| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
a Daresbury Laboratory, Daresbury, Warrington, Cheshire WA4 4AD, United Kingdom
b Department of Endocrinology and Reproduction, Erasmus University, Rotterdam, The Netherlands
ABSTRACT
Androgen-induced calcium fluxes and gap junctional intercellular communication (GJIC) were studied in three different cell types. A transient (2–3 min duration) increase in intracellular calcium levels was observed within 20–30 sec of androgen addition, which was followed by a plateau phase with steroid concentrations higher than 1 nM. The kinetics of the calcium responses were similar in immature rat Sertoli cells, which contain normal nuclear receptors; the human prostatic tumor cell line, LNCaP, which contains a mutated nuclear receptor; and the human prostatic cell line, PC3, which does not contain a nuclear receptor. The human A431 tumor cell line did not respond to androgens. Concentrations of testosterone and the synthetic androgen, R1881, between 1–1000 pM induced transient calcium increases with ED50 values near 1 pM and 1 nM, whereas dihydrotestosterone (DHT) was not active at these concentrations. At concentrations higher than 1 nM, testosterone, R1881, and DHT were equipotent in stimulating an increase in calcium that lasted for more than 10 min, with ED50 values between 5 and 20 nM. Testosterone covalently bound to albumin was also active, whereas 11 related androstane compounds as well as progesterone and estradiol-17ß were inactive at 1000 nM. The calcium response induced by the three androgens (10 nM) was abolished in all cell types by hydroxyflutamide (1000 nM) and finasteride (1000 nM), but not by cyproterone acetate (1000 nM). The calcium response was also abolished in the absence of extracellular calcium and strongly inhibited by the presence of verapamil. Exposure of the responsive cells to brief (150-sec) pulses of androgens generated calcium responses that were similar to those after continuous exposure. After exposure of Sertoli cells for only 30 sec to 100 nM testosterone, the calcium response lasted for at least 50 min. Although nuclear binding of androgens could be demonstrated, there was no evidence for tight binding to the plasma membrane under similar conditions. When protein synthesis was inhibited, an enhancement of GJIC between rat Sertoli cells, but not between LNCaP cells or PC3 cells, was observed within 15 min of the addition of 10 nM testosterone. Because nuclear androgens are not present in PC3 cells and many functional properties of the responsive system are different from the nuclear receptor in all three cell types, we postulate the existence of an alternative cell surface receptor system with biphasic response characteristics (high and low affinity). The calcium signals are probably coupled to the regulation of gap junctional efficiency between Sertoli cells. The low-affinity receptors may convey complementary androgen signals at elevated local levels such as in the testis, when nuclear receptors are (over)saturated.
calcium, hormone action, prostate, Sertoli cells
INTRODUCTION
Spermatogenesis occurs when testicular levels of androgens are much higher than peripheral androgen levels [1, 2]. Although the discussions about the minimal amount of testosterone that is required for maintenance of spermatogenesis are still ongoing, there is no doubt that concentrations higher than the normal peripheral levels of testosterone are required [3, 4]. Because Sertoli cells are central to the regulation of spermatogenesis, it is possible that specific Sertoli cell functions can be sufficiently stimulated only by these high concentrations of androgens. This hypothesis is supported by the observation that androgen binding protein (ABP) secretion in vitro by a combination of rat Sertoli cells and peritubular myoid cells can be stimulated only by free testosterone doses of between 4 and 20 nM [5], whereas androgen-responsive cells derived from peripheral organs, such as LNCaP cells and cell lines that have been transfected with the androgen receptor, show a full response to androgens between 0.1 and 1 nM [6].
The major mechanism of action of androgens is via high-affinity receptors in the nucleus through activation of DNA transcription. The saturation level of nuclear high-affinity androgen receptors is close to 1 nM, which is in accordance with many observed biological responses. It is difficult to envisage how these high-affinity receptors can sense and respond further to androgen levels above their saturation point. An independent low-affinity transducing system, acting in conjunction with the high-affinity receptors, may explain the odd response characteristics of testicular cells [7]. Supporting such an alternative receptor system is the observation that in vitro spermatogonial multiplication in eels can be stimulated only with 11 keto-testosterone, for which no nuclear receptor has been identified, whereas testosterone and dihydrotestosterone (DHT), for which nuclear receptors have been identified, were not active [8].
In recent years, a variety of membrane actions of steroids have been reported for otherwise nuclear oriented ligands, such as progesterone [9, 10], estrogens [11, 12], corticosterone [13], aldosterone [14, 15], and vitamin D [16, 17]. This is also discussed in a general review by Wehling [18] and, more specifically, for reproductive organs by Revelli et al. [19]. Many of these steroids can induce transient elevations of intracellular calcium within 1 min, and many authors have explained these phenomena by postulating the existence of receptors for steroids in the plasma membrane that differ from the nuclear receptors.
