Androgens Increase Intracellular Calcium Concentration and Inositol 1,4,5-Trisphosphate and Diacylglycerol Formation via a Pertussis Toxin-sensitive G-protein*

Bone is a target tissue of androgens, but the mechanisms by which they act on bone are still unclear. This study examines the early (5-60 s) effects of 1 PM to 1 J ~ M testosterone on cytosolic free Ca2+ concentration ([Ca2+li) and inositol 1,4,5-trisphosphate (InsP3) and dia- cylglycerol (DAG) formation in confluent male rat osteoblasts. 10 PM to 10 m testosterone increased [Ca2+li within 5 s via Ca2+ influx as shown by the effects of EGTA and the Ca2+ channel blockers nifedipine and verapamil and via Ca2+ mobilization from the endoplasmic reticulum as shown by the effects of thapsigargin and neo- mycin. 10 PM to 10 m testosterone increased InsP3 and DAG formation within 10 s. Testosterone immobilized on bovine serum albumin (testosterone (0- carboxymethy1)oximehovine serum albumin) and its de-rivative, (0-carboxymethyl)oxime, rapidly increased [Ca2+Ii and InsP3 and DAG formation and were full ago- nists, although they were less potent than the free steroid. Cyproterone acetate, a nuclear antagonist, did not block the increase in [Ca2+Ii and on [Ca2+l, of confluent male rat osteoblasts. Second, we investigated whether the action of these steroids on [Ca2+li was due to an influx of Ca2+ from extracellular milieu and/or to Ca2+ mobilization from intra- cellular stores. Two types of blocking experiments were performed. 1) A small excess of EGTA (2 mM) was added to the cuvette medium (18). Replenishment of Ca2+ to a 1.25 m Ca2+ excess following EGTA treatment the basal level. 1 min after EGTA addition, when a new steady-state level of [Ca2+], had (18, 19), was added. 2) The selective blockers of CaZ+ entry, nifedipine and verapamil, were added to give a final concentration of 1 PM. They induced a decrease in [Ca2+l, by blocking Ca2+ entry via voltage-dependent Ca2+ channels. Testosterone was added 1 min after. Third, we examined what part of the [Ca2+l, transient was due to Ca2+ release from intracellular stores. The naturally occurring sesquiterpene lactone thapsigargin was used to inhibit the endoplasmic reticulum ATP-dependent Ca2+ pump in the steroid CMOBSA preparation, the steroid CMOBSA preparation was treated with charcoal to remove noncovalently bound steroid or steroid CMO (9). The results showed that charcoal treatment had no effect on the ability of steroid CMOBSA to increase [Ca2+l, (data not shown). The observed effects of steroid CMOBSA on [CaZ+], were due to covalently bound steroid and not to free steroid or steroid CMO in the steroid CMOBSA preparation.

Bone is a target tissue of androgens, but the mechanisms by which they act on bone are still unclear. This study examines the early (5-60 s ) effects of 1 PM to 1 J~M testosterone on cytosolic free Ca2+ concentration ([Ca2+li) and inositol 1,4,5-trisphosphate (InsP3) and diacylglycerol (DAG) formation in confluent male rat osteoblasts. 10 PM to 10 m testosterone increased [Ca2+li within 5 s via Ca2+ influx as shown by the effects of EGTA and the Ca2+ channel blockers nifedipine and verapamil and via Ca2+ mobilization from the endoplasmic reticulum as shown by the effects of thapsigargin and neomycin. 10 PM to 10 m testosterone increased InsP3 and DAG formation within 10 s. Testosterone immobilized on bovine serum albumin (testosterone (0-carboxymethy1)oximehovine serum albumin) and its derivative, (0-carboxymethyl)oxime, rapidly increased [Ca2+Ii and InsP3 and DAG formation and were full agonists, although they were less potent than the free steroid. Cyproterone acetate, a nuclear antagonist, did not block the increase in [Ca2+Ii and InsP, and DAG formation induced by testosterone. Finally, neomycin and pertussis toxin totally abolished the effects of testosterone on InsP, and DAG. These results suggest that male rat osteoblasts bear nongenomic unconventional cell-surface receptors for testosterone that belong to the class of the membrane receptors coupled to a phospholipase C via a pertussis toxin-sensitive G-protein.
