Nongenomic Testosterone Calcium Signaling

Steroid hormones exert genotropic actions through members of the nuclear receptor family. Here, we have demonstrated genotropic actions of testosterone that are independent of intracellular androgen receptors (iAR). Through plasma membrane androgen receptors (mAR), testosterone induces a rapid rise in the intracellular free Ca2+ concentration of iAR-free murine RAW 264.7 macrophages. This nongenomic testosterone signaling, which is independent of both iAR and estrogen receptors, does not in itself activate either the mitogen-activated protein kinase (MAPK) families ERK1/2, p38, and JNK/SAPK, the stably and transiently transfected c-fos promoter, or NO production. In the context of lipopolysaccharide (LPS) signaling, however, testosterone attenuates LPS activation of the c-fos promoter and NO production, which is abolished by the intracellular Ca2+chelator BAPTA. Testosterone also attenuates the LPS activation of p38 but not that of ERK1/2 and JNK/SAPK, and this attenuation is abrogated by BAPTA. Moreover, the p38 inhibitor, SB 203580, largely reduces LPS activation of the c-fos promoter and NO production, and the remaining levels are no longer regulated by testosterone. This study is the first to provide information on genotropic actions of mAR-mediated nongenomic testosterone Ca2+ signaling by cross-talk with the LPS signaling pathway through p38 MAPK with impact on cell function.

Steroid hormones play a key role as signaling molecules in the development, differentiation, and physiology of multicellular organisms. According to the traditional view, steroids exclusively exert their effects on cells through members of the intracellular nuclear receptor family (1). The classic genotropic mode of action of these ligand-inducible transcription factors is to bind as homo-or heterodimers to specific response elements in target gene promoters, causing subsequent activation or repression of transcription (2). Previous studies (3,4) indicate that hormone-activated nuclear receptors are also able to interact with other transcription factors on target gene promoters without direct binding to DNA. Recent findings (5,6) also indicate that nuclear steroid receptors are able to mediate steroid-induced activation of other signaling molecules, such as the mitogen-activated protein kinase (MAPK) 1 -family ERK1/2 by transcription-independent mechanisms.
Attention has recently been paid to a totally different mode of steroid action, i.e. action through cell surface receptors (7)(8)(9)(10)(11)(12). These receptors in the plasma membranes of cells, though mostly not yet defined in molecular terms, mediate nongenomic actions that have been described for different steroids, including the brassinosteroids of the plant Arabidopsis thaliana (13,14). The latter is remarkable in that the plant genomes do not seem to encode any transcription factors of the nuclear receptor family (15). Nongenomic actions of steroids through surface receptors often become evident as rapid rises in the intracellular free Ca 2ϩ concentration ([Ca 2ϩ ] i ) (16 -20). Nonetheless, the Ca 2ϩ signals, though steroid-specific, generally exhibit low amplitude and short duration so that their consequences for cell functioning are not immediately obvious. Hence, the nongenomic steroid actions are often disregarded as meaningless.
Testosterone has been described as exerting both genotropic and nongenomic actions. Genotropic actions are mediated through the classic intracellular androgen receptor (iAR), which is a 110-kDa protein with domains for androgen binding, DNA binding, and transactivation (21,22). The ligand-activated iAR is not only able to act on transcription but has also been recently reported to exert nongenomic actions such as activation of ERK1/2 and p21-activated kinases (23)(24)(25). By contrast, nongenomic actions of testosterone through membrane androgen receptors (mAR) on cell surfaces become evident as intracellular Ca 2ϩ signaling, which varies with cell type. In murine T-cells, for example, mAR mediate ligandinduced Ca 2ϩ influx through non-voltage-gated, Ni 2ϩ -blockable Ca 2ϩ channels (26,27). In rat osteoblasts, testosterone induces both the influx of extracellular Ca 2ϩ via voltage-gated Ca 2ϩ channels and Ca 2ϩ release from intracellular stores through G-protein-coupled receptors activating phospholipase C via a pertussis toxin-sensitive G-protein (28). Murine macrophages of the cell line IC-21 respond to testosterone with predominantly intracellular Ca 2ϩ mobilization mediated through Gprotein-coupled receptors for testosterone (29,30). However, the functional importance of testosterone-induced nongenomic Ca 2ϩ signaling, in particular with regard to gene expression and cell function, is not yet understood.