Androgens can also induce rapid calcium fluxes in a variety of cell types, including human prostatic cancer cells (LNCaP; [20]), rat heart myocytes [21], male (but not female) rat osteoblasts [22], and mouse T cells, in which the presence of a functional classical androgen receptor could not be demonstrated [23, 24]. Recently, it was shown that human granulosa cells show a calcium response to androstenedione but not to testosterone [25]. Furthermore, Gorczynska and Handelsman [26] have shown rapid calcium fluxes following addition of high concentrations of androgens to freshly isolated immature rat Sertoli cells. However, the doses of testosterone required for a response (between 300 and 3000 nM) were far in excess of the already high local testicular levels (estimated free testosterone 20–60 nM [2]). This poor response of the freshly isolated cells may have been influenced by membrane damage during the enzymatic isolation procedure, as has been shown for glucocorticoid receptors in the plasma membrane of lymphoma cells [27].
In an effort to understand the mechanisms of action of androgens, we have investigated various characteristics of androgen-induced calcium signaling and as an immediate functional expression of this regulatory pathway, gap junctional intercellular communication (GJIC). Because there are observations that estrogen receptors (perhaps modified) can activate signaling pathways at the plasma membrane [28, 29] and, because suggestions have been made that this could hold also for androgen receptors [30], we have not only investigated rat Sertoli cells, but also the human prostate cancer cell line, LNCaP, which possesses a mutated androgen receptor [6], and the androgen-insensitive human prostate cancer line, PC3, in which mRNA for the androgen receptor could not be detected [31].
MATERIALS AND METHODS
Primary Culture of Seminiferous Tubules
Testes from 21- to 25-day-old Wistar rats (Biological Services Unit, University of Manchester, UK) were collected in 10 ml Dulbeccos PBS containing 1 mM Ca2+ and 1 mM Mg2+. The tunica albuginea was removed and the tissue was incubated in 1 mg/ml collagenase (type 1A, activity >125 collagen digestion units per mg solid; Sigma Chemical Company, St. Louis, MO) in PBS containing 1 mM Ca2+ and 1 mM Mg2+ at 37°C for 30 min. Incubation was terminated by addition of 10 ml culture medium containing 1% fetal calf serum (FCS; Gibco, Paisley, Strathclyde, UK) and pieces of seminiferous tubules were collected after sedimentation at gravity. One drop of a concentrated suspension of tubular fragments was added to a glass coverslip in a culture dish with RPMI 1640 medium without phenol red (Sigma) containing 1% FCS. Tubules were cultured for 1–4 days at 37°C in a humidified atmosphere of 5% CO2. During this period the Sertoli cells migrated away from the explants. Close to the remnants of the tubular fragments, clusters of germ cells remained attached to the flat monolayer of Sertoli cells and, at the periphery, only a monolayer of Sertoli cells existed [32].
Culture of Cell Lines
LNCaP cells [33] and PC3 cells [34] were grown in normal RPMI 1640 medium without phenol red (Sigma) supplemented with 7.5% FCS. The human epidermal carcinoma cell line, A431 [35], was grown in normal DMEM without phenol red (Sigma) supplemented with 10% FCS. All cells were incubated for 2–3 days at 37°C in a humidified atmosphere of 5% CO2 before they were used for measurements.
Chemicals
Testosterone (4-androsten-17ß-ol-3-one), androstenedione (4-androstene-3,17-dione), epitestosterone (4-androsten-17
-ol-3-one), 5ß-dihydrotestosterone (5ß-androstane-17ß-ol-3-one), 5-dihydrotestosterone (5-androstane-17ß-ol-3-one), 5-androst-16-en-3-ol, estradiol-17ß (1,3,5(10)-estratriene-3,17ß-diol), progesterone (4-pregnene-3,20-dione), 11ß-hydroxyandrostenedione (4-androsten-11ß-ol-3,20-dione), dehydroepiandrosterone (5-androsten-3ß-ol-17-one), 5ß-androstan-3ß-ol-17–one, 5-androstane-3,17-dione, 5-androst-3-one, 5-androst-16-en-3-one, and 5-androst-16-en-3ß-ol were all obtained from Steraloids (Newport, RI). R1881 (17ß-hydroxy-17-methyl-estra-4,9,11-trien-3-one) was a gift from Roussel UCLAF (Romainville, France). Cyproterone acetate and hydroxyflutamide (HF) were gifts from Schering (Berlin, Germany). Human recombinant hFSH (Org 32489) was a gift from Organon (Oss, The Netherlands). Finasteride, testosterone 3-(O-carboxy methyl)oxime:BSA (29 mol steroid/mol BSA), testosterone 3-(O-carboxy methyl) oxime BSA-fluorescein isothiocyanate conjugate (FITC), bovine albumin-fluorescein isothiocyanate, epidermal growth factor (EGF), cycloheximide, lucifer yellow CH, and the ionophore A23187 were obtained from Sigma. All compounds were dissolved in ethanol. Fluo 3 and Fura Red acetoxymethyl (AM) ester (Molecular Probes, Leiden, The Netherlands) were dissolved in DMSO. All dilutions were made in buffer solutions so that the final concentration of ethanol or DMSO was less than 0.1%.