The sex steroids (androgens and estrogens) are major regulators of bone metabolism in males and females, respectively (1). Both hormones interact with growth hormone in the control of adolescent growth spurt (2). In males, hypogonadism is associated with bone loss, which is stabilized by testosterone administration (3). Recently, androgen receptors have been identified in osteoblasts (4). But the mechanisms by which androgens exert their effects on bone are still unclear.
Most of the known effects of steroid hormones are mediated by receptors in the cell nucleus that, upon ligand binding, act to modulate the transcriptional activity of the responsive cells.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$To whom correspondence should be addressed: CNRS URA 583, Tour Lavoisier 6eme Etage, HBpital des Enfants Malades, 149, rue de Sevres,75015 Paris, No information is available about the early effects of androgens on intracellular calcium and on the turnover of membrane phosphoinositides, except one work showing an increase in intracellular calcium levels via Ca2+ influx through L-type channels in a prostatic cell line (LNCaP) (14).
We examined the early (5-60 s ) effects of androgens on intracellular calcium concentration and phospholipid metabolism of male rat osteoblasts. Testosterone covalently bound to high molecular weight molecules (7, 91, which did not enter the cell, were also used to examine the possible involvement of a plasma membrane receptor.
Isolation and Cell Culture-Osteoblasts were isolated from parietal bones of 2-day-old male rats by sequential enzymatic digestion (15). These cells had the following osteoblast characteristics: high alkaline phosphatase activity, high type I collagen synthesis, a CAMP response to parathyroid hormone, and an osteocalcin response to 1,25-dihydroxyvitamin DS. Osteoblasts were grown on rectangular glass coverslips (16) for 4 days or in Petri dishes (25 ern2) in phenol red-free a-minimal essential medium with 10% FCS. Cells were then incubated for 72 h in phenol red-free medium containing 1% heat-inactivated FCS and then transferred to serum-free medium 24 h before use.   The Fura-2 fluorescence response to intracellular calcium concentration ([Ca2+li) was calibrated from the ratio of 340l380 nm fluorescence values after subtraction of the background fluorescence of the cells at 340 and 380 nm as described by Grynkiewicz et al. (17). The dissociation constant for the Fura-2Ca2+ complex was taken as 224 ILM (17). The values for R,,, and Rmi, were calculated from measurements using 25 p~ digitonin and 4 mM EGTAand enough Tris base to raise the pH to 8.3 or higher. Each measurement on Fura-2-loaded cells was followed by a parallel experiment under the same conditions with non-Fura-2-loaded cells.
First, we studied the direct effects of androgens (1 PM to 1 y) on [Ca2+l, of confluent male rat osteoblasts. Second, we investigated whether the action of these steroids on [Ca2+li was due to a n influx of Ca2+ from extracellular milieu and/or to Ca2+ mobilization from intracellular stores. Two types of blocking experiments were performed. 1) A small excess of EGTA (2 mM) was added to the cuvette medium (18).
Replenishment of Ca2+ to a 1.25 m Ca2+ excess following EGTA treatment restored the basal level. 1 min after EGTA addition, when a new steady-state level of [Ca2+], had been reached (18,19), testosterone was added.
2) The selective blockers of CaZ+ entry, nifedipine and verapamil, were added to give a final concentration of 1 PM. They induced a decrease in [Ca2+l, by blocking Ca2+ entry via voltage-dependent Ca2+ channels. Testosterone was added 1 min after. Third, we examined what part of the [Ca2+l, transient was due to Ca2+ release from intracellular stores. The naturally occurring sesquiterpene lactone thapsigargin was used to inhibit the endoplasmic reticulum ATP-dependent Ca2+ pump and to release Ca2+ from the associated store (20,21). Cell Labeling-The action of androgens on the generation of inositol phosphates was measured in cells incubated for 72 h with or without myo-[2-3Hlinositol(10 pCi!ml) in phenol red-free medium containing 1% heat-inactivated FCS. Its action on the formation of diacylglycerol (DAG), phosphatidic acid (PA), and monoacylglycerol (MAG) was assessed in cells incubated with ['4Clarachidonic acid (0.25 pCi/ml) for the last 24 h of the 72-h incubation in phenol red-free medium with 1% heat-inactivated FCS. The labeled cells were washed five times with serum-free medium and incubated a t 37 "C in fresh medium without heat-inactivated FCS for 4 h; ethanol solvent (0.01%) or androgens (0.1 PM to 10 m) were then added for 5-120 s.
