| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Developmental Origins of Health and Disease Research Division, University of Southampton, Princess Anne Hospital, Southampton SO16 5YA, United Kingdom
Address all correspondence and requests for reprints to: Malcolm C. Richardson, Developmental Origins of Health and Disease Research Division, University of Southampton, Level F, Princess Anne Hospital, Coxford Road, Southampton SO16 5YA, United Kingdom. E-mail: mcr2{at}soton.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Important aspects of insulin action are mediated via a group of transcription factors termed sterol regulatory element-binding proteins (SREBPs) (reviewed by Horton et al. in Ref. 5). These control the expression of more than 30 genes concerned with the synthesis and uptake of cholesterol, fatty acids, and other lipids, many of which are likely to be important in the function of the granulosa cell, providing substrates for the production of lipid intermediates intended for transfer to the developing oocyte and steroid synthesis after luteinization. There are three members of the SREBP family, designated SREBP-1a, SREBP-1c, and SREBP-2. Unlike SREBP-2, which is encoded by a separate gene, both SREBP-1 isoforms are encoded by the same gene using different promoter systems and transcription start sites that lead to two alternative forms of exon 1. Once formed, the two SREBP-1 isoforms are incorporated into the endoplasmic reticulum and, under conditions of low cellular cholesterol, cleaved to release structurally distinct N-terminal fragments with different capacities to activate gene expression. For example, the SREBP-1c isoform, generally a weaker activator of gene expression, preferentially enhances the transcription of genes required for fatty acid rather than cholesterol synthesis. A picture emerges of two distinct physiological roles for the two SREBP-1 isoforms influenced by their relative abundance in particular tissues (6). Moreover, the two promoter systems used by the two isoforms provide a mechanism for the differential control of their expression through hormonal influences. Thus, in liver, higher exposure to circulating insulin, associated with insulin resistance, leads to a specific increase in SREBP-1c expression, which increases hepatic fatty acid synthesis (7).
There is evidence for the expression of SREBP isoforms in the ovary. In a study that examined a wide range of tissues, Shimomura et al. (6) recorded a 2.8-fold preponderance of SREBP-1c over SREBP-1a expression in the human ovary, acknowledging that different cell types within such a complex tissue could be expressing different ratios of the two transcripts. Also, the N-terminal SREBP-1a fragment has been demonstrated in rat ovary (8) and human granulosa cells (9), its cleavage from the membrane-bound SREBP altered, as expected, through the availability of cholesterol. Accumulating evidence thus points to a functional SREBP system in ovarian steroidogenic cells, which would control cholesterol and fatty acid synthesis. Additional regulation by SREBP of steroidogenesis through sterol regulatory element (SRE) sites in the promoter region of the gene encoding steroid acute regulatory protein (StAR) has also been proposed (8, 9). Clearly, our present knowledge would be extended by an understanding of the differential expression of the mRNAs encoding the separate isoforms of SREBP in granulosa cells under various conditions of hormonal stimulation. Whether insulin action on granulosa cells involves such a differential effect on expression is of particular interest because the liver and ovary may be similarly affected by higher circulating insulin associated with whole-body insulin resistance, a common feature of PCOS.
Because of differences in the proportion of individual SREBP isoforms expressed by individual cell types (6), we developed a method for the preparation of human granulosa cells free of contaminating white blood cells, a group of cells expressing their own specific array of SREBP isoforms with a predominance of SREBP-1a over -1c (10). Using our newly developed method, involving removal of white cells with anti-CD45-linked magnetic beads, granulosa cells were prepared and exposed in culture to gonadotropin and/or insulin under conditions in which a demonstrable and sensitive effect of insulin was apparent as judged by several parameters. This methodology enabled the specific measurement of the SREBP-1c to -1a ratio within granulosa cells under a range of culture conditions, without the influence of background levels of these isoforms in contaminating cell types. With reference to the paradigm established in liver in which insulin causes a shift toward expression of the SREBP-1c isoform (7), a hypothesis was formulated that differential modulation of the SREBP-1c to -1a ratio by insulin (and gonadotropin) may also occur in granulosa cells. Our results suggest some commonality in the underlying features of the control by insulin of SREBP isoforms in liver and ovary and may have implications for the impact of clinically defined insulin resistance on the ovary.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Follicular aspirates were obtained at ovum collection for IVF according to a procedure approved by our local ethical committee. Patients selected for the study had apparently normal ovarian function with either tubal blockage, had a partner with male infertility, or were acting as egg donors. The treatment protocol was adapted from a previously described method (11) that involved down-regulation of pituitary function with GnRH analog (nafarelin: 400 µg intranasally twice daily) starting in the previous luteal phase. From d 4 of the IVF cycle, 150–600 IU FSH was administered daily. When the leading two follicles had a diameter of more than 18 mm and serum estradiol concentration was more than 300 pmol/liter for each follicle larger than 14 mm, human chorionic gonadotropin (hCG) (10,000 IU) was given and oocyte collection arranged 34 h later. Follicles aspirated were more than 15 mm in diameter.