Loading With Fluo 3 and Fura Red AM Esters
Cultures were washed twice with a buffer containing 130 mM NaCl, 5 mM KCl, 1 mM Na2HPO4, 1 mM CaCl2, 1 mM MgCl2, and 25 mM Hepes (pH 7.4). Cells were loaded with the calcium-sensitive dyes by incubation with 3 µM Fluo 3 and 3 µM Fura Red AM esters for 1 h in the buffer at 37°C. Subsequently, the cultures were washed three times with buffer and the glass coverslip carrying the cells was mounted in a temperature-controlled microperfusion chamber constructed of aluminum with a volume of 1 ml. It was possible to rapidly change conditions with a perfusion rate of 2 ml/min. At a constant perfusion, a steady state concentration of ligands could be maintained. This is especially important when ligands can be metabolized by the cells. In some experiments, cells were preincubated for 20 min with inhibitors prior to calcium measurements.
Ratiometric Measurement of Calcium
The fluorescence intensity from a 100-µm2 area, corresponding to approximately 10 fluorescently loaded cells, was measured using a high-sensitivity low-irradiance microfluorimeter [36]. Fluo 3 and Fura Red [37–39] were excited at 488 nm using light from the Daresbury Laboratory synchrotron radiation source, which provided long-term stability and low irradiance. Fluorescence emissions were simultaneously recorded at 525 nm and 660 nm. Fluorescence intensities were integrated every 2 sec for periods of 15 min with hormones being added 2 min after the start of the measurement. The data were analyzed using in-house data analysis software (http://www.dl.ac.uk/SRS/NCD/manual.otoko.html) to generate the Fluo 3/Fura Red fluorescence ratios. Because ratios could be calculated immediately, a group of cells that gave a constant ratio could be easily positioned under the objective lens.
Correlation of the Fluo 3/Fura Red fluorescence ratios to absolute calcium concentrations was performed using a 2-point in situ calibration method similar to that described by Foskett et al. [40]. The minimum fluorescence ratio (Rmin) for Fluo 3 and Fura Red in the absence of calcium was determined from the fluorescence of Fluo 3/Fura Red-loaded cells exposed to 5 µM ionophore A23187 in a zero Ca2+ buffer containing 10 mM EGTA. Normal intracellular calcium levels were restored within 30 sec when cells exposed to zero calcium were perfused again with normal buffer containing 1 mM calcium in the absence of ionophore. This indicates that the effects of the ionophore under these conditions are reversible. The maximum fluorescence ratio (Rmax) was determined from the fluorescence of cells exposed to 5 µM ionophore A23187 in a 10 mM Ca-EGTA buffer with 600 nM free-Ca2+.
Binding of Androgens
Binding of tritium-labeled R1881 was measured in intact cells as described previously [41]. With this technique, specific, low-capacity binding to the nuclear receptor can be discriminated from nonspecific high-capacity binding to cellular proteins by comparing cellular binding in the absence and presence of an excess, unlabeled steroid.
Scrape Loading-Dye Transfer Technique
Cultures were washed in PBS and incubated in 0.05% lucifer yellow CH in PBS. Scrape lines were made with a needle in the cell monolayer and the cultures were left undisturbed for 3 min. The cultures were washed several times with PBS and examined with a Zeiss epifluorescence microscope. The fluorescent dye enters the primary cells on either side of the scrape line and is subsequently transferred through gap junctions into adjoining cells. The extent of dye transfer from the primary cells to the adjoining cells gives a measure of GJIC [42]. The percentage of fluorescently labeled cells in a defined area was used as a quantitative measure of GJIC in control and treated cells.
Data Analysis
Measurements of Ca2+ are presented as mean values ± SEM of n independent experiments, each performed in triplicate. Significance of differences was determined by an unpaired Student's t-test and the differences were considered significant if P < 0.001.
RESULTS
Measurement of Calcium and Effects of Protein Hormones
At the start of the experiment, the Fluo 3/Fura Red fluorescence ratio values varied between 0.3 and 1.3. This could indicate either cellular heterogeneity in calcium levels or variations in dye loading. Therefore, various cell areas were prescreened to select a group of approximately 10 cells that gave the standard starting ratio value of 0.5, which was the mean of the distribution.
