| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Lipid Research, Vol. 45, 2287-2302, December 2004
Copyright © 2004 by American Society for Biochemistry and Molecular Biology
* Institut National de la Santé et de la Recherche Médicale U488, 94276 Kremlin-Bicêtre, France
Department of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77, Stockholm, Sweden
Published, JLR Papers in Press, September 1, 2004. DOI 10.1194/jlr.M400244-JLR200
1 To whom correspondence should be addressed. e-mail: liere{at}kb.inserm.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
1 nmol/g) are much higher than those of their unconjugated counterparts. They are distinct from fatty acid esters, and preliminary data do not favor structures such as sulfolipids or sterol peroxides. Noncovalent interactions between steroids and proteolipidic elements, such as lipoproteins, could account for some experimental data. Given their abundance in rodent brain, the structural characterization and biological functions of L-PREG and L-DHEA in the central nervous system merit considerable attention.
Abbreviations: apoD, apolipoprotein D; CNS, central nervous system; DHEA, dehydroepiandrosterone; DHEA-HFB, dehydroepiandrosterone heptafluorobutyrate; DHEA-S, dehydroepiandrosterone sulfate; ESI, electrospray ionization; EtOH, ethanol; HFBA, heptafluorobutyric anhydride; LC-MS, liquid chromatography-mass spectrometry; L-DHEA, lipoidal dehydroepiandrosterone; L-PREG, lipoidal pregnenolone; MeOH, methanol; MSTFA, N-methyl-N-trimethylsilyltrifluoroacetamide; PREG, pregnenolone; PREG-HFB, pregnenolone heptafluorobutyrate; PREG-3-HFB-20-MO, pregnenolone-3-heptafluorobutyrate-20-methoxime; PREG-20-MO, pregnenolone-20-methoxime; PREG-S, pregnenolone sulfate; PREG-TMS, pregnenolone trimethylsilyl ether; SIM, selected ion monitoring; SPE, solid-phase extraction; TEA, triethylamine
Supplementary key words pregnenolone sulfate • dehydroepiandrosterone sulfate • solid-phase extraction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
-hydroxy-5-pregnan-20-one).
Neurosteroids are synthesized by glial cells and neurons from cholesterol or from blood-borne precursors (4). In particular, PREG-S and DHEA-S are known to be neuroactive, and their major effects are the modulation of several neuronal membrane receptors, such as 
-aminobutyric acid (5, 6), N-methyl-D-aspartate (7), and 
receptors, by affecting neuronal excitability and behavior (8, 9). Notably, a physiologic function of PREG-S was suggested by the positive correlation between PREG-S levels in the hippocampus of aged rats and their spatial memory performance (10). However, some studies do not favor the concept that PREG-S and DHEA-S are actually neurosteroids. Indeed, contradictory results have been obtained concerning the presence and activity of the hydroxysteroid sulfotransferase, implied in the sulfation of free steroid. In some studies, the enzyme was not detected in human (11, 12), mouse (13), or rat (12) brain, although some groups have found the enzyme in rat (14, 15) and in frog (16) brain. Other studies have reported a very low hydroxysteroid sulfotransferase activity in rat (17) and human (18) brain.
These results are in apparent contradiction to the reported high levels of PREG-S and DHEA-S measured in mammalian brain by RIA (19–22) and the higher concentrations of PREG-S and DHEA-S in rodent brain than in blood (20–22). Moreover, PREG-S and DHEA-S were found to be still present in the brains of castrated and adrenalectomized male rats, whereas they disappeared from plasma. These latter results, together with the low permeability of the blood-brain barrier for sulfated steroids (18, 23), are consistent with their in situ synthesis. In agreement with the concentration range for PREG-S and DHEA-S in these reports, their average levels in rat brain measured by GC-MS were 8.3 ± 0.80 ng/g and 2.5 ± 0.27 ng/g, respectively, and were higher than the levels of the unconjugated steroids (1). However, several experimental oddities in terms of reproducibility and linearity were observed for PREG-S and DHEA-S measurements. We also became aware of the fact that the chemical identity of the sulfated steroids had never been established. The analytical protocol was thus modified to improve the reliability of PREG-S and DHEA-S quantification. Our results demonstrate that the sulfated forms of PREG and DHEA are not present in rat and mouse brain. The previous detection of steroid sulfates resulted from the indirect method of analysis used in earlier studies that was based on the measure of unconjugated PREG and DHEA after hydrolysis of fractions supposed to contain the endogenous sulfated conjugates. We show here that large amounts of PREG and DHEA are derivatized from lipoidal complexes by acylation derivatization reaction in the presence of triethylamine (TEA). The presence of these lipoidal derivatives of PREG and DHEA in rodent brain could explain the variations observed for PREG-S and DHEA-S measurements with direct (24–27) and indirect (1, 19–22) methods of analysis. The chemical identities of these lipoidal derivatives still need to be elucidated.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
Extraction
The standard extraction procedure
The extraction procedure is roughly similar to that reported in our previous paper (1). The entire brains (2 g) of 2 month old male Sprague-Dawley rats [Centre d'Elevage R. Janvier, (CERJ), Le Genest St-Isle] were removed and weighed. Rat blood was sampled in heparinized polypropylene Falcon tubes through a funnel containing a heparin-treated gauze filter. Blood was centrifuged at 3,000 g for 10 min at 4°C, and the plasma was collected. Steroids were extracted from tissues by adding 10 volumes (w/v) of methanol (MeOH), and the samples were sonicated and left overnight at room temperature and centrifuged at 3,000 g for 5 min. The supernatants were collected, and the residues were extracted with 10 volumes (w/v) of MeOH/CHCl3 (1:1, v/v) and centrifuged again. The proportions of solvent in all solvent mixtures are given as volume proportions throughout the article. The supernatants were combined, and aliquots corresponding to 100 mg of rat brain, unless otherwise stated, were analyzed. The same protocol was used for analysis of 1 ml of plasma. One hundred milligram aliquots of the anterior brain and the entire hippocampus (144.5 ± 13.0 mg; n = 5), liver (340.8 ± 77.4 mg; n = 5), and adrenal glands (38.9 ± 6.2 mg; n = 5) were also used in this study. Two nanograms of [2H4]PREG-S in MeOH was added as an internal standard to aliquots of brain and plasma and to the total extracts of hippocampus, liver, and adrenals for quantification of PREG-S and DHEA-S. The extracts were dried, taken up in 1 ml of MeOH, and sonicated. This protocol was also applied in the analysis of PREG and DHEA conjugates in 150 mg of whole brain from 2 month old Swiss male mice. One hundred milligram aliquots from the cerebellum of a 91 year old man and from the frontal cortex of an 88 year old woman were also analyzed. These patients were hospitalized and died in a geriatric unit (Service de Gérontologie, 3 Hôpital Emile Roux, Limeil-Brévannes, France), and the postmortem delay before dissection was less than 24 h. These brain tissues were obtained from a previous study concerning neurosteroid quantification in human brain in Alzheimer's and nondemented patients (2). These investigations were in agreement with French legislation (the French bioethic Huriet law from 1994) requiring the explicit consent of patients or the family for medical research purposes. A blood sample from a healthy adult man (50 years old) was analyzed as a positive control for validation of the steroid sulfate methodology.