Granulosa cell preparation and purification
Follicular aspirates and washes were combined and centrifuged (100 x g for 10 min). Cell pellets were resuspended in about 20 ml Hank’s balanced salt solution (HBSS) and incubated for 30 min with hyaluronidase (0.5 mg/ml; Boehringer, Lewes, UK). The mixture was then layered on to 45% Percoll (Pharmacia, Milton Keynes, UK) diluted in HBSS and centrifuged for 20 min at 100 x g. Cells on the interface were collected, washed twice by centrifugation and resuspension, and diluted in a mixture (50:50) of DMEM and Ham’s F12 (Invitrogen, Paisley, UK), containing 10% fetal bovine serum, glutamine (2 mmol/liter), penicillin (100,000 IU/liter), streptomycin (100 mg/liter), and amphotericin (0.25 mg/liter). Cells were distributed among four individual 50-ml culture flasks that were then incubated at 37 C overnight in the presence of 95% air-5% CO2. On the following day, adherent cells were rinsed with HBSS and then trypsinized, pooled, and reestablished in culture for 5 h in one 50-ml flask. The medium was then removed and 1 ml HBSS added together with 50 µl of a suspension of paramagnetic microbeads linked to anti-CD45 antibody (Miltenyi Biotec, Bisley, UK). The mixture was maintained at 4 C with occasional gentle rocking for 15 min allowing attachment of the beads before addition of 1 ml HBSS containing 10% fetal bovine serum and 4 mmol/liter EDTA adjusted to pH 7.2. Gentle agitation at 4 C for a further 10 min resulted in the removal of the cells from the culture surface so that they could be aspirated through a fine pastette before application to the magnetized column matrix (Miltenyi Biotec). White blood cells were retained on the column, whereas the granulosa cell population was allowed to run through into a collecting tube containing excess culture medium. The white cells could be harvested separately by elution after removal of the column from the magnet.
Monitoring of white cell removal
Various cell fractions, before and after separation, were fixed for 20 min in 4% paraformaldehyde, treated for 1 h with antihuman CD45 antibody (1:200 dilution; Becton Dickinson, Oxford, UK) and then exposed for 1 h to a fluorescein isothiocyanate-labeled goat antimouse antibody (dilution 1:200; Sigma, Poole, UK). After washing in PBS, the cell fractions were examined under the fluorescence microscope. In this way, the original cell suspension (Fig. 1A
), unseparated cells cultured overnight (Fig. 1, B and C), the freshly separated granulosa cell fraction (Fig. 1D), and cultured granulosa and white cell fractions (Fig. 1, E and F) were monitored. The inclusion of propidium iodide (1 µg/ml) confirmed good nuclear preservation, without obvious apoptosis, in the final granulosa cell cultures used for the study (Fig 1, E and F).
|
After centrifugation and resuspension in culture medium, the cells in the granulosa fraction were counted using a hemocytometer and adjusted to approximately 105 cells/ml. Viability (routinely > 90%) was checked by trypan blue exclusion. Dispersion of Matrigel (25 µl/ml; Becton Dickinson) into the cell suspension was carried out immediately before plating of the cells into a 96-well culture dish. After culture overnight to allow attachment, the culture medium was removed and replaced by a defined medium consisting of the Hams F12/DMEM base medium with antibiotics as before but with additions of transferrin (5 mg/liter), selenite (25 nmol/liter), and albumin (500 mg/liter; ELISA grade with low insulin concentration; Sigma). A further 8 h in culture preceded a final change into defined medium with additions of hCG alone (100 ng/ml), insulin alone (1000 ng/ml), or both hormones combined. A 16-h experimental culture period followed, allowing the effects of the hormonal additions to be investigated. This short culture period did not result in differences in cell number between the treatment groups as assessed by DNA assay (4). Culture medium was taken directly for assay of lactate or stored at –20 C for later assay of IGFBP-1 or progesterone. Adherent cells were taken for total RNA extraction.