Lipid Extraction and Chromatography-For lipids, the reaction was stopped by removing the medium and adding cold methanol. The lipids were extracted according to Bligh and Dyer (23) with a final amount of 2 ml of methanol, 2 ml of chloroform, and 1.6 ml of aqueous salt solution (0.74% KCI, 0.04% CaCI2, 0.034% MgC1,). The first chloroform extract was removed, and the remaining methanovwater phase was acidified with 10 mM HCl (final concentration). Phospholipids were extracted at acid pH for the next two steps, except for neutral lipids (DAG and MAG), for which acidification was omitted. The chloroform phases were combined and dried in a rotary evaporator and dissolved in 200 pl of chloroform/methanol(2:1, vlv), and an aliquot was taken for thin-layer chromatography.
Statistical Analysis-The data on the changes in [Caz+l, and the release of inositol phosphates and lipids were analyzed by one-way analysis of variance. The individual contrasts between treatment pairs were made by Dunnett's method. Differences ofp < 0.05 were considered significantly different. Avalue of n represents n different cultures for a specific time and concentration.
Androgens-Androgens were dissolved in ethanol; the final concentration of ethanol never exceeded 0.01%. This ethanol concentration was without effect on intracellular calcium concentration. In inositol phosphates and diacylglycerol experiments, there were no significant differences between the ethanol control values between 5 and 120 s, although the ethanol control values were lower (4%, p < 0.001, n = 25) than those of untreated controls.
To eliminate the possibility that there was no free steroid or steroid CMO in the steroid CMOBSA preparation, the steroid CMOBSA preparation was treated with charcoal to remove noncovalently bound steroid or steroid CMO (9). The results showed that charcoal treatment had no effect on the ability of steroid CMOBSA to increase [Ca2+l, (data not shown). The observed effects of steroid CMOBSA on [CaZ+], were due to covalently bound steroid and not to free steroid or steroid CMO in the steroid CMOBSA preparation. T-CMO and T-CMOiBSA (Fig. 1) induced a smaller increase in [Ca2+Ii; the time course of the T-CMOBSA effect was similar to that of testosterone or T-CMO. The concentration-dependent effects of testosterone were bell-shaped, with maximal activity at 1 n~ (Fig. 2). Testosterone was more potent than either T-CMO and T-CMOBSA, although these latter products were equipotent (Fig. 2). Sa-Dihydrotestosterone was as active as testosterone, while dehydroepiandrosterone had no effect (data not shown). Blockade of Androgen-induced Changes in Intracellular Calcium Concentration-A small excess of EGTA (2 m~) was first used. EGTA caused a marked decrease in basal [Ca2+Ii (52 f 2%, mean '' S.E., n = 25). The calcium entry blockers nifedipine and verapamil(1 p~) triggered rapid drops of 31 f 2 and 41 2 3%, respectively (mean i: S.E., n = 25). The steady-state level reached within 20 s was higher than that obtained with EGTA. 1 n~ testosterone was added 1 min after EGTA or calcium entry blocker ( Table I). EGTA and calcium entry blockers not only diminished (30 f 2%, means S.E., n = 25) the transient increase induced by testosterone, T-CMO, or T-CMOBSA (Table   I), but totally abolished the sustained plateau phase (data only shown for EGTA) (Fig. 3A).

Effects of High K' Buffer on Intracellular Calcium Response
to Testosterone-The membrane depolarization with 25 mM KC1 increased [Ca2+Ii in osteoblasts (Fig. 3B), and a further addition of 2 mM EGTA brought [Ca2+Ii back to the basal level. The plateau phase induced by 1 nM testosterone was abolished

Time (seconds)
when the steroid was added 30 s after EGTA. Pretreatment of the PTX-treated cells with neomycin totally Characterization of Intracellular Organelle Responsible for abolished the remaining first increase, but did not modify the Calcium Increase-Thapsigargin, which modifies calcium sequestration by the endoplasmic reticulum, was used at 100 nM as this concentration had the greatest effect on [Ca2+],. Fig. 4A shows the [Ca2+li response to thapsigargin and testosterone. The rise in [Ca2+li induced by thapsigargin reached a peak within 90 s and then slowly decayed. Testosterone was added 10 min after thapsigargin. Pretreatment with thapsigargin totally abolished the increase, but not the plateau phase. The responses to T-CMO and T-CMOBSA were comparable to those to testosterone (data not shown). However, when cells were pretreated with both EGTA and thapsigargin, the [Ca2+lj response to testosterone (Fig. 4B), T-CMO, or T-CMOBSA (data not shown) was totally abolished.