Extraction procedure for the comparative study
A protocol derived from the techniques used in our laboratory for RIA analysis in the period 1980–1995 (21, 22, 28, 29) was used for comparison with the standard method with regard to the detection of PREG-S and DHEA-S in rat brain. The halves of rat anterior brain were dissected and weighed (650 mg). One-half was extracted by the standard method, as described above, and the other was homogenized in 1 ml of PBS buffer (10 mM, pH 7.4) containing 0.8% NaCl in a Teflon/glass homogenizer at 4°C. [2H4]PREG-S and 2 volumes of ethyl acetate-isooctane (1:1) were added. The mixture was left overnight at room temperature and then centrifuged at 3,000 g for 5 min. The organic phase was removed and the extraction was repeated twice. The aqueous phase, containing steroid sulfates, was taken to dryness, and the residue was dissolved in 1 ml of MeOH.
Incubation experiments
These experiments were performed with a homogenate of rat brain in a Tris hypotonic buffer (pH 7.4). Aliquots of 300 mg of homogenate were incubated with [3H]PREG for 1, 4, and 48 h at 37°C. Control experiments were carried out by incubating [3H]PREG with a boiled homogenate for 4 h and in the absence of rat brain homogenate. Analogous experiments were carried out by adding tritiated PREG and PREG-S to a methanolic extract of rat brain for 0 and 24 h.
SPE methodology
The first purification and fractionation step was achieved by SPE on C18 columns (500 mg, 6 ml; International Sorbent Technology, Mid Glamorgan, UK). The protocol has been changed relative to that described in our previous paper (1) for several reasons explained below.
Original SPE with addition of a water washing step This procedure was the same as that described in our recent paper (1) except that a washing step with 5 ml of water was introduced before the elution of sulfated and free steroids with 5 ml each of MeOH/water (1:1) and MeOH/water (9:1), respectively.
SPE with sequential recycling and stepwise elution In this system, the C18 cartridge was conditioned with 5 ml of MeOH, 5 ml of water, and 5 ml of MeOH. The brain extract in MeOH was applied to the cartridge, which was then washed with 5 ml of MeOH. The flow-through and the washing were combined and diluted with water to MeOH/water (8:2) and again applied to the same cartridge. Then, the flow-through was diluted to MeOH/water (6:4) and reapplied to the cartridge, and the same process was repeated with MeOH/water (4:6), MeOH/water (2:8), and finally water. This sequence of adsorptions was defined as a sample recycling by SPE. The second part of the procedure was the stepwise recovery of steroids by eluting the C18 cartridge with increasing concentrations of MeOH in water from 0% to 100% in 5% increments.
SPE with simplified recycling and fractionation A simplified recycling/elution SPE protocol was established for quantifying sulfated, unconjugated, and lipoidal PREG and DHEA (Fig. 1) . Samples were dissolved in 1 ml of MeOH, applied to the cartridge, and 5 ml of MeOH/water (85:15) was added. The flow-through was collected and dried. The column was reconditioned with 5 ml of water, and samples were dissolved in MeOH/water (2:8) and applied to the cartridge. After a preliminary wash with 5 ml of water, sulfated, free, and lipoidal steroids were eluted with 5 ml of MeOH/water (1:1), 5 ml of MeOH/water (9:1), and 5 ml of MeOH/CHCl3 (1:1), respectively. In this paper, lipoidal PREG and DHEA (L-PREG and L-DHEA, respectively) designate derivatives of unknown structure eluted in the MeOH/CHCl3 (1:1) fraction. Fatty acid esters of PREG and DHEA were also eluted in the same fraction. The recovery of L-PREG and L-DHEA was increased by additional elution with 5 ml of hexane.
|
SPE for lability study For this purpose, the lipoidal fraction obtained after SPE fractionation was dissolved in different solvents for 30 min or 24 h at different temperatures and in the presence or absence of light. Then, the samples were dried and taken up in 1 ml of MeOH for a second C18 SPE fractionation. The sulfated, unconjugated, and lipoidal fractions were collected as described above.