Lactate assay
Immediately after culture, 10 µl culture medium were added to 225 µl assay reagent (lactate oxidase, peroxidase, and chromogen precursor; Sigma Diagnostics, Poole, UK), contained within a 96-well plate. Lactate was quantified by reading the plate at 540 nm and comparing values with a 2-µg lactate standard, assuming that the absorbance at this wavelength was directly proportional to the lactate concentration.
IGFBP-1 assay
An ELISA for IGFBP-1 was developed using the Duoset reagents supplied by R&D Systems (Abingdon, UK). Briefly, 96-well plates were coated overnight at room temperature with capture antibody (100 µl/well of a 4 µg/ml solution of mouse antihuman IGFBP-1 reconstituted in PBS) and then blocked with a solution of 5% (vol/vol) Tween 20 and 5% (wt/vol) sucrose in PBS containing 0.05% (wt/vol) sodium azide. Standards were prepared using serial dilution in reagent diluent [5% (vol/vol) Tween 20 in PBS] to give a range of 125-4000 pg IGFBP-1 per milliliter. One hundred-microliter aliquots of conditioned media (diluted 1:1 with reagent diluent) or standards were added in duplicate and the plates incubated at room temperature for 2 h. After the addition of detection antibody (100 µl of 400 ng/ml biotinylated goat antihuman IGFBP-1 in reagent diluent), the plates were incubated for a further 2 h. Color was developed using a streptavidin-horseradish peroxide conjugate with subsequent addition of a hydrogen peroxide/tetramethylbenzidine substrate solution (R&D Systems). After stabilization of the color with 2 N H2SO4, ODs were read using a plate reader set at 450 nm with correction wavelength at 540 nm. The following factors at 50 ng/ml were assayed and exhibited no cross-reactivity or interference: IGF-I, IGF-II, IGFBP-2, IGFBP-3, IGFBP-4, IGFBP-5, and IGFBP-6. The intra- and interassay variation was approximately 6 and 12%, respectively, and the limit of detection was about 60 pg/ml.
Progesterone assay
Progesterone was assayed in culture medium using an automated, solid-phase, chemiluminescent enzyme immunoassay (Immunolite analyzer system; Diagnostic Products Corporation, Llanberis, UK). Over a working range of 20–120 nmol/liter, the intra- and interassay variations were less than 7 and less than 12%, respectively.
Total RNA preparation
After removal of culture medium, cells were scraped into ice-cold Tri reagent (75 µl/well; Sigma) providing a monophase solution of guanidine thiocyanate and phenol. Replicate wells (five per treatment) were combined and 75 µl chloroform added. After mixing and centrifugation for 15 min at 13,000 x g, the upper aqueous layer was transferred to a new tube and 40 µg glycogen added as carrier. After addition of isopropanol (0.5 ml for each milliliter of Tri-reagent used), the tubes were allowed to stand at room temperature for 10 min and then centrifuged as before. Resulting pellets were washed in 75% ethanol and then air dried before solution in 20 µl RNase-free water. The yield of total RNA was quantified by spectrophotometry.
cDNA synthesis
Total RNA (
0.5 µg) in 12.5 µl RNase-free water was denatured at 70 C for 5 min and then chilled to 4 C. Additions were then made to give 0.5 µmol/ml deoxynucleotide triphosphate mix, 25 µg/ml random hexamer, 1000 U/ml RNAsin, and 10,000 U/ml Moloney murine leukemia virus reverse transcriptase in buffer [50 mmol/liter Tris-HCl (pH 8.3), 75 mmol/liter KCl, 3 mmol/liter MgCl2, and 10 mmol/liter dithiothreitol; all reagents supplied by Promega (Southampton, UK)]. The reverse transcriptase reaction was carried out at 42 C for 1 h and stopped by heating at 95 C for 5 min.
mRNA quantification
This was carried out using a quantitative, competitive PCR method as previously described (12, 13, 14, 15). Briefly, a 385-bp standard DNA was prepared that shared identical PCR binding sequences with the target cDNA. The standard DNA was quantified by spectrophometry and diluted appropriately for inclusion in the PCR. For each target cDNA, primer sequences were designed to yield a PCR product approximately 60–80 bp either shorter or longer than the standard DNA. Each primer pair developed (see Table 1) yielded a PCR product that corresponded to a sequence running across a boundary between exons, minimizing the possibility of expansion of areas of genomic DNA.