To get information for possible genotropic effects of nongenomic testosterone Ca 2ϩ signaling, we have chosen as a model system mouse RAW 264.7 macrophages transiently or stably transfected with a c-fos promoter linked to human secreted alkaline phosphatase (SEAP) as a reporter, because the immediate early gene c-fos is known to be Ca 2ϩ -sensitive (31,32). Here, we have demonstrated that testosterone-induced nongenomic Ca 2ϩ signaling through mAR cannot directly activate the c-fos promoter. However, nongenomic testosterone signaling is able to exert genotropic actions in context with the LPS signaling pathway through p38 MAPK with impact on cell function.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-The macrophage cell line RAW 264.7 (ATCC no. TIB-71) was grown in phenol red-free, low endotoxin Iscove's modified Dulbecco's medium (IMDM) (Invitrogen) supplemented with 5% heat-inactivated, low endotoxin FCS (PAA Laboratories, Cölbe, Germany). The mouse c-fos promoter from Ϫ520 to ϩ109 was cloned into the EcoRI site of pSEAP2-Basic (CLONTECH, Heidelberg, Germany) and either transiently or stably transfected into RAW 264.7 cells using FuGENE 6 (Roche Molecular Biochemicals). Stable transfection was achieved by co-transfection of the cells with pEGFPN3 (CLONTECH) followed by selection of the cells using 250 g/ml geniti-FIG. 1. Testosterone-unresponsiveness of MAPK and c-fos promoter. A, RAW 264.7 cells were treated with 10 nM testosterone (T) for different periods. Protein extracts were subjected to immunodetermination as described under "Experimental Procedures." The upper panels represent the activated forms of p-ERK1/2, p-JNK/SAPK, or p-p38MAPK as detected by anti-phospho-antibodies. The lower panels indicate total ERK1/2, JNK/SAPK, or p38MAPK as detected with antitotal MAPK antibodies. A 30-min stimulation with 10 g/ml anisomycin (AN) was used as positive control. Representative blots are shown, and the results were verified by at least two independent experiments. B, RAW 264.7 macrophages were cultured in IMDM medium containing 5% stripped FCS for 18 h, transiently transfected with c-fos promoter-SEAP, and stimulated with 10 nM testosterone or 1 g/ml LPS for 3 h. The c-fos promoter activity of each group was determined relative to the activity of vehicle-treated control. The results shown are means Ϯ S.E. of at least three independent experiments performed in triplicate. C, RAW-fos13 cells stably transfected with the c-fos promoter-SEAP construct were stimulated with 10 nM testosterone, 5␣-DHT, or 1 g/ml LPS for 3 h.

FIG. 2. Detection of iAR and mAR in RAW-fos13 cells.
A, absence of iAR in RAW-fos13 cells. RT-PCR was performed with the primer pairs ARS1, ARS2, and ARS3 spanning 365, 560, and 281 bp of the steroid-binding domain of iAR, respectively, and the primer pair ARD1 spanning 511 bp of the DNA-binding domain of iAR. The predicted bands in a 2% agarose gel were only revealed in RNA from murine testes. The primer pairs ER␣S1 and ER␣S2 amplified iER␣ fragments of 385 and 608 bp, respectively, with RNA from murine uteri. The negative control ϪRT was RNA from RAW-fos13 cells and the primer pair ER␣S1 without reverse transcriptase. B, flow cytometry of intact RAW-fos13 cells labeled with testosterone-BSA-FITC (T-BSA-FITC) and BSA-FITC. C, confocal laser-scanning microscopy of RAW-fos13 cells labeled with T-BSA-FITC or ConA-rhodamine (ConA-Rh). Bars represent 10 m. cin disulfate. Clone 13 of the stably transfected RAW 264.7 cells (RAW-fos13), which does not express detectable enhanced green fluorescent protein, was used for the experiments (33). All the cells used for steroid stimulation were passaged not more than eight times.