Effects of Pertussis Toxin (PTX) on Intracellular Calcium Response to Testosterone-Osteoblasts were incubated for 24 h with 100 ng/ml PTX. Fura-2/AM loading and [Ca2+Ii measurements were carried out with 100 ng/ml PTX. PTX partially blocked the increase, while the plateau phase was unchanged.
Effects of Cyproterone Acetate on Intracellular Calcium Response to Testosterone-Osteoblasts were incubated for 5,10, or 45 min or 24 h with cyproterone acetate (10 n M and 1 VM), and 1 n~ testosterone was then added. Preincubation with cyproterone acetate did not modify basal [Ca2+Ii and did not inhibit the effects of testosterone on intracellular calcium whatever the incubation time and the concentration of the nuclear antagonist (Table 11).
The formation of lipids was followed by measuring DAG, MAG, and PA. The percentage of radioactivity incorporated into each was 2 4 % for DAG, 0.1-0.3% for MAG, and 1-3% for PA.
The action of testosterone (1 PM to 10 VM) on inositol phosphates and lipid formation was dose-dependent in a bellshaped manner. The bell-shaped dose dependence was found for each inositol phosphate or lipid studied (data not shown, except for InsP3 and DAG) (Fig. 6, A and B ) . Fig. 7 shows the response profiles of InsP,, InsP2, and InsP t o 1 nM testosterone. The InsP4, InsP2, and InsP responses showed one stimulation

Time (seconds)
peak, which was later than that of InsP3: at 30 s for InsP4, 40 s for InsP,, and 40-50 s for InsP. T-CMO (100 nM) and T-CMO/ BSA (100 nM) gave comparable results, but with a lower amplitude (data not shown). The profiles of MAG and PA formation in response t o l m testosterone are shown in Fig. 8. Effects of Cyproterone Acetate on DAG Response to Androgens-Osteoblasts were incubated for 45 min with cyproterone acetate (10 nM and 1 w), and 1 nM testosterone was then added. Cyproterone acetate did not modify the level of incorporated radioactivity and did not inhibit the testosterone effects on DAG formation (data not shown).
Effects of Neomycin and Pertussis Toxin on InsP3 and DAG Responses to Testosterone-Osteoblasts were preincubated for 1 or 5 min with neomycin (2 mM); 1 r m testosterone was then added. Neomycin, whatever the incubation time, inhibited the increasing effect of testosterone on InsP, and DAG (data not shown).
Osteoblasts were preincubated with 100 ng/ml PTX for 24 h. Incubations with 1 m testosterone were carried out with PTX.
PTX totally abolished the increasing effect of testosterone on InsPB and DAG (data not shown).
Specificity of Testosterone Effects on Intracellular Calcium-1 PM to 100 nM 17P-estradiol and 1 PM to 100 nM progesterone had no effect on [Ca2+lj in male osteoblasts. 1 PM to 100 nM testosterone had no effect on [Ca2+Ij in female osteoblasts, which were isolated from parietal bones of 2-day-old female rats and cultured in the same medium as described above (a-minimal essential medium without phenol red). DISCUSSION This is, to our knowledge, the first study showing very rapid (5-60 s) effects of testosterone on cytosolic free calcium and membrane phospholipid metabolism in male rat osteoblasts. These effects are produced by physiological concentrations as low as 10 PM and are bell-shaped, with a maximum at 1 nM. This bell-shaped dose-dependent action can be compared to that found for calcitriol, the hormonally active form of vitamin D, in different cell types (11,13).
