HPLC and an enzymatic assay for 17&hydroxy- and IFoxo-steroids to provide separate profiles of unconjugated, glucuronidat~, and sulfated androgens in human, male serum.
INTRODUCTION Many methods have been described for the extraction
of steroids from biological materials [l]. A convenient, reliable and efficient method for the extraction of steroids from aqueous solutions employs small disposable cartridges packed with silica gel containing an octadecylsilane bonded phase. Methods have heen described using Sep-Pak C,, or similar commercially available materials for the extraction of conjugated and unconjugated androgens, estrogens, corticoids, progestins and bile acids from urine, milk, tissue, cell culture and serum [2-6]. However, none of these methods is designed to use the same cartridge to extract groups of conjugated and unconjugated steroids into separate pools although this would be of use in studies where both groups must be measured. For example, many recent studies have suggested that the measurement of conjugated as well as unconjugated androgens is important in understanding the etiology of hirsutism in women. Such studies would benefit from a convenient, simultaneous method for the extraction of both conjugated and unconjugated androgens. Therefore a simple and unified scheme was devised for the Sep-Pak cartridge extraction of serum C9 steroids. Using this procedure, groups of *Address all correspondence to: D. W. Payne, Dept. of Obstetrics and Gynecology, University of Maryland School of Medicine, 655 W. Baltimore St, Baltimore, MD 21201, U.S.A. 289
sulfates, glucuronides and unconjugated steroids are obtained by nearly identical means. In addition to the simplicity and relative rapidity of this method, comparisons of conjugated and unconjugated steroids in biological samples should be improved because of the elimination of procedural differences during extraction. EXPERIMENTAL Materials
fl-Glucuronidase (Type VIII from E. cd), glucuronidase-sulfatase (Type H-5, from I% pomatia), dehy~~piandrosterone-3-glucuronide (DHEA-G), androsterone-3-glu~uronide (androsterone-G), and testosterone-17-glucuronide (testosterone-G) were obtained from Sigma (St Louis, MO.). Unlabelled unconjugated steroids were obtained from Steraloids (Wilton, N.H.) and [)H]-labelled steroids from Amersham (Arlington Heights, Ill.) or New England Nuclear (Boston, Mass.). Sera were obtained from normal adult, human males and stored at -20°C until used. Extraction
procedure
Serum or plasma (1 ml), with or without urea/heattreatment (1 ml serum plus 1 ml of 4 M urea, 60°C 30 min), was applied to Sep-Pak CIB cartridges (Waters/Millipore, Milford, Mass.) prepared by washing with 2 ml methanol (HPLC grade) and 5 ml deionized, distilbd H,O. The following fractions
290
D~NNAW.
PAYNE
~uREA~HEAT . Tfemd SEdUM I
Ssp-P&,s Dirurd Effluent --I Hz0 Dirard Wash --I 47% hOOH
~HYDAOLYZED
SULFATES[
Fig. 1. Extraction/hydrolysis
were then collected in order (see Fig. 1): serum effluent; 5ml H,O, 3-5 ml 47% methanol (conjugated steroids); 3-5 ml 100% methanol (unconjugated steroids). Serum and all eluents were passed through the cartridge either by syringe or by suction using an &place vacuum manifold (Waters/Millipore, Milford, Mass.). The limiting factor in the time required to perform the extraction(s) is the size of the vacuum manifold since the time required (less than 5 min) to pass all the eluents through the cartridges under vacuum is the same whether one, eight, or more cartridges are present.
procedure (see text for details).
HPLC and steroid ~~y~i~ Dried Sep-Pak fractions containing steroids were dissolved in a small volume of 2-propanol: hexane (5:95) and chromatographed on a Lichrosorb Diol column (EM Reagents, Gibbstown, N.J.) using a concave gradient of 2-propanol : hexane as described earlier [8]. Steroids were detected and quantified using highly specific bacterial hydroxysteroid dehydrogenases [9]. 17@-Hydroxy- plus 17-oxo-steroids were measured by a ~anshydrogenase (TH) assay [lo]. 3a -, 38-, and 17@-Hydroxysteroids were measured by enzymatic oxidation as described previously [8].