LPS Stimulation of Cells-For stimulation with LPS, cells were incubated in IMDM with 5% stripped FCS for 18 h before adding 1 g/ml LPS from Salmonella typhosa (Sigma) in the presence or absence of steroids for the indicated intervals. FCS stripping was done twice with 1 g of charcoal and 0.1 g of dextran/500 ml of FCS for 30 min at 56 and 37°C, respectively, followed by a final filtration through 0.22-m pore-sized membranes (Nalgen, Wiesbaden, Germany).
Assay for c-fos Promoter Activity-For reporter gene assays, 2.5 ϫ 10 5 RAW 264.7 cells were cultured in 24-well plates overnight and then transiently transfected with the c-fos promoter, SEAP, and the control SV40 SEAP plasmids. LPS and steroid stimulation was performed 24 h post-transfection. For experiments with c-fos promoter stably transfected cells, 5 ϫ 10 5 of RAW-fos13 were cultured in 24-well plates overnight followed by stimulation with LPS and steroids. Just before stimulation, 100 l of the supernatants were collected from each well (RLU 0 h ) and another 100 l after 3 h of stimulation (RLU 3 h ). LPS and steroids were added in a volume of 100 l of medium. SEAP activity was determined using a Phospha-Light kit (Tropix, Weiterstadt, Germany) with a Lumat LB 9507 luminometer (EG&G Berthold, Bad Wildbad, Germany) as described previously (34). The relative activity of the c-fos promoter was calculated by the equation, RLU 3 h Ϫ 0.9 ϫ RLU 0 h , and normalized to controls.
NO Determination-NO was measured as nitrite using the Griess reagent as described previously (34).

Testosterone-unresponsiveness of MAPK and c-fos-Previous
studies (5,23,24) have shown that steroid hormones, including testosterone, are able to activate ERK1/2 in different cell types. We therefore first investigated the possible direct effects of testosterone on ERK1/2 and the two other MAPK families, p38 and JNK/SAPK, in RAW 264.7 macrophages. However, testosterone at the physiological concentration of 10 nM is not able to activate any of the three MAPK families at different time points over 3 h as detected by specific anti-MAPK antibodies used in Western blotting (Fig. 1A). The reason for this testosterone-unresponsiveness of the three MAPK families is not an inherent failure of the cells, because anisomycin (35,36) stimulates phosphorylation of all three kinases ( Fig. 1A; compare also Fig. 4).
We then studied a possible direct effect of testosterone on activation of the c-fos promoter, which is used here only as a molecular marker for early genotropic signaling because the immediate early gene c-fos is inducible by a wide variety of stimuli (37,38). RAW 264.7 cells were transiently transfected with a c-fos promoter linked to a SEAP reporter gene and then stimulated with 10 nM testosterone for 3 h. However, this stimulation is not sufficient to induce any significant response of the c-fos promoter in comparison with non-stimulated control cells (Fig. 1B). In addition, testosterone is not able to induce any change in c-fos promoter activity in the cell clone RAW-fos13, which was derived from RAW 264.7 through a stable transfection with the same reporter construct (see "Experimental Procedures") (Fig. 1C). This unresponsiveness is not due to a possible defect in the ability of the transfected c-fos promoter construct to be stimulated, because LPS causes a significant induction in both transiently and stably transfected cells (Fig. 1, B and C).