Testosterone triggers a transient increase in [Ca2+li, followed by a sustained plateau phase. Testosterone modulates [Ca2+l, via two mechanisms: Ca2+ influx from the extracellular milieu and Ca2+ mobilization from the endoplasmic reticulum. On the one hand, EGTA, nifedipine (dihydropyridine-type blocker), and verapamil (phenylalkylamine-type blocker) decrease the increase response to testosterone by -30% and totally abolish the sustained response. The testosterone-induced increase in [Ca2+li implicates K+-dependent channels as shown by depolarizing the membrane. These results show that calcium influx occurs via voltage-gated calcium channels, which have been described in these cells (30). They are in agreement with what has been observed in LNCaP cells, in which androgens increase intracellular calcium levels in 1-2 min (14). On the other hand, thapsigargin (which modifies calcium sequestration by the endoplasmic reticulum) and phospholipase C inhibitors (indirect as neomycin or direct as U-73122) partially or totally block the increase without effect on the sustained plateau phase. Here is the first evidence that androgens trigger the release of calcium from the endoplasmic reticulum, which is mediated through phosphoinositide breakdown.
170-Estradiol and progesterone, used from 1 PM to 100 m, have no effect on intracellular calcium, leading to a specific action of androgens in male rat osteoblasts. Moreover, 1 PM to 100 nM testosterone does not elicit any increase in [Ca2+lj in female rat osteoblasts, while 170-estradiol increases [Ca2+], in

Cell Signaling and Androgens in Male Rat Osteoblasts
these cells at concentrations a s low as 1 PM (31). This suggests that the rapid effects of androgens and estrogens in rat osteoblasts are sex-dependent instead of due to the absence or presence of an aromatic A ring. In addition, the rapidity of the androgen effect excludes any possibility of a further conversion of testosterone to estradiol. No information is available on the rapid effect of testosterone on the rapid turnover of membrane phospholipids. This work shows for the first time that testosterone induces a concomitant   10 DM rapid (within 10 s ) increase in the cellular content of inositol 1,4,5-trisphosphate and DAG formation in male rat osteoblasts. The concomitant increase in InsPB and DAG formation plus the inhibition of the two products formed by neomycin and U-73122 provide further support for the activation of the phospholipase C linked to phosphatidylinositol 4,5-bisphosphate. The response of the osteoblasts to testosterone seems to follow the pattern described for agonist-stimulated phosphoinositide turnover in general (32,33). There are also significant increases in InsP,, InsPz, and InsP formation. In addition to the increase in DAG formation, this study also documents rises in MAG and PA levels in response to testosterone. The maximal formation of DAG and MAG preceding that of PA may indicate the sequential action of a specific phospholipase C and DAG kinase.
The data point to direct interactions of androgen-specific membrane steroid recognition moieties. First, androgens immobilized by covalent linkage to BSA, which do not enter the cell, also increase [Ca2+li via calcium influx and calcium mobilization and InsPs and DAG formation. The effects of T-CMO/ BSA are due to covalently bound steroid and not to free steroid or to T-CMO in T-CMOiBSA. Second, direct or indirect inhibitors of phospholipase C totally block the rapid responses induced by testosterone. Third, cyproterone acetate, which com- petes with testosterone at the nuclear level in these cells in inhibiting cell proliferation stimulated by testosterone,' does not inhibit the rapid responses. These results suggest that the androgen receptor responsible for mediating these rapid effects resides on the outer surface of the osteoblasts and is distinct from the intracellular genomic steroid receptor from the standpoint of agonist and antagonist specificity. Although our results provide no conclusive evidence as to the nature of testosterone interactions with the plasma membrane, the time course of the effect on phospholipid metabolism is not compatible with a "liponomic" action. Further research is needed to assess the membrane association of testosterone.
A large number of physiological experiments have led to the postulation that there are two classes of heterotrimeric guanine nucleotide-binding proteins coupling specific receptors to M. Lieberherr and B. Grosse, unpublished data, the activation of phosphoinositide-phospholipase C: one class of G-protein-mediated responses is pertussis-sensitive, and the other is insensitive (34). Preincubation of the osteoblasts with PTX totally abolishes InsPs and DAG formation. The toxin seems to uncouple the androgen nongenomic receptor from its G-protein by blocking the signal transduction that activates the phospholipase C.
Taken together, these data suggest that male rat osteoblasts display membrane nongenomic androgen receptors that belong to a class of membrane receptors linked to intracellular effector coupled to phospholipase C via a pertussis toxin-sensitive Gprotein.
In conclusion, these finding may open up entirely new areas of investigation in the field of bone metabolism since these rapid changes may represent a mechanism whereby osteoblasts integrate different input signals, whereas the genomic nuclear pathway regulates the long-term osteoblast adaptation to the needs of bone turnover.