|
7 ng/tube), standard DNA (0.04 attomoles/tube for 28 cycles and 0.01 attomoles/tube for 30 cycles), primers (0.6 µmol/liter; sequences shown in Table 1
), 1 U/tube of a hot-start Taq DNA polymerase (HotGoldstar, Eurogentic, Romsey, UK), 2 mmol/liter MgCl2, 15 mmol/liter Tris-HCl (pH 8.0), 50 mmol/liter KCl, and 0.01% (vol/vol) Tween 20. Amplification conditions were as previously described (15) with a preincubation step of 10 min at 95 C to allow activation of the hot-start enzyme. PCR products were separated by electrophoresis, visualized with ethidium bromide (0.5 µg/ml), and analyzed using an image analyzer package (Phoretix, Newcastle upon Tyne, UK). Comparison of the two adjacent but separate bands of standard and target (exemplified by that seen in Fig. 5A) enabled quantification of the target mRNA in the original extract, which was expressed as attomoles per microgram total RNA. A representative intraassay coefficient of variation between replicates for the PCR was 9%.
|
To obtain sufficient cellular material for this analysis, the original, unfractionated granulosa cell preparation was cultured in flasks for 3 d in serum-containing medium and then 3 d in defined medium containing hormonal additions. At the end of this period, medium was removed and cells extracted with chloroform-methanol (2:1, vol/vol), extracts dried under nitrogen, and methyl esters prepared and then analyzed by gas chromatography using flame ionization detection (16). Individual, identified fatty acid peaks were integrated and amounts calculated with reference to an internal standard of methyl heptadecanoate. Addition of these values gave a total fatty acid content (nonesterified plus esterified) for each original cell sample. The most prevalent fatty acids measured were palmitic (16:0), stearic (18:0), oleic (18:1; most abundant), elaidic (18:1 trans), arachidonic (20:4), and lignoceric (24:0) acids.
Statistics and EC50 calculations
One set of cultures was established for each subject. Each set of results illustrated (where N value gives number of subjects) was inspected by ANOVA and individual comparisons made mostly by paired t test or Wilcoxon signed rank test in which the original data failed a normality test (Sigmastat, Sigma). Comparisons in Fig. 5 were carried out on a nonpaired basis using the Mann-Whitney U rank sum test as advised by the statistics package after inspection of the variance in the two groups. Estimates of EC50 values were calculated from data (see Fig. 3) using GraphPad Prism (San Diego, CA). This was straightforward for Fig. 3A in which the effect showed a plateau between 30 and 100 ng/ml of insulin. For Fig. 3B, a further marginal decrease of about 8% in IGFBP-1 levels (over and above that seen at 100 ng/ml) occurred as concentrations of insulin were raised to 1000 ng/ml (data not shown in figure). In this case, the complete data set up to 1000 ng/ml were used to calculate the EC50 value.
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
An assessment of the efficiency of white cell removal from the granulosa cell preparation is provided by Fig. 1, showing the CD45-positive staining (typical of leukocytes) in the various cell fractions. We estimate that the overall level of white cell contamination in the original cell suspension (Fig. 1A) was typically about 40–50% although variable between preparations. After culture overnight (but before separation), white cells often appeared as small clusters of rounded cells adjacent to granulosa cells now flattening and spreading across the culture surface (Fig. 1, B and C). After separation, the granulosa cell fraction, before and after culture, was virtually free of CD45-positive cells (Fig. 1, D and E) and was quite distinct from the white cell fraction, which clearly stained positively for this antigen (Fig. 1F). A quantitative estimate, through cell counting, determined that the resulting level of contamination of the granulosa cell fraction with white cells was less than 1%.