Testosterone-induced Ca 2ϩ Signaling through mAR-The non-inducibility of MAPK and c-fos promoter activity by testosterone may be due to the absence of functionally active iAR and/or mAR, which normally mediate the actions of testosterone on cells. We have therefore examined the occurrence of mAR and iAR in RAW 264.7 and RAW-fos13 cells. These cells do not express any significant amounts of iAR as proven by RT-PCR. When primer pairs flanking a region of the DNAbinding domain or three different regions of the steroid-binding domain were used, all four predicted PCR products could be amplified using RNA from mouse testes. Expression of iAR could not be detected with RNA from RAW 264.7 or RAW-fos13 cells, although the quality of the RNA and the cDNA from these cell lines was proven by amplification of the intracellular estrogen receptor ␣ using two different primer pairs (Fig. 2A). These data are consistent with our previous results in IC-21 macrophages (29,30).
For detection of mAR, the plasma membrane-impermeable testosterone-BSA-FITC conjugate was used to label the cells for 1 min, resulting in an increase in cellular fluorescence as analyzed by flow cytometry (Fig. 2B). No binding was detectable with BSA-FITC (Fig. 2B). The fluorescence of the bound testosterone-BSA-FITC was localized exclusively on the cell surface, as revealed by confocal laser-scanning microscopy (Fig.  2C). The surface binding of testosterone-BSA-FITC was also corroborated by colocalization with the red fluorescent cell surface marker, ConA-rhodamine (Fig. 2C). Moreover, the mAR on RAW cells are functionally active. This becomes evident as a testosterone-induced rapid rise in [Ca 2ϩ ] i of Fura-2-loaded RAW-fos13 cells. The increase in [Ca 2ϩ ] i amounts to about 40 -70 nM at 10 nM testosterone (Fig. 3A). Remarkably, 5␣dihydrotestosterone (5␣-DHT) could also induce an increase in [Ca 2ϩ ] i by about the same amount as testosterone, whereas 5␤-dihydrotestosterone (5␤-DHT) and 1-dehydrotestosterone (1-DeHT) were ineffective in raising [Ca 2ϩ ] i (Fig. 3, A and B). Moreover, the testosterone-induced rise in [Ca 2ϩ ] i was prevented neither by the iAR blocker, cyproterone, nor by raloxifene, tamoxifen, or ICI 182,780, which are inhibitors of the intracellular estrogen receptor (iER) through which testosterone could possibly have been acting after aromatization to 17␤-estradiol (Fig. 3, C and D). An effective rise in [Ca 2ϩ ] i was also inducible by the plasma membrane-impermeable testosterone-BSA but not by BSA alone (Fig. 3E). The testosteroneinduced Ca 2ϩ increase is predominantly because of intracellular Ca 2ϩ mobilization rather than Ca 2ϩ influx. This rise in [Ca 2ϩ ] i is mediated by an agonist-sequestrable mAR linked to a pertussis toxin-sensitive G-protein coupled to phospholipase C (data not shown), as recently also described in mouse macrophages of the cell line IC-21 (29,30).