Derivatization reaction
As shown in Fig. 1, sulfated, unconjugated, and lipoidal steroids were derivatized separately. Simultaneous hydrolysis/derivatization of PREG-S and DHEA-S with HFBA was investigated. Touchstone and Dobbins (31) reported that some sulfated steroids can react with HFBA for their conversion into heptafluorobutyrates suitable for GC-MS analysis, and Murray and Baillie (32) showed that this single-step reaction was quantitative only in the case of aromatic sulfated steroids and steroids with a 3ß-sulfoxy-
5 structure, such as PREG-S and DHEA-S. The potential usefulness of this reaction was of interest because of its simplicity and the selectivity of the hydrolysis/acylation reaction. The reaction mechanism proposed by Murray and Baillie (32) is schematized in Fig. 2
.
|
Lipoidal steroids The MeOH/CHCl3 (1:1) fraction from the SPE containing lipoidal steroids was derivatized with HFBA or TEA/HFBA. The reaction with HFBA was performed as described above for sulfated and free steroids except that the temperature was kept at 70°C. Concerning the reaction with TEA/HFBA, the residue was dissolved in 100 µl of acetone, 20 µl of TEA and 20 µl of HFBA were added, and the mixture was heated for 30 min at 70°C. This reaction has been optimized according to several parameters, such as the relative amount of reagent, reaction duration, and temperature. A liquid-liquid partition step between water and hexane was performed. The organic phase was removed, and the extraction was repeated. The combined organic extracts were dried under nitrogen at 70°C.
The heptafluorobutyrates obtained in both the HFBA and TEA/HFBA reactions were purified by C18 SPE with the CH3CN/water (1:1) system described above.
Two other derivatization reagents were used for analyzing L-PREG and L-DHEA by GC-MS. Fifty microliters of MSTFA or 100 µl of methoxyamine hydrochloride (2% in pyridine) was added to the lipoidal fraction at 70°C for 30 min. The reaction product with MSTFA [i.e., trimethylsilyl ether of PREG (PREG-TMS)] was purified with SPE using the CH3CN/water (1:1) system, whereas the PREG-20-methoxime (PREG-20-MO) synthesized with methoxyamine hydrochloride was submitted to SPE with simplified recycling and fractionation. In the latter case, MeOH/water (9:1) and MeOH/CHCl3 (1:1) fractions were collected, and 20 µl of HFBA and 100 µl of acetone were added to these fractions for 30 min at 70°C to form the 3-heptafluorobutyrate-20-methoxime of PREG (PREG-3-HFB-20-MO).
In some experiments, PREG palmitate, DHEA stearate (10 ng each), and 100 ng of PREG tosylate were added to the SPE lipoidal fraction from 100 mg of rat brain extract to test the acylation power of the TEA/HFBA derivatizing reagents toward different PREG conjugates.
Saponification
Saponification was performed on the SPE MeOH/CHCl3 (1:1) fraction. This was dried and dissolved in 1 ml of a solution of 40% (w/v) KOH in ethanol (EtOH)/water (95:5) under nitrogen. The vial was kept at 80°C for 1 h and at room temperature for 12 h. Five milliliters of water was added, and the solution was neutralized with hydrochloric acid. Free steroids were extracted with 5 ml of ethyl acetate. The extraction was repeated twice, and the organic phases were combined and dried. The samples were submitted to SPE fractionation to analyze free and lipoidal steroids as described before.
GC-MS analysis
GC-MS analysis was carried out using an Automass Solo mass spectrometer (Thermofinnigan, Les Ulis, France) interfaced with a TraceGC (Carlo Erba, Milan, Italy) gas chromatograph. Samples were injected with AS 2000 autosampler (Carlo Erba) to the injection chamber maintained at 250°C in the splitless mode. The analytical capillary column was BPX35 (35% phenyl/65% dimethyl polysiloxane; SGE, Victoria, Australia), 30 m long with an inner diameter of 0.25 mm and a 0.25 µm film thickness. The initial oven temperature was kept at 50°C for 1 min and was increased to 175°C at 30°C/min and to 320°C at 10°C/min. The transfer line and ionization chamber temperatures were maintained at 300°C and 180°C, respectively. The flow rate of helium carrier gas was kept constant at 0.8 ml/min all along the temperature profile. The mass spectrometer was operated in the electron-impact mode with an ionization energy and an emission current of 70 eV and 1,000 µA, respectively.
Identification of PREG heptafluorobutyrate (PREG-HFB), DHEA heptafluorobutyrate (DHEA-HFB), and tetradeuterated PREG heptafluorobutyrate ([2H4]PREG-HFB) was achieved in the full-scan mode (m/z range 50–550), and quantification was carried out in the selected ion monitoring (SIM) mode in a time-programmed manner. The diagnostic ions for the quantification of steroids in the sulfate fraction were m/z 270 for DHEA-HFB (M+ 484), m/z 298 for PREG-HFB (M+ 512), and m/z 302 for [2H4]PREG-HFB (M+ 516). Unconjugated and lipoidal PREG and DHEA were detected in SIM mode, and PREG-HFB and DHEA-HFB were identified by their GC retention times and the ratio between two specific diagnostic ions. The latter are given in Table 1 for all of the analyzed steroid derivatives in the SPE lipoidal fraction from rat brain.
|
Experiments with tritiated tracers and radioactivity analysis
Tritiated PREG, PREG-S, and DHEA-S were used for recovery calculations in extraction processes and the SPE methodology with simplified recycling and fractionation. [3H]PREG palmitate was used to validate the saponification reaction.
Incubations of MeOH extracts of rat brain were made with [3H]PREG and [3H]PREG-S, and incubations with rat brain homogenates in buffer were carried out with [3H]PREG.
Dried radioactive samples were dissolved in 5 ml of Picofluor 15 scintillation liquid and counted in a Packard Tricarb liquid scintillation spectrometer model 4660 equipped with quench correction (Packard Instruments, Downers Grove, IL).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
It is notable that steroid sulfates so far had been assayed by indirect methods of analysis. PREG-S and DHEA-S were isolated from brain extracts by chromatography on Sephadex LH-20 (20) or by liquid-liquid extraction (21, 22, 29) and then submitted to acidic solvolysis before RIA using anti-PREG and anti-DHEA antibodies.