Hormonal effects on lactate, IGFBP-1, and progesterone production
To provide a measure of the effectiveness of insulin (in relation to gonadotropin) within the purified granulosa cell preparation, several potential indicators of insulin action were estimated in culture medium after the final 16 h incubation period (Fig. 2). Lactate production (Fig. 2A) was significantly increased (P < 0.001) in the presence of a maximal concentration (1000 ng/ml) of insulin, reaching a level somewhat less than that achieved with 100 ng/ml hCG, a concentration of hormone previously shown to be maximal for stimulation of progesterone production (4). Increased lactate output in the presence of insulin and hCG together, showed a degree of synergism between the effects of the two hormones. The sensitivity of the effect of insulin alone on lactate production was investigated in two representative experiments shown in Fig. 3A, demonstrating an EC50 of about 1–10 ng/ml (average calculated EC50 for the two experiments, 2.6 ng/ml). Significant inhibitory effects (P < 0.001) of insulin (1000 ng/ml) on IGFBP-1 production by the granulosa cells, in the presence and absence of maximal hCG, are shown in Fig. 2B. The sensitivity of the inhibitory response to insulin alone investigated in two representative experiments (Fig. 3B), was shown to be similar to that shown for the lactate response (average calculated EC50 value for experiments shown, 1.5 ng/ml). As expected, 100 ng/ml hCG caused a more than 4-fold increase in progesterone production in either the presence or absence of insulin (P < 0.001 for both conditions; Fig. 2C). Although there was no consistent effect of insulin alone on progesterone production, there was a positive effect of insulin in 10 of 12 patients (see expanded scale in Fig. 2D showing an average stimulatory effect of about 19%), albeit relatively small, compared with the response to gonadotropin.
|
As a positive control, establishing that meaningful and consistent changes in specific mRNA levels can be measured by the methodology developed, mRNA for StAR was estimated within total RNA extracts prepared from the granulosa cell cultures at the end of the final 16 h culture period (Fig. 4). As expected, mRNA for StAR was significantly increased (P < 0.05) by maximal hCG (with or without insulin) by about 4-fold, providing a very similar pattern of responses to those obtained for progesterone production (Fig. 2C). There was no significant effect of insulin alone on induction of mRNA for StAR, a result again consistent with measurements of progesterone production (Fig. 2C). Linear regression analysis of individual results for insulin action on StAR vs. insulin effects on progesterone revealed no significant correlation between the two parameters (R value, 0.43).
|
Results of the measurement of the two isoforms of SREBP-1 (1a and 1c) in the granulosa cells, cultured under our standard conditions, are shown in Fig. 5, both as representative electrophoretograms (Fig. 5A) and in graphical form (Fig. 5B). Whereas there was no significant effect of either insulin or hCG on the SREBP-1a isoform, there were increases in mRNA for SREBP-1c with insulin alone (about 6-fold), hCG alone (about 10-fold), and in combination (about 15-fold; P < 0.01 for each intervention vs. control). Equivalent measurement of mRNA for SREBP-2 showed no significant effect over the four treatment modes [means with SEM (n = 4) as attomoles mRNA per microgram RNA: control, 1.26 ± 0.53; hCG, 1.94 ± 0.47; insulin, 1.42 ± 0.56; hCG/insulin, 1.85 ± 0.64].
Changes in expression of a target gene for SREBP-1c
A representative target gene responsive to SREBP-1c action was chosen for study. Fatty acid synthase was selected as being an important regulatory enzyme in the control of fatty acid synthesis. Measurements of the expression of this enzyme in response to maximal hCG and insulin concentrations are shown for three experiments (Fig. 6). On average with hCG, there was a 24-fold increase in the expression of fatty acid synthase (P = 0.006), whereas insulin was somewhat less effective, eliciting an average increase of 19-fold (P = 0.024).
|
The average value for total fatty acid content for the cells incubated without hormonal addition was 88.0 µg/106 cells. Treatment with insulin (1000 ng/ml) caused an elevation of this fatty acid content to 134 ± 10.4% of control (the elevation calculated for each of four experiments and then averaged with SEM). This effect of insulin was significant (P = 0.03 vs. control; paired t test). In general, the changes in the individual fatty acids measured were similar to those recorded for the total content. For example, arachidonic acid changed to 133 ± 12% of control after insulin treatment.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
, D and E) was virtually free of contaminating white cells and provided an excellent basis for further work.