Attenuation by Testosterone of LPS-activated p38 MAPK-LPS stimulation of macrophages is known to activate numer- The RAW-fos13 cells were stimulated with 1 g/ml LPS and 10 nM testosterone for different periods. Protein extracts were subjected to immunodetermination as described under "Experimental Procedures." The upper panels for p-ERK1/2, p-JNK/SAPK, or p-p38 MAPK represent the phosphorylated forms. Lower panels, total ERK1/2, JNK/ SAPK, or p38 MAPK, as detected with the corresponding antibodies. Representative blots are shown. Stimulation of the cells with 10 g/ml anisomycin (AN) was used as positive control. Relative activation of p38 was densitometrically evaluated. The results were verified in at least two independent experiments. B, requirement of intracellular calcium mobilization for testosterone-induced depression of p38 MAPK activation by LPS. RAW-fos13 cells were preincubated with 10 M BAPTA for 10 min and then stimulated for 15 min with 1 g/ml LPS combined with 10 nM testosterone. ous signaling events, including the three MAPK families p38, ERK1/2, and JNK/SAPK (39 -43). In accordance, 1 g/ml LPS activates all three MAPKs in RAW-fos13 cells after 15 min (Fig. 4A). Activation of JNK/SAPK and p38 MAPK declined after 60 min, whereas the decline of ERK1/2 activation is obvious only after 180 min. Conspicuously, however, co-stimulation with 10 nM testosterone significantly attenuates LPSinduced activation of p38 MAPK. The maximal inhibition is about 50% after 15 min. By contrast, testosterone does not affect the LPS-induced activation of ERK1/2 and JNK/SAPK at all. The attenuation of LPS-induced activation of p38 MAPK is dependent on the testosterone-induced rise in [Ca 2ϩ ] i because it is completely prevented by BAPTA (Fig. 4B) under conditions blocking the testosterone-induced Ca 2ϩ increase (Fig. 3F). BAPTA does not influence the LPS-induced activation of p38 MAPK (Fig. 4B).
Testosterone Attenuated LPS Activation of the c-fos Promoter through p38 MAPK-In addition to MAPK, LPS is also known to activate the immediate early gene c-fos in macrophages (44). In accordance, we can show that 1 g/ml LPS induces activation of the c-fos promoter-SEAP construct transiently transfected in RAW 264.7 cells. Similar to the situation with p38 MAPK, 10 nM testosterone attenuates the LPS-induced activation of the c-fos promoter by ϳ35% (Fig. 5A). This depressive effect of testosterone is specific, because under the same experimental conditions 17␤-estradiol (E 2 ) causes an increase in c-fos promoter activity. In addition, we can exclude the hypothesis that the opposite effects of testosterone and E 2 are due to their possible effects on translation and/or secretion of SEAP. Neither testosterone nor E 2 influence the transport of SEAP into medium when SEAP is under the control of the SV40 promoter (Fig. 5B).
The testosterone-attenuated LPS response of the c-fos promoter is also observed in stably transfected RAW-fos13 cells, which are still more sensitive to LPS and testosterone stimulation. Incubation of RAW-fos13 cells with 1 g/ml LPS for 3 h resulted in a 5-fold induction of c-fos promoter activity (Fig.  5C). This induction, however, is significantly reduced by about 60% upon co-incubation with 10 nM testosterone (Fig. 5C). About the same reduction can be induced with 10 nM 5␣-DHT, whereas 10 nM 5␤-DHT and 10 nM 1-DeHT were ineffective in influencing the LPS-induced increase in c-fos promoter activity. Moreover, the iAR blocker, cyproterone, and the estrogen receptor blocker, ICI 182,780, cannot prevent the depressive testosterone effect on LPS-stimulated c-fos promoter activation (Fig. 5C). By contrast, when RAW 264.7 macrophages transiently transfected with the c-fos promoter were preincubated with BAPTA, testosterone lost its ability to exert a depressive effect on the c-fos promoter (Fig. 5D). This indicates that the depressive testosterone effect requires a rise in [Ca 2ϩ ] i , similar to the attenuating effect of testosterone on LPS-induced p38 MAPK phosphorylation. To find out whether the Ca 2ϩ -dependent down-regulation by testosterone of LPS-induced c-fos promoter activation involves the p38 MAPK pathway, we have used SB 203580, a specific inhibitor of p38 MAPK. At 50 M SB FIG. 5. Effects of testosterone on LPS