Conjugate hydrolysis The 47% methanol fraction was evaporated to dryness (Speed-Vat, Savant Instr., Hicksville, N.Y.), and the residue reconstituted in 1 ml 75 mM potassium phosphate buffer (pH 7). The glucuronides were hydrolyzed at 37°C overnight with I?. coli /?glucuronidase (5 Units) and extracted in the 100% methanol fraction of a second Sep-Pak cartridge (Fig. 1). The 47% methanol fraction of the second cartridge (Fig. 1) contained unhydrolyzed sulfates which were dried, reconstituted in 3 ml ethyl acetate containing 2.5 ~1 concentrated HCl and incubated for 1.5 h at 37°C [7]. In some ex~~rnen~ (not part of standard protocol) where it was convenient to simultaneously hydrolyze sulfates and glucuronides, the conjugates were reconstituted in 1 ml 25 mM sodium acetate buffer (PH 4.5) containing glucuronidase/ sulfatase and hydrolyzed at 37°C for 24 h. Preliminary experiments were performed to determine the optimal conditions of hydrolysis which are described above.
RESULTS AND DISCUSSION
Comparison ofpretreatment md extraction procedures Initially, extraction of endogenous serum and plasma steroids by Sep-Pak cartridges was compared to extraction by a variety of organic solvent extraction procedures. The enzymatic assay used (TH Assay) can detect any steroid containing a 178hydroxy or 17-0~0 function (& steroids), and thus proved very useful in the reliable evaluation of the maximum quantities of endogenous serum steroids which could be extracted by the various methods. Whereas extraction by Sep-Pak cartridges was as good as, or better than, organic solvent extraction for most unconjugated steroids (not shown), this was not so for testosterone and some other androgens. 17/?Hydroxysteroids strongly interact with serum sexbinding globulin and many steroids bind (less stroigly) to serum albumin [ll]. Presumably this prevents their quantitative extraction from Sep-Pak
Extraction of serum steroids
291
Fig. 2. Effect of pretreatment with urea and/or heat on the extraction of endogenous plasma steroids. Four 1 ml aliquots of normal human male plasma were diluted with either 1 ml of phosphate buffered saline pH 7 (C, 60”) or with 1 ml 4 M urea (U, 60” + U) and incubated at either room temperature of 60°C for 30 min as noted. Each sample was then extracted and conjugates hydrolyzed as described in Fig. 1. The total pool of hydrolyzed glucuronides and hydrolyzed sulfates were each assayed in duplicate for 171-hydroxy and 17-oxo-steroid content by the TH assay. The pool of unconjugated steroids was subjected to HPLC and TH assay in duplicate to individually quantify androstenedione, testosterone, 3a-dial, and DHEA. Therefore, a variety of pretreatments (urea, heat, acid, base, methanol) used alone or in combination were tested and the amounts of endogenous conjugated and unconjugated steroids extracted were compared. Urea/heat treatment (Fig. 2) proved best, giving results similar to ethyl acetate
cartridges.
extraction for unconjugated steroids (not shown). Urea (2 M final concentration) plus heat (60°C 30 min) in combination, significantly improved the Sep-Pak cartridge extraction of plasma testosterone, androstenedione, and Sa-androstane-3a,l7fl-dial (3cr-diol) compared to untreated plasma or either
60_
-
DHEA 7
40-
20-
DHEA-G
60.
%-DHEA-s
4020-
20
30
40
50
60
70
60
% METHANOL
Fig, 3. Recoveries of steroids from Sep-Pak cartridges eluted with increasing methanol concentration. DHEA (12.5 nmol), DHEA-G (5 nmol), or [‘HIDHEA-S (5 pmol) were added to 1 ml aliquots of phosphate buffered saline, applied to a cartridge, and eluted with successive 4 ml fractions of 20% to 80% methanol. DHEA was assayed in duplicate by enzymatic oxidation with 3/?-hydroxysteroid dehydrogenase after hydrolysis if necessary (DHEA-G) or by scintillation counting (DHEA-S).