Several parameters, measurable in culture medium, were used to assess the hormonal responsiveness of the culture system. First, lactate provided a quick and reproducible measure of both gonadotropin and insulin action (Fig. 2A). The finding that gonadotropin and insulin separately stimulated lactate production by granulosa cells is consistent with previous work (1, 19, 20). However, our observed synergism between the two hormones (Fig 2A) is a novel finding. An improved sensitivity to insulin in the present study (Fig. 3A), compared with that shown by Lin et al. (1), is notable and establishes that these responses to insulin occur within ranges of insulin encountered physiologically. As a second parameter for measurement in the culture medium, IGFBP-1 was chosen because the production of this binding protein is well established in granulosa cells (21) and is known to be inhibited by insulin (22). Our findings (Fig. 2B) are entirely consistent with this earlier work and show a strong inhibitory action of insulin in both the presence and absence of hCG. Evidence that low, physiological levels of insulin are able to elicit the inhibitory action on IGFBP-1 output is provided by Fig. 3B. These effects of low levels of insulin, both on lactate and IGFBP-1 production, are consistent with the filling of insulin receptors by insulin (23). However, where a maximal level of insulin (1000 ng/ml) was used (Figs. 2, 4, 5, and 6), the possibility of additional interaction of insulin with IGF-I receptors [discussed by Steele-Perkins et al. (24)] is acknowledged.
Hormonal responsiveness was finally assessed through the output of progesterone by the granulosa cells (Fig. 2, C and D). As expected, progesterone production was very responsive to maximal hCG and this was associated with elevations in the expression of StAR (Fig. 4), a transport protein responsible for the mitochondrial translocation of cholesterol that is rate limiting for steroidogenesis (25), and induced by gonadotropin (26). The rather marginal effect of insulin on progesterone output in the present study (present in 10 of 12 preparations, Fig. 2D) is consistent with evidence showing a considerable variation in the reported effectiveness of insulin on granulosa cells in this regard (22, 27, 28). It has been suggested that provision of exogenous cholesterol by prior exposure to serum in culture may preload the cells with cholesterol ester (an integral part of the luteinization process) reducing the potential impact of insulin-induced changes in cholesterol synthesis on substrate supply for progesterone production (27), and this may have applied to the present study. The lack of significant effect of insulin alone on the expression of StAR (Fig. 4) is consistent with a firm link between this rate-limiting intermediate and steroidogenesis (25, 29).
Our finding that SREBP is expressed in human granulosa cells is consistent with previous work that has established the presence of the SREBP-1 protein in human granulosa cells (9), porcine granulosa cells (30), and hCG-treated rat ovaries (8). The present study on the separate isoforms of SREBP builds on the work of Shimomura et al. (6), who investigated the expression of the separate mRNA transcripts for SREBP-1a and -1c in a range of organs including the ovary. However, instead of providing a composite picture of isoform prevalence for the combined cell types in the whole ovary, our work provides novel information on a purified steroidogenic cell type. The importance of the removal of contaminating white blood cells, with their own characteristic balance of SREBP isoforms (10), is worth emphasizing in this regard. Our study has the additional benefit of being able to monitor changes in SREBPs that occur as a result of hormonal stimulation (Fig. 5).
Shimomura et al. (6) found high SREBP1c to -1a ratios in liver and in steroidogenic tissues (human adrenal, 5.5:1; human ovary, 2.8:1), emphasizing the importance of the SREBP-1c isoform, which has a preferred action on fatty acid rather than cholesterol synthesis (5). The predominance of the SREBP-1c isoform was understandable in liver, which synthesizes more fatty acids than cholesterol, but difficult to explain in steroidogenic tissues in which a greater need for cholesterol would be anticipated. The present study now shows that, when we consider the granulosa cell alone, the expression rates of SREBP-1a and -1c are broadly similar (Fig. 5) and depend on the state of hormonal stimulation (control, SREBP-1a > -1c; with hCG/insulin, SREBP-1c > -1a). The situation remains, however, that hCG, known to increase steroidogenesis, has the effect of increasing SREBP-1c with a predominant effect on fatty acid synthesis.
Our finding that it is the SREBP-1c isoform that is subject to hormonal control in our system (Fig. 5) is consistent with the general view that SREBP-1a appears to be constitutively expressed at low levels in most tissues of adult animals (5). Moreover, there is a strong parallel between the observed stimulatory effect of insulin on SREBP-1c in the present study and a similar action of insulin in liver, in which increased insulin action leads to increased fatty acid synthesis [reviewed by Horton et al. (5)]. In insulin-resistant states, it appears that high circulating insulin continues to stimulate SREBP-1c in the liver, despite resistance to insulin action in some peripheral tissues (5). Our results now indicate that this paradigm established in liver may apply more widely, reflecting some commonality in the underlying mechanisms of insulin action. In conformity with this pattern, we speculate that the ovary, like the liver, may respond to the high level of circulating insulin found in insulin-resistant conditions such as PCOS through an up-regulation of the expression of SREBP-1c with consequent local effects on ovarian function.