activation of c-fos promoter. RAW 264.7 macrophages were cultured in IMDM medium containing 5% stripped FCS for 18 h and then transiently transfected with c-fos promoter-SEAP (A) or SV40-SEAP (B) for 24 h before stimulation with 1 g/ml LPS in combination with 10 nM testosterone or 1 nM E 2 for 3 h. The c-fos promoter activity of each group was determined relative to the activity of the vehicle-treated control, which was set at 100% for the LPS group. The results shown are means Ϯ S.E. of at least three independent experiments performed in triplicate. C, testosterone attenuation of the LPS-activated c-fos promoter. Cells were stimulated for 3 h with 1 g/ml LPS in combination with 10 nM testosterone, 5␣-DHT, 5␤-DHT, 1-DeHT. Before stimulation the cells were preincubated for 1 h with 1 M cyproterone (Cyp) and 1 M ICI 182,780 (ICI), respectively. c-fos promoter activity from each group was determined and expressed relative to the c-fos promoter activity induced by LPS, which was set at 100%. The results shown are means Ϯ S.E. from at least three independent experiments performed in triplicate. D, cells were first preincubated with 10 M BAPTA for 10 min or 50 M SB 203580 (SB) for 30 min and then stimulated with 1 g/ml LPS or LPS plus 10 nM testosterone for 3 h. Results are presented as in panel C. 203580, LPS stimulation of p38 is reduced by about 50%, and the remaining promoter activity is no longer regulated by testosterone (Fig. 5D). These results indicate that p38 is involved in the LPS-induced activation of the c-fos promoter and that testosterone attenuation of the LPS-stimulated c-fos promoter activity is achieved predominately, if not exclusively, by interference with p38 MAPK phosphorylation.
Down-regulation by Testosterone of LPS-stimulated NO Production through p38 MAPK-Activation of p38 MAPK is known to be important for induction of genes associated with macrophage activation manifesting, for example, as increased production of the key immune effector molecule NO (45). We therefore suspected that the testosterone-induced rise in [Ca 2ϩ ] i could also influence NO production. Testosterone in itself is not able to affect NO production. However, testosterone caused a significant decrease of NO production in LPS-stimulated RAW-fos13 cells (Fig. 6A). This decrease was not sensitive to the iAR blocker, cyproterone (Fig. 6A). The suppressive effect of testosterone can be prevented by BAPTA under conditions blocking the testosterone-induced rise in [Ca 2ϩ ] i (Fig. 6B). Moreover, LPS-induced NO production can be reduced by about 60% with 50 M SB 203580 (Fig. 6B). This is consistent with previous reports that LPS-stimulated NO production of macrophages is mediated through p38 MAPK (46). When LPS-induced NO production is down-regulated by SB 203580, the remaining NO production is no longer regulated by testosterone (Fig. 6B). This indicates that testosterone exerts its attenuative effect on LPS-stimulated NO production largely through p38 MAPK. DISCUSSION Nongenomic actions of steroids on cells are obviously much more complex than anticipated to date. During recent years, a number of reports have appeared describing nongenomic steroid effects that are mediated through the classic intracellular steroid receptors. For instance, iER-mediated nongenomic E 2 effects can activate members of the MAPK family such as ERK1/2 (5,(47)(48)(49) or endothelial NO synthase (50,51) and can induce an increase of [Ca 2ϩ ] i (47). Androgens have been found to activate ERK1/2 (52) and p21-activated kinases (25) by a nongenomic pathway via the classic iAR. Very recently, testosterone has been shown to activate the ERK signaling pathway nongenomically through either iAR or even iER in diverse cells such as osteoblasts, osteocytes, embryonic fibroblasts, and HeLa cells (24). By contrast, the present study has demonstrated that intracellular steroid receptors are not a necessary precondition to mediate nongenomic effects of steroids. Like the macrophage cell line IC-21 (29,30), the murine RAW 264.7 macrophages investigated here do not express any significant amounts of iAR. Nevertheless, in RAW 264.7 macrophages, testosterone is able to induce iAR-independent nongenomic actions, which can in turn even exert genotropic actions with impact on cell function.