In order to test the analytical system and the responsiveness of the cells, calcium responses were measured after exposure of the cells to protein hormones because it is known that Sertoli cells respond to FSH and EGF [43], whereas LNCaP [44] and A431 cells [35] respond only to EGF. Figure 1a shows the typical rapid response pattern of Sertoli cells to high concentrations of EGF and FSH: a transient calcium peak followed by a plateau phase higher than basal level, but lower than the peak level.
|
Kinetics and Dose-Response Properties of Androgens
Rapid increases in calcium levels were also found following addition of androgens. Typical responses for testosterone are shown in Figure 1b for rat Sertoli cells. Similar responses were seen for LNCaP cells and PC3 cells. A431 cells did not respond to androgens and served as controls. In all responding cells, different kinetics were observed, depending on the dose of androgen used. In general, low concentrations of androgens generated a transient peak lasting 2–3 min after a delay of approximately 20 sec. The lowest dose of testosterone and R1881 to show a response was 1 pM. Higher androgen concentrations caused an increase in the peak height. At concentrations higher than 10 nM testosterone, R1881, or DHT for rat Sertoli cells and 1 nM testosterone, R1881, and DHT for LNCaP and PC3 cells, a persistent elevation of calcium after the transient peak was observed. The response pattern of cells stimulated with DHT was different from that of testosterone or R1881, because concentrations of DHT below 1 nM did not affect intracellular calcium. We constructed two dose-response curves from the kinetic data of each cell type, one describing the transient peak levels at approximately 2 min after hormone addition, and one describing the dose-dependent plateau levels after approximately 10 min (Fig. 2). The dose-response curves of the three cell types were very similar and a biphasic stimulation pattern could be recognized. Approximate ED50 values of the two phases of stimulation in the three cell types are given in Table 1.
|
|
Steroid Specificity
The specificity of the response system in the three cell types was tested by challenging the cells with various compounds at a concentration of 1000 nM and measuring the intracellular calcium levels for a period of 15 min. The androstane-derived compounds epitestosterone, 5ß-dihydrotestosterone, androstenedione, 11ß-hydroxyandrostenedione, dehydroepiandrosterone, 5-androst-16en-3-ol, 5ß-androstan-3ß-ol-17–one, 5-androstane-3,17-dione, 5-androst-3-one, 5-androst-16-en-3-one, and 5-androst-16-en-3ß-ol did not stimulate calcium signaling. Similarly, progesterone and estradiol-17ß; two antiandrogens, cyproterone acetate (CPA) and HF; and the inhibitor of 5-steroid reductase, finasteride, had no effect on calcium levels.
Inhibition of Response
Inhibitory effects of various compounds were investigated by preincubating the cells with 1000 nM HF or CPA for 20 min prior to perfusion with 10 nM testosterone. The same responses were observed in the three cell types: HF completely blocked the androgen-induced calcium increases, whereas CPA did not interfere with the calcium signaling system. Finasteride, a specific inhibitor of 5-steroid reductase [45], inhibited the response induced by 10 nM testosterone, R1881, and DHT when present at 1000 and 100 nM, but not at 10 nM (Fig. 3, Table 2).
|
|
Interactions of Androgens with the Cell Surface
To determine whether testosterone could induce a calcium response directly at the cell surface, testosterone covalently bound to BSA (T:BSA), which cannot enter the cells, was used. Figure 4, a and b, shows that the calcium response to 10 nM testosterone in Sertoli cells is the same for unbound testosterone and for testosterone covalently bound to albumin. No calcium response was found to unconjugated albumin. When Sertoli cells and LNCaP cells were perfused with T:BSA conjugated to FITC, an increase in fluorescence intensity was seen when T:BSA-FITC was in the perfusion chamber, but the fluorescence disappeared completely after washing out of the conjugate (Fig. 4c). BSA conjugated to FITC but without testosterone showed the same result (Fig. 4d), indicating that the presence of testosterone has no effect on the retention of the albumin conjugate on the plasma membrane. It was not possible to evaluate the biological activity of the FITC-labeled T:BSA because the FITC fluorescence made it impossible to measure the fluorescence of the calcium dyes. Thus, transient and reversible interactions between the androgen conjugate and the plasma membrane appear sufficient for generating at least the transient response.
|
We have also investigated interactions between unconjugated steroids and the plasma membrane of intact LNCaP cells by using tritiated R1881 and classical conditions for binding studies in intact cells. Because CPA can bind to the nuclear androgen receptor but does not interfere with the calcium signaling pathway, it was used to suppress binding of R1881 to the nuclear receptor without influencing the membrane system. In the presence of 1000 nM CPA, specific binding of 3H R1881 after 1 h incubation could not be demonstrated. This was because residual binding of 5 nM 3H R1881 after 1 h incubation and 15 rapid washings with medium containing 1000 nM CPA could not be discriminated from nonspecific 3H R1881 binding in the presence of 1000 nM unlabeled R1881. However, specific nuclear binding of R1881 was 10–20 times higher than nonspecific binding after equilibration of the cells with 5 nM 3H R1881 for 1 h in the absence of CPA or unlabeled R1881 at either 4 or 37°C. With immunocytochemical techniques the presence of the nuclear receptor could be shown in the LNCaP cells and Sertoli cells, but not in the PC3 cells (data not shown).