DONNAW. PAYNE
292
treatment alone. Extraction of other conjugated and unconjugated steroids was relatively unaffected (Fig. 2). Differential elution of conjugated and unconjugated steroids from Sep-Pak
Conjugated and unconjugated steroids are retained by Sep-Pak Cl8 cartridges and may be quantitatively eluted with methanol [2,4,5]. Based on the widely different polarities of most conjugated as compared to unconjugated steroids, differential elution of the two groups from the Sep-Pak cartridges was expected. Figure 3 demonstrates the validity of this concept, using selected steroid standards. Total recovery was greater than 90% in each case. On the basis of preliminary experiments (not shown) 47% methanol was chosen to elute conjugated steroids followed by 100% methanol to elute unconjugated steroids. Sample pH affected the elution of DHEA-S. [3H]DHEA-S was extracted by Sep-Pak cartridges from buffered solutions of pH4.7, 7.0, or 8.5 (not shown). Only 22% of recovered radioactivity eluted in 47% methanol (78% in 100% methanol) when the sample pH was 4.7. In contrast, 97% and 99% of [3H]DHEA-S was recovered in the 47% methanol fraction at pH 7.0 and 8.5, respectively. This suggests the importance of adjusting the pH to 7 or greater,
if necessary, in order to elute conjugates in the 47% methanol fraction. Using the procedure shown in Fig. 1, the quantitative extraction of endogenous and added steroids from serum was examined. Figure 4 shows the recoveries and distribution of radioactive steroids, with wide ranging polarities, added to serum and recovered in the four Sep-Pak fractions. Recoveries ranged from 82 to 97% in the expected fraction with all the unconjugated steroids (Nos 1-l 1) eluting primarily in 100% methanol, and DHEA-S (No. 12) in 47% methanol. A small amount (7-12%) of the most polar unconjugated steroids tested (estradiol and cortisol) eluted in the 47% fraction. Other experiments (not shown) with micromolar quantities of unlabelled A4-3-0~0 steroids and androsterone-, DHEA- and testosterone-glucuronides gave similar results. Since recoveries of both tracer (i.e. radiolabelled) and micromolar quantities of steroid were excellent, it seems probable that quantitative extraction is achieved throughout the usable range of steroid concentration. Although one Sep-Pak is adequate for the quantitative extraction of serum steroids, use of two Sep-Pak cartridges (joined by a short piece of tubing) as used in Fig. 4, reduces somewhat the loss of steroid in the initial breakthrough volume (“serum effluent”).
H,O
Serum Effluent
Wash
20--
I23456769lOlll2
I234s6769101112
1 8
47%
Methanol
loo
SO
T
100%
Methanol _.u
z
t
i
12345676910
12346671
STEROID Fig. 4. Recoveries of steroids added to serum and extracted by Rep-Pak cartridges. One of the following radiolabelled steroids (13.000-16.000 cnm) was added to identical I ml-aliquots of pooled normal human, male serum: (1) Progesterone, (2) androstenedione, (3) pregnenolone, (4j androsterone, (5) DHEA, (6) ZOa-dihydroprogesterone, (7) estrone, (8) testosterone, (9) 5a-androstane-3/3,17/I-diol, (10) 17/3-estradiol, (11) cortisol, (12) DHEA-S. The sera were incubated at 37°C for 15 min to equilibrate steroids, treated with urea and heat, and eluted throuh two joined Rep-Pak cartridges as described in Fig. 1. Radioactivity in duplicate 100~1 aliquots was determined by scintillation counting for the initial serum et&rent and subsequent 5 ml fractions of H,O, 47% methanol, and 100% methanol. The results are represented as a percent of radioactivity recovered in each fraction to the total radioactivity recovered in all fractions. The latter value was 93.9% f 6.4 (mean f SD) for the twelve samples shown.
Extraction of serum steroids
293
Because of the unavailability of many steroid conjugated standards, the Sep-Pak cartridge elution distribution of endogenous serum steroid conjugates as well as two available added standards was studied (Fig. 5). Serum containing endogenous steroids as well as added testosterone-G and androsterone-G was applied to a Sep-Pak cartridge and the conjugates eluted in successive l-ml fractions of 47% methanol. These steroids were then hydrolyzed by glucuronidase/sulfatase so that both steroid glucuronides and sulfates would be detected in the subsequent assay. Figure 5 shows that none of the tested serum steroid conjugates were eluted in any appreciable quantity in the unconjugated (100% methanol) fraction. Since enzyme assays capable of detecting any 17/I-, 3a-, or 3fl-hydroxysteroid present were used, it was shown that: Top panel (Fig. 5)-most endogenous 17/I-hydroxysteroid glucuronides and sulfates, as well as added testosterone-G are eluted in the first 2 ml of 47% methanol.
Middle panel. (Fig. 5)-most 3a-hydroxysteroid glucuronides and sulfates, as well as added androsterone-G are eluted with 5 ml of 47% methanol. The approximately 4 nmol/ml of endogenous khydroxysteroid conjugates detected in excess of the 3 nmol/ml of androsterone-G added is much greater than the quantity of Cl9 steroid glucuronides likely to be present in serum (c$ Fig. 2; also Ref. 12). The 3a-hydroxysteroids detected in the 47% methanol fraction are most likely the highly polar bile acids and bile acid conjugates which are 3cr-hydroxy compounds that are detected by this enzyme assay and which are known to be quantitatively extractable by Sep-Pak cartridges [3, 131. Lower panel (Fig. 5F most 3/I-hydroxysteroid conjugates, of which the most abundant in serum is DHEA-S [12], are eluted in the first 2 ml of 47% methanol. In summary, Fig. 5 demonstrates that most endogenous steroid glucuronides and sulfates, as well as bile acids, are eluted in 47% methanol with negligible quantities (O-4%) appearing in the 100% “unconjugated” fraction.