The levels of SREBP-2 observed in the present study were broadly equivalent to the combined levels of SREBP-1a and -1c and did not change dramatically with hormonal treatment. SREBP-2, which has a predominant effect on cholesterol synthesis (5), in combination with SREBP-1a, would thus provide sufficient flux through the cholesterol pathway for granulosa cells irrespective of hormonal action. The stimulatory effects on SREBP-1c observed in the present study, potentially influencing fatty acid synthesis, are thus superimposed on this background level of control provided by the other SREBPs.
Several studies have examined the possibility that SREBP isoforms act as transcriptional regulators of the StAR gene. For example, Christenson et al. (9) reported that high levels of SREBP-1a (but not SREBP-2) up-regulated StAR promotor activity, although changes in culture conditions expected to alter mature SREBP-1a levels did not affect StAR promotor function. In the present study, it is possible that increased expression of StAR mRNA seen in the presence of hCG (Fig. 5) could have been affected by concomitant increases in the expression of SREBP-1c. However, increased levels of SREBP-1c induced by insulin (Fig. 5) did not up-regulate StAR transcription (Fig. 4). Our evidence therefore casts doubt on whether SREBP isoforms serve to act as transcriptional modulators of StAR expression under changing physiological situations in granulosa cells.
The potential importance of the hormonal regulation of SREBP-1c was investigated in terms of fatty acid synthase, an important regulatory enzyme, controlled by SREs, which is involved in the control of fatty acid synthesis. Our effects on fatty acid synthase are novel and fit into a wider picture of a range of SRE-responsive genes up-regulated by SREBP isoforms either via cAMP-responsive or insulin-sensitive mechanisms. For example, a recent report by Sekar and Veldhuis (31) showed the interaction of insulin with gonadotropin to amplify low-density lipoprotein-receptor expression in granulosa cells, and this effect was mediated via SREBP activity. Our observed increases in the expression of fatty acid synthase imply an up-regulation of fatty acid synthesis under the hormonal conditions tested. The functional significance of this is substantiated by the observed significant increase in total fatty acid content of granulosa cells when exposed to insulin alone. This further underlines our case for considering the events occurring in the ovary consequent on exposure to high insulin, as being analogous to the situation in liver in which a switch toward expression of SREBP-1c causes accumulation of lipid (5). The additional possibility that release of intracellular arachidonic acid after insulin action may lead to further changes in cellular metabolism through release of prostanoids will require further investigation.
Taken together, our results indicate a major shift in the balance of SREBP isoforms toward SREBP-1c in human granulosa cells as a result of gonadotropin and insulin action. Our particular findings refer to work on granulosa cells luteinized in culture (see introductory text) so that further work may be required to extend these studies to preovulatory granulosa cells within developing follicles. Nevertheless, our unique data are consistent with the view that ovarian cells follow the generally accepted model that describes the SREBP-1c isoform as the version susceptible to hormonal modulation [see Horton et al. (5)] and pose a dilemma in that a switch toward an isoform favoring fatty acid synthesis is occurring in a tissue in which the greater need would appear to be for cholesterol production as a substrate for steroidogenesis. Under conditions of clinical insulin resistance, hepatic SREBP-1c continues to be stimulated by high circulating insulin (5). If this were to occur in the steroidogenic cells of the ovary, we suggest that clinical states involving insulin resistance (such as PCOS) could lead to changes in ovarian cellular function associated with high expression of SREBP-1c and associated overproduction of fatty acids and their metabolites.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online March 15, 2005
Abbreviations: HBSS, Hanks’ balanced salt solution; hCG, human chorionic gonadotropin; IGFBP, IGF binding protein; IVF, in vitro fertilization; PCOS, polycystic ovary syndrome; SRE, sterol regulatory element; SREBP, SRE-binding protein; StAR, steroid acute regulatory protein.
Received October 18, 2004.
Accepted March 3, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-, ß- and 
-fibrinogen genes in rat liver lobes. Biochem J 321:769–775
—implications of white blood-cells in culture. Obstet Gynecol 84:121–127This article has been cited by other articles:
 |
|
 |
|
  B. Vasarhelyi, A. Cseh, I. Kocsis, A. Treszl, B. Gyorffy, and J. Rigo Jr Three mechanisms in the pathogenesis of pre-eclampsia suggested by over-represented transcription factor-binding sites detected with comparative promoter analysis Mol. Hum. Reprod., January 1, 2006; 12(1): 31 - 34. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