Aspects of Calcium Signaling
To study whether the androgen-induced calcium increases are due to an extracellular influx of calcium or to calcium mobilization from intracellular stores, experiments were carried out without calcium in the medium. When 5 mM EGTA was added to the medium, the basal intracellular calcium level was reduced and addition of androgens did not induce an increase in calcium levels (Fig. 5b, Table 3). Following preincubation with a voltage-gated calcium channel blocker, verapamil (100 µM), for 20 min, only a small increase in calcium levels was observed after androgen addition (Fig. 5c, Table 3). These results show that calcium channels in the plasma membrane play an important role in calcium signaling.
|
|
Effects of Pulse Exposures to Testosterone
Because the binding studies showed that androgens interact reversibly with the plasma membrane, we have used the perfusion system to investigate the calcium response after transient exposure (150 sec) to androgens. Under these conditions, the persistent elevation of calcium after the transient peak was only 5.21% ± 1.33, for 10 nM testosterone, and 5.65% ± 1.91, for 100 nM testosterone (n = 6), lower than during continuous exposure to testosterone. The similarities in the response pattern between continuous and pulse exposure is an indication that the response is mainly determined by the initial changes in steroid concentrations. We therefore tested the response after two or three brief 30-sec pulses of 10 nM testosterone (Fig. 6, a and b), and found that the effect of the second or third pulse was less than the initial response. On one hand, the duration of the response after a 30-sec pulse of 10 nM testosterone (Fig. 6d) is transient, and calcium levels return to basal levels within approximately 30 min; on the other hand, a long-term response lasting for at least 50 min was obtained after a 30-sec stimulation with 100 nM testosterone (Fig. 6c).
|
Gap Junctional Communication
At least part of the calcium signaling pathways are coupled to phosphorylation cascades. Because GJIC can be modulated via phosphorylation, we have investigated whether androgens can rapidly affect GJIC. GJIC was measured using scrape loading-dye transfer in Sertoli cells because it is known that Sertoli cells communicate with each other via gap junctions [46]. Limited dye transfer was observed between Sertoli cells in the control situation (dye transfer 5.92% ± 0.36 of the area). The dye transfer was markedly increased 15 min after addition of testosterone to the cells (dye transfer 29.59% ± 1.06 of the area, n = 4, P < 0.0005), indicating an increase in GJIC (Fig. 7). A similar response to testosterone (dye transfer 28.31% ± 0.62 of the area, n = 4, P < 0.0005) was observed in the presence of cycloheximide (100 µM), which completely inhibits protein synthesis. LNCaP cells and PC3 cells showed no GJIC, indicating the specificity of the method.
|
DISCUSSION
The majority of androgen effects are mediated by the nuclear androgen receptor that is already maximally stimulated when free steroid concentrations in the target cells are 1 nM or lower [6]. It is difficult to envisage how these receptors can respond to higher androgen concentrations when they are already saturated. As yet there is no evidence that androgen receptors can produce specific signals depending on the degree of oversaturation. Nevertheless, biological effects at such high concentrations of androgens have been described [1–6, 44]. Thus, a low-affinity sensory system must operate to signal the changes in androgen concentration at these relatively high concentrations [7]. Many so-called nongenomic cell surface effects of steroids, recently described in the literature [8–26], possess this property and this alternative low-affinity sensory system could operate in conjunction with the nuclear high-affinity receptor [1, 3–5]. The results of the present investigation confirm that alternative androgen actions can take place at the plasma membrane level and that there is not only a response to high androgen levels up to 100 nM, but also to androgen concentrations of 1 pM. Moreover, they show a steroid specificity that has certain similarities with the nuclear receptor, but which also shows clear differences, especially with respect to effects of antiandrogens and finasteride.
Steroid Specificity
The biologically active ligands, testosterone, DHT, and R1881, which can act via the nuclear receptor, were also active in rapid calcium signaling. Other steroids such as estradiol-17ß, progesterone, 17-testosterone, 5ß-dihydrotestosterone, androstenedione, 11ß-hydroxyandrostenedione, 5-androst-16en-3-ol, dehydroepiandrosterone, 5- androstane-3,17-dione, 5-androst-3-one, 5-androst-16-en-3-one, 5-androst-16-en-3ß-ol, HF, CPA, and finasteride were not active even when tested at concentrations of 1000 nM. The testosterone-, R1881-, and DHT-induced calcium responses could be inhibited by the androgen antagonist HF; however, in contrast to effects on the nuclear receptor, the testosterone-, R1881-, and DHT-activated calcium transients were not inhibited by CPA but could be inhibited by finasteride, which is a specific inhibitor of 5-steroid reductase [45].