Three years ago, we published a paper on the validation of an analytical protocol to quantify unconjugated steroids such as PREG, DHEA, PROG, allopregnanolone, and sulfated steroids such as PREG-S and DHEA-S in the same brain tissue with high selectivity and sensitivity by GC-MS. This protocol allowed the simultaneous identification and quantification of several steroids without the need for specific antibodies or for pooling of brain areas from several animals, as is usually required with RIA methods because of their lack of sensitivity. Hence, GC-MS is the most sensitive detector and a powerful tool to precisely describe the distribution of neurosteroids, present in trace amounts in rodent brain, to give a better understanding of their physiological functions. An important disadvantage of GC-MS is its inability to detect intact steroid conjugates. For example, the sulfate moiety of PREG-S and DHEA-S is cleaved by pyrolysis as they are introduced in the injection chamber at 250°C. For this reason, the sulfated steroids must be hydrolyzed and derivatized before their analysis by GC-MS. Some groups have evaluated the potential of direct detection methods such as high performance liquid chromatography-electrospray ionization-tandem mass spectrometry (HPLC-ESI-MS/MS) (35) and capillary column HPLC-nanoelectrospray ionization-MS/MS (27, 33, 36) for quantification of sulfated steroids in rat brain. Furthermore, ELISAs were developed to measure DHEA-S (24) and PREG-S (25, 26) levels in rat brain by using anti-DHEA-S and anti-PREG-S antibodies, respectively. All of these direct methods represent reference methods to validate the quantitative results obtained by indirect detection methods such as GC-MS to ascertain that solvolysis/hydrolysis of fractions containing steroid sulfates is specific.
Effect of initial water washing on SPE separation
In our previous study by GC-MS (1), a critical step in the analysis of steroid sulfates was their isolation from other less polar steroids by SPE. The fraction containing steroid sulfates was then solvolyzed with ethyl acetate containing H2SO4 and repurified by SPE before derivatization of the released free steroids with HFBA. However, it was noted that the coefficients of variation in the measurements of PREG-S and DHEA-S were relatively high (21% and 23%, respectively). Indeed, GC-MS analysis indicated many interfering compounds that could explain these results. The peak area calculations and detection levels were obscured by this "chemical noise." Furthermore, we observed that the solvolysis reaction was incomplete as the mass of brain tissue was increased, particularly when it exceeded 50 mg. When [3H]PREG-S was added to a methanolic extract of 100 mg of rat brain and sulfated steroids were isolated by SPE, only 17% of the radioactivity was recovered in the organic phase after the solvolysis step (ethyl acetate extraction) and 83% remained in the aqueous phase. This latter fraction contained unhydrolyzed [3H]PREG-S. This indicated that the SPE fraction, containing endogenous or exogenous PREG-S, also contained brain constituents that inhibited the solvolysis. For these reasons, a water wash was added to the SPE methodology between the brain sample deposition onto the cartridge and the elution of sulfated steroids (35). Under these conditions, the SPE recovery of [3H]PREG-S was excellent (90%), and 96% of the radioactivity was recovered in the organic phase after solvolysis of the MeOH/water (1:1) fraction from 100 mg of rat brain. This marked improvement indicated that additional sample purification by the water wash was necessary for the efficiency of the subsequent solvolysis reaction.
Hence, we tried to detect endogenous PREG-S and DHEA-S by GC-MS in 100 mg of rat brain by introducing the new purification step. The brain extract was dissolved in 1 ml of MeOH/water (1:1) and applied to the cartridge. After a water wash, steroid sulfates were eluted with MeOH/water (1:1). The conventional solvolysis step described previously (1) was replaced by a hydrolysis/derivatization reaction with HFBA (see Experimental Procedures). Under these conditions, little or no endogenous PREG-S and DHEA-S was detected by GC-MS (data not shown). However, the signal of [2H4]PREG derived from the internal standard [2H4]PREG-S was constant and independent of the tissue amount, suggesting that the hydrolysis/derivatization reaction was complete. The total recovery of the standard was 90%, and the GC-MS baseline was suitable for precisely quantifying trace amounts of PREG-S and DHEA-S in rat brain. Moreover, no PREG-S and DHEA-S could be detected in the MeOH/water (1:1) fraction by direct LC-MS analysis.
The SPE water wash was analyzed by GC-MS to search for the expected levels of PREG-S and DHEA-S previously detected in rat brain (1). Substantial amounts of PREG-HFB and DHEA-HFB were detected, suggesting the presence of PREG and DHEA derivatives more hydrophilic than PREG-S and DHEA-S. These derivatives are distinct from PREG-S and DHEA-S, which were not detected by direct LC-MS analysis of the aqueous eluate. It is noteworthy that the water wash contained predominantly phospholipids (phosphatidylcholine, phosphatidylserine, phosphatidylinositol, phosphatidic acid, some corresponding lysophospholipids, and plasmalogens) (data not shown).
The presence of large amounts of lipids in the very polar eluates suggested an inadequacy of the SPE procedure. Although this method seemed to be reliable for PREG and DHEA and their sulfated counterparts (1), its deficiency was demonstrated by the presence of less polar lipids in the polar eluates. This behavior of nonpolar lipids in SPE extraction has previously been obtained in analyses of fecal sterols and bile acids (37). When the dried extract is taken up in MeOH/water (1:1), the brain lipids are not dissolved and are poorly or not at all adsorbed by the C18 cartridge. Hence, we believe that the water wash preceding the MeOH/water (1:1) elution does not elute adsorbed material but only displaces the dispersed lipids through the C18 bed. The massive amounts of phospholipids in the SPE polar fractions probably inhibit the solvolysis as the amount of brain extract is increased. As a consequence, reliable conclusions cannot be drawn regarding the polarity of PREG and DHEA derivatives present in the aqueous eluates. However, PREG and DHEA derivatives eluted by MeOH/water (1:1) according to our previous method (1) were found in the same fraction as steroid sulfates, explaining the reports of so-called endogenous PREG-S and DHEA-S in rat brain.