I?@HYDROXYSTEROID
Application of the extraction procedure to the detection of serum C,p steroids
lnmal
TESTOS7ERONE*
nmol
total
G
I
n___l,
I
ii ii
ADOED (3.9
2 E
CONJUGATES
plu*
3
3a-HYDROXYSTEROID
i
1
CONJUGATES
plu* lnmol
% 3 .1E 5 2 .!Z p I ..
ADDED
ANDROSTERONE
t 7nmol
-
total
1
nmol
total)
l
G
8 8
ID-ttYOROXYS7EROIO
:
3
‘.
8 N
2
‘.
( 2.7
CONJUGATES
0% 111
I
2nd
3rd
4th
SUCLSSIVE Iml FRACTIONS 47K MIon
5th OF
2rnl loon YIOH
Fig. 5. Recoveries and elution order of endogenous and added serum steroid conjugates from Sep-Pak cartridges. Normal human, male serum (1 ml) was incubated with 3 nmol each of testosterone-G and androsterone-G and eluted through one Sep-Pak cartridge. Successive l-ml fractions of 47% methanol were dried and the steroids therein hydrolyzed with ghtcuronidase/sulfatase from H. pornaria. The individual groups of 178, 3cr-, and 38hydroxysteroids were sequentially assayed by enzymatic oxidation using highly specific bacterial hydroxysteroid dehydrogenases as previously described [8]. SB 33,2-J
An example of the results obtained using the extraction and analysis procedure shown in Fig. 1 is provided in Fig. 6. Normal adult human, male serum contains significant picomolar quantities of at least four unconjugated and four glucuronidated Cl9 steroids, and nanomolar quantities of at least two sulfated C,,, steroids. Probable identities of these steroids are indicated (arrows) by coelution with steroid standards. The numbers inside the peaks indicate the quantity of steroid present. These results are consistent with those of others using entirely different methodologies [12, 14, 151. In order to check the reproducibility of the extraction scheme for serum, four identical 1 ml aliquots of adult male serum were extracted on Sep-Pak cartridges (not shown) as described above, and the hydrolyzed sulfates, hydrolyzed glucuronides and unconjugated steroids subjected to HPLC. For each group, the DHEA peak was collected and assayed by the TH assay. The values obtained (mean k SD) for DHEA, DHEA-G and DHEA-S were 24.5 f 2.1, 11.5 + 1.3, and 7110 + 300 pmol/ml (coefficient of variation = 8.6, 11.3, 4.2%) respectively. The method described herein extends earlier work on the Sep-Pak cartridge exraction of unconjugated and conjugated steroids [2,4,5] by providing a means to obtain maximum extraction of protein-bound steroids and by introducing the differential and sequential extraction of steroid conjugates and unconjugated steroids. Mainly the extraction of serum steroids was studied. However, since Sep-Pak cartridges have been used to extract steroids from tissues [ 161, milk [6], cell culture [17] and urine [2] it is likely that the differential extraction of conjugated and unconjugated steroids from other biological materials
294
DONNA
W.
PAYNE
19998 (E.Y.A., D.W.P.) from the National Institutes of Health. E.Y.A. is the recipient of N.I.H. Research Career Development Award l-KO4-HD-00697. A preliminary account has been published: D. W. Payne, W. D. Holtzclaw and E. Y. Adashi: Endocrinology Suppl. (1988) 122, 207. REFERENCES 1.
Sjiivall J. and Axelson M.: Newer approaches to the isolation, identification, and quantitation of steroids in biological materials. In Vitamins and Hormones (Edited by P. L. Munson, E. Diczfalusy, J. Glover and R. E. yll$. Academic Press, New York, Vol. 39 (1982) pp.
2. Shackleton C. H. L. and Whitney J. 0.: Use of Sep-Pak cartridges for urinary steroid extraction: evaluation of the method for use prior to gas chromatographic analysis. Clin. chim. Acta 107 (1980) 231-243. 3. Whitney J. 0. and Thaler M. M.: A simple liquid chromatographic method for quantitative extraction of hydrophobic compounds for aqueous solutions. J. liquid Chrotnat. 3 (1980) 545-556.