The inhibition of the testosterone-induced effects by finasteride could, in theory, be explained by an inhibition of a very active 5-reductase. However, R1881 is a metabolically stable androgen, testosterone is covalently bound to albumin, and DHT cannot be activated by intracellular 5-reductase. Thus, it appears unlikely that the action of finasteride is due to inhibition of 5-reductase activity. A more likely explanation is that finasteride interacts with the androgen binding site of the receptor at the cell membrane. The results with LNCaP cells also show that the binding site of the membrane system differs from the receptor in the nucleus because CPA, HF, progesterone, and estradiol, which can stimulate cell growth due to a mutation in the classical androgen receptor in this cell line [6], do not activate the receptor in the membrane.
Rapid effects of various androgens on LNCaP cells have been reported previously by Steinsapir et al. [20]. Effects of testosterone and DHT on calcium signaling in freshly isolated Sertoli cells have been reported by Gorczynska and Handelsman [26], but the lowest dose to show a response was 300 nM. Enzymatic and mechanical damage during the isolation procedure of the cells may have caused this low sensitivity because it has been shown that trypsin treatment reduces the membrane binding of monoclonal antibodies and cellular responses to glucocorticoids in lymphoma cells [27]. Gorczynska and Handelsman described that HF and finasteride blocked the effect of testosterone, whereas finasteride could not affect the response to DHT, leading to the conclusion that 5-reduction of testosterone is necessary for the effect. Our data do not support this interpretation. Audy et al. [47] reported calcium fluxes in LNCaP cells induced by estradiol (100 pM–100 nM) and 10 nM DHT within 6 min. Flutamide completely inhibited the androgen response but had no effect on the estradiol response. Their data suggest that separate binding sites on the plasma membrane mediate the responses for the different sex steroids. Our responses showed completely different kinetics and steroid specificity without an effect of estradiol. Different LNCaP sublines could explain these discrepancies.
More than One Receptor
It is difficult to envisage how one binding site can sense and transmit information from a range of androgen concentrations that spans 5 orders of magnitude. A receptor clustering model has recently been proposed for chemotactic receptors on Escherichia coli bacteria to explain that one receptor for aspartate can sense gradients of attractants over 5 orders of magnitude [48]; however, there is no evidence that such a system operates in cells other than bacteria. Our calcium response data are compatible with a two-membrane receptor model in spermatozoa for progesterone, as shown by Luconi et al. [49]. Similar to their observations in spermatozoa with progesterone, we observed in three different cell types comparable biphasic dose-response curves generated by testosterone or R1881, but our responses occurred at approximately 1000-fold lower concentrations (pM and nM, respectively) when compared to progesterone. Similar to the restricted action of derivatives from progesterone on spermatozoa, DHT could activate only the low-affinity site of the responsive membrane system in our cells. It is unlikely that metabolism of DHT caused the absence of low-dose effects because under the perfusion conditions, the extracellular concentration of DHT is kept constant. Moreover, if metabolism had occurred, the entire dose-response curve would have been shifted to the right.
At the present time there is no structural information of the receptor for androgens in the membrane. Results from recent experiments show that classical nuclear receptors (perhaps modified), especially estrogen receptors, can also activate the mitogen-activated protein kinase pathway [28]. It has also been shown that a small fraction (4%) of the nuclear receptor can associate with the plasma membrane and that these membrane-associated receptors also bind the same ligands as the nuclear receptor with a similar high affinity. They can also activate signal transduction pathways via G proteins, although with a different steroid specificity, when compared with the nuclear receptor [29]. For the androgen receptor, there is also evidence for alternative processing. Immunoreactivity of androgen-receptor proteins have been shown in T cells but no androgen binding activity could be shown [23, 24]; however, there was no evidence that such alternatively processed products were present in the plasma membrane of T cells. The membrane receptor in our cells could, in theory, also originate from alternative processing of the nuclear receptor, but the many similarities in the response characteristics independent of the nuclear receptor status; no receptors in PC3 cells, mutated receptors with very different steroid specificity in LNCaP cells, and normal receptors in Sertoli cells, make this possibility very unlikely. Moreover, the interactions between androgens and receptors on the plasma membrane and in the nucleus are so different that, with classical binding techniques (including several washing steps), we could not detect any specific retention of R1881 to the plasma membrane of LNCaP cells, whereas nuclear binding was normal. A high dissociation rate constant for these membrane interactions could explain the poor retention of R1881.
As a result of these findings, we propose that proteins not coded by the mRNA for the nuclear receptor most probably mediate the calcium signaling.
Characteristics of Calcium Signaling
High calcium levels persisted after brief exposure (150 or 30 sec) to high concentrations of androgens when cells were perfused with medium without androgens for periods up to 45 min. Because androgens do not bind tightly to the membrane, this retention of the calcium response occurs without the presence of steroids on the membrane receptors. Thus, a brief exposure to androgens can induce a long-term response. All we can conclude from experiments with calcium channel blockers and zero calcium in the incubation medium is that external calcium is required for the calcium signaling, but we do not know how this activated state is initiated by the presence and maintained in the absence of androgens.