Systematic recycling and stepwise fractionation by SPE To extract and fractionate all PREG and DHEA derivatives present in rat brain, the extract from 100 mg of rat brain was first dissolved in MeOH to ascertain that all steroids were solubilized. Then, lipids and steroid derivatives were adsorbed by recycling through the C18 cartridge in MeOH/water mixtures of increasing polarity. Subsequent elution was performed with solvent mixtures of increasing concentrations of MeOH (see Experimental Procedures). Fractions were collected and analyzed by GC-MS in SIM mode. In the initial exploration experiments, only PREG derivatives were analyzed because they are present at higher concentrations in rat brain than DHEA derivatives. In parallel, experiments were performed with [3H]PREG-S and [3H]PREG. As shown in Fig. 3 , endogenous PREG-S (similar to DHEA-S; data not shown) could not be detected. Endogenous free PREG was detected in eluates with MeOH/water from 6:4 to 8:2, and a less polar derivative of PREG was detected with MeOH/water from 8:2 to pure MeOH. Hence, the unknown derivative of PREG (now called L-PREG) is a nonpolar form, clearly separated from PREG-S.
|
Measurements of PREG-S and DHEA-S with the new SPE method and validation
Steroid sulfates were extracted from a whole brain and plasma of an adult male rat. All SPE fractions were treated and analyzed as described in Fig. 1. The results are shown in Fig. 4
. PREG-S (and DHEA-S; data not shown) was below the detection limit in rat brain and plasma, and the internal standard, [2H4]PREG-S, was detected with high sensitivity. Endogenous PREG and L-PREG were detected, and the latter was more abundant in both rat brain and plasma. To validate the quantification of PREG-S and DHEA-S, linearity experiments were carried out by extracting steroids from increasing amounts of rat brain (10, 25, 50, 100, and 200 mg). PREG-S and DHEA-S were not detected in any of the brain samples, signifying that the concentrations of endogenous PREG-S and DHEA-S are less than 0.1 ng/g in whole rat brain. The detection limit was calculated from the detection limits for PREG-HFB and DHEA-HFB in the GC-MS analyses (2 pg), the dilution factor (1:6), and the recovery of steroid sulfates through the entire analytical process (75%). Finally, the accuracy was checked by the addition of PREG-S (0.5, 1, 2, and 5 ng) to 100 mg samples of rat brain. The results are satisfactory, and the accuracy is independent of the mass of analyzed brain tissue. Results validating the steroid sulfate analysis are summarized in Table 2.
|
|
|
This indicates that the discrepancies between the results of direct and previous indirect methods for sulfated steroids are of a methodological nature. According to the results shown in Figs. 3, 4, the explanation may involve a contamination of the fraction containing the sulfated steroids by endogenous lipoidal forms of PREG and DHEA present in rat brain that can be readily hydrolyzed and acylated by HFBA. If the unknown lipoidal derivatives can be totally or partially solubilized or form micelles in an aqueous medium without being extractable with apolar organic solvents, this may explain both the presence of so-called steroid sulfates in previous reports (21, 22, 28, 29) and the poor reproducibility of measurements. Therefore, we reinvestigated the influence of the extraction method on the distribution of PREG derivatives in male rat brain. Rat brain tissue was homogenized in isotonic PBS buffer (see Experimental Procedures), and unconjugated and fatty acid esters of steroids were isolated by a differential extraction procedure with ethyl acetate/isooctane. The water phase was fractionated by SPE and derivatized according to the protocol shown in Fig. 1. This experiment clearly illustrates that PREG-S and DHEA-S are not detected by using this methodology, similar to the finding with recycling SPE of a methanolic extract (Fig. 5) . The most striking result was the presence of a lipoidal form of PREG, and to a lesser extent of DHEA, in the aqueous phase of the rat brain homogenate. These results confirm our hypothesis above concerning a contamination with L-PREG and L-DHEA when using this extraction process. Brain lipids, including L-PREG and L-DHEA, are probably not dissolved in the buffer, and some of them may form micelles or other lipid aggregates in aqueous medium. The presence of lipoidal derivatives of PREG and DHEA in the aqueous phase provides a coherent explanation for the identification of PREG and DHEA by GC-MS after acidic solvolysis of the water phase (20, 21). Free PREG and DHEA were indeed released by the solvolysis, but they came from brain lipoidal and not sulfated derivatives.
|
Therefore, our data and the study of Liu, Sjövall, and Griffiths (27) demonstrate that PREG-S and DHEA-S are not present at levels above 0.3 ng/g in rat brain. These results are consistent with the difficulties in detecting and measuring the activity of the hydroxysteroid sulfotransferase enzyme in rat and human brain and indicate that PREG-S and DHEA-S are not endogenous steroid conjugates in rat brain. These findings suggest that they may not be directly involved in cognitive abilities in rat, as proposed by Vallée et al. (10), even if PREG-S (45, 46) and DHEA-S (47, 48) display pharmacological neuroactivity, especially to enhance memory performance.
L-PREG and L-DHEA: artifact or reality?
The unexpected presence of high levels of a lipoidal form, hydrolyzable under acidic conditions, of PREG and DHEA in rat brain and plasma raises the question of whether it is a real endogenous derivative of PREG and DHEA or if it is formed as an artifact during the extraction process in MeOH or in buffer. Incubation experiments were carried out to resolve this question. First, incubations in MeOH were done by adding tritiated PREG-S and PREG with the brain extract for 24 h at 20°C (i.e., the experimental conditions used in our analytical protocol). No complexation was observed between exogenous [3H]PREG-S and [3H]PREG and lipid components from the rat brain in MeOH. The distribution from [3H]PREG-S and [3H]PREG in the different SPE fractions (sulfated, unconjugated, and lipoidal) was the same regardless of the incubation duration (0 or 24 h) and the presence or absence of brain extract. These results are consistent with those reported above showing that the signal intensity of [2H4]PREG-HFB, added as [2H4]PREG-S as an internal standard for PREG-S and DHEA-S, was the same regardless of the amount of brain tissue and corresponding to that expected from the recovery of the whole process. These results rule out the possibility of ion-pair formation between sulfated steroids and brain cationic lipids and the possibility that no interaction takes place between free steroids and lipids in MeOH.