I
SULFATES
4. Heikkinen R., Fotsis T. and Adlercreutz H.: Reversedphase C,, cartridge for extraction of estrogens from urine and plasma. Clin. Chem. 27 (1981) 1188-1189. 5. Axelson M. and Sahlberg B. L.: Group separation and gas chromatography-mass spectrometry of conjugated steroids in plasma. J. steroid Biochem. 18 (1983) 313-321. 6. Sahlberg B. L. and Axelson M.: Identification and quantitation of free and conjugated steroids in milk from lactating women. J. steroid Biochem. 25 (1986)
I I I
t
379-391. 5
IO
I5
20
25
30
35
40
45
50
ELUTIONVOLUME(mL) Fig. 6. Profiles of unconjugated, glucuronidated, and sulfated C,, steroids in normal male serum. Normal adult human, male serum (1 ml) was urea/heat treated, extracted on Sep-Pak cartridges, hydrolyzed and subjected to HPLC (concave gradient of from 2 to 6% 2-propanol in hexane) as described in Fig. 1. Each 1 ml HPLC fraction was assayed for l’lr!I-hydroxy- or 17-oxosteroid content by TH assay. The arrows indicate elution volumes of standard steroids. The numbers in the peaks represent the quantity of steroid (pmol per ml serum) based on a testosterone standard and calculated by taking into consideration the differential reactivity of some of the steroids compared to testosterone (androsterone, 3cc-diol: 49%; DHEA: 59%; see Ref. 10) in the TH assay.
7 Aso T., Aedo determination terone and immunoassay
A. R. and Cekan S. Z.: Simultaneous of the sulphates of dehydroepiandrospregnenolone in plasma by radiofollowing a rapid solvolysis. J. steroid
Biochem. 8 (1977) 110>1108.
8. Payne D. W., Holtzclaw W. D. and Adashi E. Y.: The steroidogenic characteristics of primary testicular cell cultures from adult hypophysectomized rats: enhanced formation of C,, steroids. Biol. Reprod. 39 (1988) 581-591. 9 Payne D. W. and Talalay P.: Isolation of novel microbial 3a-3P- and 17/I-hydroxysteroid dehydrogenases. J. biof. Chem. 257 (1982) 1301-1308. 10. Payne D. W. and Talalay P.: A one-step enzymatic assay for the measurement of 17p-hydroxy- and 17-0x0steroid profiles in biological samples. J. steroid Biochem. 25 (1986) 403-410.
11. Burton R. M. and Westphal U.: Steroid hormonebinding proteins in blood plasma. Metubolism 21 (1972) also be achieved provided a neutral pH is maintained during elution (see above). The extraction scheme (Fig. 1) provides groups of steroids which have been processed in a similar
could
manner. This should be an advantage in directly comparing steroid profiles. Although the steroids may be assayed by any techniques following extraction, the extraction/hydrolysis scheme (Fig. 1) in
combination with the separating power of HPLC and the specific enzyme assay for Cn steroids provides unique and exhaustive steroid profiles of biological samples. Acknowledgemerrrs-The authors express their gratitude to Dr Paul Talalav (Johns Hopkins University) for his contin-
ued support a& hisccming~advice,and to Theresa Yeshnik for preparation of the manuscript. This work was supported in part by Grants AM074222 (D.W.P., W.D.H.) and HD
253-276.
12. Barden C. W. and Santen R. J.: Androgen assay in blood and other biological fluids. In Steroid Hormones (Edited by R. I. Dorfman). American Elsevier, New York (1975) pp. 6&65. 13. Setchell K. D. R. and Worthington J.: A rapid method _.___ for the quantitative extraction-of bile acid‘s and their conjugates from serum using commercially available reverse-phase octadecylsilane bonded silica cartridges. Clin. chim. Acra 125 (1982) 135-144. 14. Belanger A., Brochu M. and Cliche J.: Plasma levels of steroid alucuronides in ~repubertal, adult and eldertv men. J. ;teroid Biochem: 24 (1986) 1069-1072. ----’ 15. Matteri R. K.. Stanczvk F. Z.. Kaufman F. R., Vii4 A. G., Anderson R. E: and Lobo R. A.: Measu$s and comparison of circulating C9 sulfates and glucuronides in normal and hirsute women and men. Society for Gynecologic Investigation Annual Mtg. (1988) Abstract No. 221, p. 170. 16. Anderson S. H. G. and Sjijvall J.: A method combining solvent and gel extraction for isolation and preliminary
Extraction of serum steroids purification of steroids in tissues. Analyt. Biochem. 134 (1983) 309-312. 17. Ramirez L. C., Millot C. and Maume B. F.: Sample purification using a &-bonded reversed-phase car-
295
tridge for the quantitative analysis of corticosteroids in adrenal cell cultures by high performance liquid chromatography or gas chromatography mass spectrometry. J. Chromat. 229 (1982) 267-281.