Lieberherr and Grosse [22] described dose-dependent effects of androgens in osteoblasts also greater than 5 orders of magnitude of androgen concentrations, but their dose response curve was bell-shaped. Their responses were also different from ours in other respects: osteoblasts responded within 5 and not 30 sec after administration of the androgen; testosterone and DHT gave similar responses; and the transient calcium response was not inhibited by the absence of extracellular calcium, whereas it was inhibited by inhibitors of phospholipase C. Together with our data, there are thus many different possibilities for activation of calcium transients by cell surface actions of androgens.
Effects on Gap Junctions
We have investigated changes in GJIC between Sertoli cells as a possible downstream event of the calcium signaling pathway because it is known that variations in second messengers such as Ca2+, H+, and cAMP can regulate the efficiency of the gap junctions via phosphorylation of the connexin proteins [50]. Androgens increased gap junctional efficiency between Sertoli cells within 15 min in the absence of protein synthesis. We have interpreted this gap junctional regulation as a nongenomic, functional response of the Sertoli cell to testosterone, which could be mediated via a calcium signal. Pluciennik et al. [51] have shown an increase in GJIC following incubation of rat Sertoli cells with FSH. The same authors also reported an interruption of the junctions after 15 min of exposure to 1–25 µM testosterone propionate or estradiol-17ß propionate [52, 53]. These effects were observed only with high concentrations of steroid esters in the absence of changes in intracellular calcium concentrations, suggesting direct effects of the steroid esters on membrane properties. Thus, there appear to be different modes for regulation of GJIC between Sertoli cells, with androgen-induced calcium signaling pathways as one of the possibilities.
Biological Relevance of Androgen-Induced Calcium Signaling
An important question is whether these androgen-activated signaling pathways can activate specific cellular functions or whether they merge with pathways already under the control of protein hormones or growth factors. As stated in the Introduction, Sertoli cells require high local androgen levels for proper functioning. It is possible that the persistent high calcium levels, which are generated only by high levels of androgens and not by EGF or FSH, are a reflection of such a specific pathway. Such a specific calcium pathway could explain why significant stimulation of ABP secretion by a combination of Sertoli and peritubular myoid cells occurs only when testosterone concentrations are higher than 4 nM [5] and not between 0.1–1 nM, which is the normal sensitivity for androgen-responsive cells [6].
In LNCaP cells, low levels of androgens stimulate cell growth, whereas concentrations higher than 1 nM (above the affinity constant of the receptor), inhibit cell growth, further stimulate the secretion of prostate specific antigen, and increase the percentage of androgen-receptor-positive cells [54]. Other investigators [6, 44, 55] using LNCaP cells have also observed growth-inhibitory effects of R1881 and DHT when present at concentrations higher than 1 nM, whereas HF, CPA, progesterone, and estradiol-17ß all stimulate cell growth even at 100 nM. The androgen receptor is essential for this growth inhibition because androgen-dependent growth inhibition also occurs in MCF-7 cells after transfection with androgen receptors, and is maximal at concentrations of R1881 below 1 nM [56]. It is puzzling that inhibition of cell growth in LNCaP cells requires higher concentrations than 1 nM R1881, testosterone, or DHT and that other ligands that stimulate the mutated androgen receptor do not induce growth inhibition. Because we have shown that only R1881, testosterone, and DHT stimulate intracellular calcium in LNCaP cells, we hypothesize that high intracellular calcium somehow plays a role in the androgen receptor-mediated growth inhibition. The growth response of muscles to pharmacological doses of androgens (drug abuse), is also difficult to explain with the properties of the high-affinity nuclear receptor alone [57, 58]. This hypertrophy of muscle cells may thus also require a complementary low-affinity response system in addition to the high-affinity androgen receptors in the nucleus.
It is known that membrane receptors for proteins can modulate different nuclear steroid receptors, they can even activate the steroid receptors without the presence of ligand [59]. The importance of such external regulatory networks for the outcome of steroid hormone action in the nucleus was recently stressed by O'Malley et al. [60], when they proposed that membrane transduction pathways, activated by growth factors, which interact with the nuclear receptor via intracellular phosphorylation cascades, may "set the nuclear receptor thermostat" for responses to steroids. In a similar fashion, the androgen-induced calcium signaling pathway can also influence the functional properties of the nuclear androgen receptor. To investigate this possibility for fine-tuning of the androgen receptor, selective agonists and more knowledge on the structure and function of the putative membrane receptor or receptors is necessary.
NOTE ADDED IN PROOF
Important new information on variability of hormone induced calcium signaling can be found in a multi-author commentary: "Calcium confusion: is the variability in calcium response by Sertoli cells to specific hormones meaningful or simply redundant?" by FFG Rommerts, FM Lyng, E von Ledebur, L Quinlan, GR Jones, JB Warchol, M Stefanini, N Ravindranath, M Joffre. J Endocrinol; 2000.