[3H]PREG was also incubated with a rat brain homogenate in buffer to determine if L-PREG could be formed by metabolism. [3H]PREG was added to a rat brain homogenate in buffer for 0, 1, 4, and 48 h at 37°C. Preliminary experiments showed a small increase of the [3H]L-PREG signal in the SPE lipoidal fraction, especially at 0, 1, and 4 h of incubation. This was less evident with a boiled brain extract homogenate (data not shown). Hence, there may be a slight in vitro complexation or conjugation in buffer between the free steroid and some lipid components from rat brain that occurs instantaneously. However, the radioactivity in the SPE lipoidal fraction decreased slightly by prolonging the incubation time to 48 h, possibly indicating a release of free PREG from L-PREG with time, perhaps reflecting an autolysis phenomenon. In conclusion, these experiments show that L-PREG and L-DHEA are not artifactually formed during the methanolic extraction and that they are present as endogenous steroid derivatives in rat brain and plasma.
Analytical approaches for detecting L-PREG and L-DHEA by GC-MS
L-PREG and L-DHEA, which can be cleaved with HFBA and detected in rat brain and plasma, have similar chemical properties as the lipophilic conjugates of steroid sulfates proposed by Oertel and colleagues (49) 30 years ago. Indeed, these latter compounds were indirectly detected in various mammalian tissues and named steroid sulfatides by analogy with phosphatides (these lipophilic sulfoconjugates were composed of steroid, sulfate, glycerol, and fatty acids in molar ratios of 1:1:1:2) (49). However, the evidence for the presence of PREG and DHEA covalently bound to a phospholipid or a sulfolipid was indirect, and no other investigator has confirmed the existence of such derivatives. On the contrary, Sjövall and Vikho (50) found no evidence for such a steroidal complex in studies using gel filtration to separate steroid sulfates and phospholipids.
Prasad et al. (51) also explored this research field by focusing on the lipophilic derivatives of cholesterol sulfate in bovine brain using indirect methods. Based on the reactivity of two model compounds, cholesteryl methyl sulfate and cholesteryl dipalmitoylglyceryl sulfate (cholesterol sulfatide), they proposed an analytical procedure with a sample fractionation isolating sulfated, unconjugated, fatty acid esters, and sulfolipid derivatives of cholesterol. The latter compounds were detected and measured by GC-MS after conversion into the pyridinium sulfate salts when treated in pyridine-containing MeOH (52). Lieberman's group also reported on the presence of considerable amounts of presumed PREG and DHEA sulfolipids in mammalian brains (43). These sulfolipids were also separated from the other conjugates and were cleaved by TEA to give a mixture of sulfate esters and free PREG and DHEA that was analyzed by GC-MS. These hypothetical sulfolipids were the major forms of PREG and DHEA present in rat and rabbit brains, compared with free, sulfated, and fatty acid-esterified steroids. The authors suggested that the difficulties in analyzing these compounds were attributable to their lability and dissociation by treatment with weak nucleophiles (such as water and alcohols) or heating during tissue processing.
We investigated the reactivity and lability of L-PREG and L-DHEA in rat brain extracts toward TEA, organic solvents such as acetone and EtOH, and also the influence of temperature and light. For this purpose, SPE MeOH/CHCl3 (1:1) fractions containing L-PREG and L-DHEA from 100 mg of male rat brain were subjected to different experimental conditions for 30 min or 24 h. A second fractionation by C18 SPE, according to the protocol illustrated in Fig. 1, was used to analyze the sulfated, free, and lipoidal derivatives after these treatments. The results are summarized in Tables 4, 5. First, the absence of PREG-S and DHEA-S after treatment with TEA does not confirm the results of Mathur et al. (43) because only free PREG and no PREG-S was detected. It is noteworthy that neither free DHEA nor sulfated DHEA was formed after TEA treatment, meaning that L-PREG and L-DHEA are characterized by distinct reactivities. Thus, there is no evidence for the presence of a sulfur-oxygen bond in L-PREG and L-DHEA. Furthermore, TEA has no specific effect on L-PREG dissociation, because EtOH or acetone had similar effects. However, temperature increase, length of reaction time, and light caused some L-PREG dissociation.
|
|
70 times greater than that of PREG-HFB formed from the unconjugated fraction after the TEA/HFBA treatment. Considering that the concentration of endogenous unconjugated PREG was in the range of 3–5 ng/g and assuming that experimental losses during the workup were roughly similar for free and lipoidal steroids, the range of L-PREG levels in the whole adult male rat brain is estimated to be 210–350 ng/g (1 nmol/g) using our methodology. These very surprising concentrations seem to support the quantitative results of Lieberman's group (43) concerning high levels of so-called PREG sulfolipid in rat brain and the release of large amounts of PREG after treatment of organic brain extracts with several chemicals (53).
|
Further studies were designated to quantitatively optimize the analysis of L-PREG and L-DHEA in rat brain. An additional SPE elution step with 5 ml of hexane was added to collect lipoidal steroids: 80% of L-PREG was collected in the SPE MeOH/CHCl3 (1:1) fraction, and 20% was recovered in the hexane fraction. In our system, hexane elutes PREG palmitate and cholesterol completely, which consequently have approximately the same polarity as L-PREG and L-DHEA. A systematic study of the derivatization reaction conditions with TEA/HFBA was performed to optimize the yield of released PREG-HFB and DHEA-HFB from the SPE lipoidal fraction. The reaction yield was decreased as the reaction time was increased from 30 min to 6 h (4-fold) and was decreased as the temperature was diminished from 70°C to 20°C (3-fold). Importantly, the reaction yield was optimal when the volumes of TEA and HFBA were the same, indicating that TEA probably acts as a catalyst for the acylation reaction.
Using the optimal reaction conditions [30 min, 70°C, TEA/HFBA (1:1)], unconjugated and lipoidal PREG and DHEA were measured in a qualitative way (no internal standard was added) in whole rat and mouse brain to obtain an estimate of their relative levels. As illustrated in Fig. 7
, the relative levels of DHEA, PREG, L-DHEA, and L-PREG are approximately similar for the two rodents. L-PREG and L-DHEA are the major forms, although the relative concentrations of unconjugated DHEA and PREG seem to be slightly higher in mouse than in rat brain. DHEA-S and PREG-S are also not detected in mouse brain. In rat plasma, L-PREG is more abundant than free PREG (2-fold), but to a much lesser extent than in rat brain (70-fold) (data not shown). This finding indicates that lipoidal derivatives of PREG and DHEA are circulating steroids.
|
For these reasons, the occurrence of the transesterification reaction with TEA/HFBA was tested with PREG palmitate and DHEA stearate. However, PREG-HFB and DHEA-HFB were not formed. Furthermore, exogenous PREG palmitate was added to a rat brain extract. After C18 SPE separation, the lipoidal fraction was derivatized with TEA/HFBA and purified for GC-MS analysis. Released PREG-HFB levels were similar in brain samples with or without the addition of PREG palmitate, meaning that transesterification with TEA/HFBA does not occur with fatty acid esters of steroids in a brain lipidic environment. These results exclude the possibility that fatty acid esters constitute the lipoidal forms of PREG and DHEA detected in this study.
Finally, the contribution of fatty acid esters of PREG and DHEA relative to L-DHEA and L-PREG in rat brain and plasma was investigated. The SPE lipoidal fraction was saponified, and the distribution of free and lipoidal steroids in a subsequent SPE fractionation was analyzed. The release of free PREG from the SPE lipoidal fraction was enhanced by saponification in rat plasma and to a lesser extent in rat brain (Fig. 8A) . This confirms the existence of fatty acid esters of PREG in rat brain and plasma. However, no proof of the existence of fatty acid esters of DHEA was obtained for any of the samples. Furthermore, the released amounts of free PREG by saponification of the SPE lipoidal fraction were negligible in rat brain (4.7%) and in rat plasma (0.5%) (Fig. 8A) compared with the respective amounts of L-PREG (Fig. 8B). These experiments confirm that L-PREG and L-DHEA are by far the major derivatives of PREG and DHEA compared with their unconjugated and fatty acid ester counterparts.
|
We performed additional experiments to evaluate the sulfolipid hypothesis. As described above, PREG-S and DHEA-S are hydrolyzed and derivatized when submitted to HFBA. However, in contrast to L-PREG and L-DHEA, the addition of TEA inhibited the hydrolysis/derivatization reaction of sulfated steroids with HFBA (data not shown). Further experiments were carried out to study the reactivity of steroid sulfonate esters. For this purpose, PREG tosylate was added to the SPE lipoidal fraction from rat brain and derivatization with HFBA or TEA/HFBA was attempted. Our results show that TEA also inhibited the hydrolysis/derivatization of PREG tosylate with HFBA. Together, these data indicate that L-PREG and L-DHEA are brain lipoidal derivatives that do not contain a covalent oxygen-sulfur bond at C-3.
The possibility that L-PREG and L-DHEA represent ion pairs between PREG-S and DHEA-S and some positively charged lipid such as phosphatidylcholine, phosphatidylethanolamine, or sphingomyelin was also considered. However, a strong argument against the formation of ion pairs is the fact that PREG-S and DHEA-S are recovered as such when they have been added to the brain extracts.
The peroxide hypothesis In 1994, Lieberman's group (53) formulated a new hypothesis concerning neurosteroids. By treating extracts of rat brain with a variety of reagents, such as TEA, imidazole, FeSO4, HCl, and FeCl3, substantial amounts of PREG and DHEA were released, notably from a nonketonic lipid fraction. The authors suggested that compounds such as sterol or cholesterol 20-hydroperoxides or 17-hydroperoxides, and possibly other uncharacterized conjugates called neurosteroid precursors, may be present in mammalian brain and may liberate PREG and DHEA after some chemical treatments. Furthermore, from these pioneering results, they postulated that the functions of cytochromes P450scc (61) and P450C17 (62) still must be evaluated and that alternative biosynthetic pathways for the in vivo formation of PREG and DHEA may exist. Particularly, the absence of P450C17 activity, protein, and mRNA in C6 rat glioma tumor cells (63) indicated that DHEA might be synthesized in the brain by P450C17-independent pathways, distinct from adrenal and gonadal steroidogenesis. These observations were extended to differentiating human (64) and rat (65) brain oligodendrocytes and astrocytes.
According to our results, it seems unlikely that TEA/HFBA treatment of the SPE lipoidal fraction, containing L-PREG and L-DHEA, would cause cleavage of sterol hydroperoxides or cycloperoxides with the formation of PREG-HFB and DHEA-HFB, respectively. Furthermore, additional preliminary data have shown that PREG-TMS and PREG-3-HFB-20-MO can be formed from the SPE lipoidal fraction by the addition of MSTFA and methoxyamine/HFBA reagents, respectively (see Experimental Procedures). Although the amounts of PREG derivatives with MSTFA and methoxyamine/HFBA reagents were lower than after TEA/HFBA, it is very likely that endogenous L-PREG already contains the PREG skeleton, which is not in agreement with the peroxide hypothesis. Then, even though L-PREG and L-DHEA are characterized by heat and light lability and are less polar than PREG and DHEA, as are sterol hydroperoxide or dioxetane compounds, there are arguments against oxygenated sterols to account for the detection of PREG-HFB and DHEA-HFB by GC-MS.