ACKNOWLEDGMENTS
We are grateful to Dr. M.L. Martin and Dr. M.J. Tobin at Daresbury Laboratory for their expert technical advice, and to Dr. F.H. de Jong for suggestions on the manuscript.
FOOTNOTES
1 This work was generously supported by the Ernst Schering Research Foundation. 
3 Correspondence: F.F.G. Rommerts, Department of Endocrinology and Reproduction, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands. FAX: 31 104 089 461;rommerts{at}endov.fgg.eur.nl
Accepted: April 28, 2000.
Received: June 7, 1999.
REFERENCES
and ERß expressed in Chinese hamster ovary cells. 1999; 13:307–319.-reductase 2 deficiency. Endocr Rev 1993; 14:577–593.This article has been cited by other articles:
|
|
 |
|
  P F Oliveira, M Sousa, A Barros, T Moura, and A Rebelo da Costa Intracellular pH regulation in human Sertoli cells: role of membrane transporters Reproduction, February 1, 2009; 137(2): 353 - 359. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J Kaitu'u-Lino, P. Sluka, C. F H Foo, and P. G Stanton Claudin-11 expression and localisation is regulated by androgens in rat Sertoli cells in vitro Reproduction, June 1, 2007; 133(6): 1169 - 1179. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. C. Gonzalez-Montelongo, R. Marin, T. Gomez, and M. Diaz Androgens Differentially Potentiate Mouse Intestinal Smooth Muscle by Nongenomic Activation of Polyamine Synthesis and Rho Kinase Activation Endocrinology, December 1, 2006; 147(12): 5715 - 5729. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X.-M. Ou, K. Chen, and J. C. Shih Glucocorticoid and Androgen Activation of Monoamine Oxidase A Is Regulated Differently by R1 and Sp1 J. Biol. Chem., July 28, 2006; 281(30): 21512 - 21525. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kampa, C. Kogia, P. A. Theodoropoulos, P. Anezinis, I. Charalampopoulos, E. A. Papakonstanti, E. N. Stathopoulos, A. Hatzoglou, C. Stournaras, A. Gravanis, et al. Activation of membrane androgen receptors potentiates the antiproliferative effects of paclitaxel on human prostate cancer cells Mol. Cancer Ther., May 1, 2006; 5(5): 1342 - 1351. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. H Walker and J. Cheng FSH and testosterone signaling in Sertoli cells Reproduction, July 1, 2005; 130(1): 15 - 28. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Fix, C. Jordan, P. Cano, and W. H. Walker Testosterone activates mitogen-activated protein kinase and the cAMP response element binding protein transcription factor in Sertoli cells PNAS, July 27, 2004; 101(30): 10919 - 10924. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Braun and P. Thomas Biochemical Characterization of a Membrane Androgen Receptor in the Ovary of the Atlantic Croaker (Micropogonias undulatus) Biol Reprod, July 1, 2004; 71(1): 146 - 155. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Braun and P. Thomas Androgens Inhibit Estradiol-17{beta} Synthesis in Atlantic Croaker (Micropogonias undulatus) Ovaries by a Nongenomic Mechanism Initiated at the Cell Surface Biol Reprod, November 1, 2003; 69(5): 1642 - 1650. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Estrada, A. Espinosa, M. Muller, and E. Jaimovich Testosterone Stimulates Intracellular Calcium Release and Mitogen-Activated Protein Kinases Via a G Protein-Coupled Receptor in Skeletal Muscle Cells Endocrinology, August 1, 2003; 144(8): 3586 - 3597. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. LOSEL, E. FALKENSTEIN, M. FEURING, A. SCHULTZ, H.-C. TILLMANN, K. ROSSOL-HASEROTH, and M. WEHLING Nongenomic Steroid Action: Controversies, Questions, and Answers Physiol Rev, July 1, 2003; 83(3): 965 - 1016. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Papakonstanti, M. Kampa, E. Castanas, and C. Stournaras A Rapid, Nongenomic, Signaling Pathway Regulates the Actin Reorganization Induced by Activation of Membrane Testosterone Receptors Mol. Endocrinol., May 1, 2003; 17(5): 870 - 881. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Heinlein and C. Chang The Roles of Androgen Receptors and Androgen-Binding Proteins in Nongenomic Androgen Actions Mol. Endocrinol., October 1, 2002; 16(10): 2181 - 2187. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Risley, I. P. Tan, and J. Farrell Gap Junctions with Varied Permeability Properties Establish Cell-Type Specific Communication Pathways in the Rat Seminiferous Epithelium Biol Reprod, September 1, 2002; 67(3): 945 - 952. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Guo, W. P. M. Benten, J. Krucken, and F. Wunderlich Nongenomic Testosterone Calcium Signaling. GENOTROPIC ACTIONS IN ANDROGEN RECEPTOR-FREE MACROPHAGES J. Biol. Chem., August 9, 2002; 277(33): 29600 - 29607. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