The lipoprotein hypothesis A new hypothesis is proposed here for the structural nature of rat brain L-PREG and L-DHEA. A complex between steroids and some lipid constituents in the brain may be envisaged. Lipoproteins, comprising steroid-binding apolipoproteins, would be possible candidates for this purpose. They are involved in specific binding to cellular receptors and in the lipid exchange and transfer processes. In addition to cholesterol and cholesteryl esters, Roy and Bélanger (66) have demonstrated the presence of low-polarity PREG and DHEA conjugates (and to a lesser extent free PREG and DHEA) in human serum lipoproteins. They suggested that these lipoproteins may deliver steroid conjugates to cells via receptor-mediated endocytotic pathways, as has been described for cholesterol (67). In our experiments, the lipophilic nature, the lability, and the presence of L-PREG and L-DHEA in rodent plasma and brain are in agreement with a possible noncovalent association of PREG and DHEA with some apolipoproteins in lipoproteins. The high abundance of L-PREG and L-DHEA in rat plasma relative to unconjugated and sulfated PREG and DHEA and the fact that L-PREG and L-DHEA are not extracted with organic solvents (isooctane and ethyl acetate) from an aqueous phase are arguments for a lipoprotein/steroid association. Furthermore, the results obtained with MSTFA and methoxyamine/HFBA derivatizing reagents indicate that the steroid skeletons of PREG and DHEA are maintained in the L-PREG and L-DHEA structures. Concerning the influence of TEA on the efficiency of the derivatization of L-PREG and L-DHEA with HFBA, this reagent could act as a catalyst but also could dissociate noncovalent interactions between steroid and apolipoprotein as a denaturant agent. Similarly, Mano et al. (68) have reported high concentrations of chenodeoxycholic acid in rat brain cytoplasmic fraction upon extraction with guanidine, indicating that this bile acid is noncovalently bound to protein. The incubations of tritiated PREG with a rat brain homogenate in buffer indicate that an in vitro association reaction, although weak, occurs between the steroid and a lipoidal brain component, even if the formation of fatty acid esters of PREG cannot be excluded under these conditions.
Another important point is the absence of lipoidal derivatives of PROG and its metabolites allopregnanolone and epiallopregnanolone, testosterone and its 5- and 35-reduced metabolites (allodihydrotestosterone and 5-androstan-3,17ß-diol), or estradiol in rat brain, which can be derivatized by TEA/HFBA. These observations suggest that these novel rat brain and plasma lipoidal derivatives are specific for 3ß-hydroxy-5 steroids.
A particular apolipoprotein, apolipoprotein D (apoD), first detected in human plasma (69), may be a candidate to bind PREG and DHEA in the rodent CNS. This 29 kDa glycoprotein, associated with HDL, binds small hydrophobic ligands such as PREG, PROG, cholesterol, bilirubin, and arachidonic acid even if it is not clear if these compounds are physiologic ligands. Very interestingly, human apoD mRNA is poorly expressed in liver and intestine, which are the major sites of synthesis of other apolipoproteins, but it is highly expressed in adrenals, kidneys, spleen, placenta, brain, and peripheral nerves, and large amounts of apoD were found in cerebrospinal fluid and serum (70, 71). The tissue distribution is different in the mouse (72) and rat (73), with apoD mRNA expression largely restricted to CNS (spinal cord, cerebellum, and brain) and with little amounts of apoD in plasma. The cellular expression of apoD is observed mainly in glial cells of the white matter, particularly oligodendrocytes and astrocytes, of the CNS and in endoneurial fibroblasts of the peripheral nervous system (73). Furthermore, apoD (and apoE) seems to be locally synthesized in the CNS and peripheral nervous system, involved in maintenance and repair processes in the CNS in neurodegenerative pathologies, and implicated in regeneration processes (74). In the CNS, PREG is synthesized by astrocytes and oligodendrocytes (75), cells that also express apoD, which could play a role in the local transport of PREG. Indeed, as suggested by Provost et al. (76), a principal function of apoD could be performed locally rather than in the circulation, and it could play a paracrine role in the distribution of its ligands. However, the possibility that the unidentified endogenous L-PREG and L-DHEA present in rodent brain could be lipoproteins containing PREG and DHEA binding apolipoproteins remains to be proved. Further studies are required to establish the chemical nature of L-PREG and L-DHEA in rodent CNS and to determine if these steroids represent single or multiple chemical entities.
In conclusion, the results presented here clearly demonstrate that PREG-S and DHEA-S are not detected (i.e., less than 0.3 ng/g) in male rat and mouse brain by GC-MS, in agreement with recent results of analyses of the intact steroid sulfates. Very high levels of lipoidal derivatives of PREG and DHEA were detected in rat and mouse brains. These derivatives seem to be the source of an "internal contamination" of the so-called steroid sulfate fractions in the majority of former studies concerning sulfated neurosteroid. L-PREG and L-DHEA are derivatized with HFBA in the same way as unconjugated PREG and DHEA, except that TEA catalyzes with high efficiency the release of PREG-HFB and DHEA-HFB. The determination of L-PREG and L-DHEA in rat brain allows us to discard some hypotheses about their chemical nature, such as fatty acid esters and sulfolipidic conjugates. However, our suggestion that PREG and DHEA may be incorporated into lipoproteins in rat CNS and plasma needs to be experimentally verified. Notably, some characteristics of apoD, such as PREG binding and tissue and cellular expression, could indicate that this apolipoprotein may bind PREG and DHEA in relatively stable complexes that may play a role in steroid transport in normal and regenerating CNS and the peripheral nervous system, as does apoE. Further investigations will be focused on the chemical nature and tissue distribution of L-PREG and L-DHEA to assess their brain specificity in mammals before studying their biological functions in the nervous system. Considering the large amounts of L-PREG and L-DHEA in rodent brain, it will be of great interest to determine whether L-PREG and L-DHEA serve as steroid precursors and/or as storage compounds or whether they have more direct physiological functions.
|
ACKNOWLEDGMENTS |
|---|
Manuscript received June 24, 2004 and in revised form July 21, 2004. and in re-revised form August 31, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