Nongenomic testosterone signaling manifests itself as a rapid rise in [Ca 2ϩ ] i in macrophages of the cell line RAW 264.7. Such rises have also been observed in other cell types (26 -28, 53), and they are well known to be inducible by other steroids as well (10,54). In RAW 264.7 cells, similar to IC-21 macrophages (29,30), the testosterone-induced rise in [Ca 2ϩ ] i is not prevented by the iAR blocker, cyproterone. We can also exclude the possible action of testosterone through the iER (24) because the Ca 2ϩ response was not sensitive to the iER blockers, raloxifene, tamoxifen, and ICI 182,780. Rather, our data demonstrate that testosterone exerts its nongenomic iAR-and iERindependent effects through mAR. An increase in [Ca 2ϩ ] i is also inducible with the plasma membrane-impermeable testosterone-BSA, and specific binding sites for testosterone can be localized on the surface of intact RAW-fos13 cells with testosterone-BSA conjugated to FITC. Though still unknown in molecular terms, the mAR mediate actions specific for different androgens; only testosterone and 5␣-DHT induce about the same rise in [Ca 2ϩ ] i , whereas 5␤-DHT and 1-DeHT are ineffective in evoking changes in [Ca 2ϩ ] i of RAW-fos13 macrophages.
The nongenomic testosterone Ca 2ϩ signaling through mAR is not able in itself to induce genotropic actions in terms of activation of the stably and transiently transfected promoter of the Ca 2ϩ -inducible gene c-fos in RAW 264.7 cells. Nevertheless, the nongenomic testosterone signaling has the potency to in- duce genotropic actions. This becomes evident in the context of co-stimulation of the LPS signaling pathway. Thus, LPS stimulates c-fos promoter activation, and this stimulation is attenuated by testosterone-induced Ca 2ϩ signaling. This attenuation can be completely abolished by BAPTA, and it is specific as supported by the following results. First, E 2 causes the opposite effect of testosterone, namely an enhanced activation of the c-fos promoter following LPS stimulation. Second, testosterone has no effect on SV40 promoter activity. Third, 5␣-DHT is just as effective as testosterone in inducing attenuation, whereas 5␤-DHT and 1-DeHT are ineffective. The nongenomic testosterone Ca 2ϩ signaling not only exerts genotropic actions in context with LPS signaling but also has a specific impact on cell function. This manifests itself as a testosterone-induced attenuation of LPS-stimulated NO production of RAW 264.7 macrophages. In accordance, a recent report (55) also has shown testosterone-induced attenuation of LPSstimulated NO production. Our data demonstrate that nongenomic testosterone Ca 2ϩ signaling is the reason for testosterone-attenuated LPS-activated NO production, because the latter can be abolished by BAPTA.
The testosterone-induced attenuation of LPS-induced c-fos promoter activation and NO production may be thought to reflect only a simple uniform dampening of all LPS signaling parameters in macrophages through the increased free Ca 2ϩ ions. However, this attenuation is specific, which has to be deduced from our results. Thus, LPS activates all three MAPK families, ERK1/2, JNK/SAPK, and p38, in accordance with previous reports (39,41). Testosterone, however, selectively down-regulates only the LPS-induced activation of p38, not that of ERK1/2 and JNK/SAPK. This p38 downregulation is indeed due to nongenomic testosterone Ca 2ϩ signaling. When the free Ca 2ϩ ions induced by testosterone are captured by BAPTA, p38 is again fully able to be stimulated by LPS. The central role of p38 MAPK is further substantiated by our finding that the p38 inhibitor, SB 203580, diminishes the LPS-induced activation of c-fos promoter activation and NO production and abolishes the responsiveness to testosterone of the remaining LPS effect. Collectively, our results indicate that there is a cross-talk of the testosteroneinduced nongenomic Ca 2ϩ signaling with the LPS signaling pathway. The testosterone-induced rise in [Ca 2ϩ ] i attenuates the LPS-induced activation of p38, which further downstream attenuates both the LPS-activated c-fos promoter and NO production.
The nongenomic iAR-independent testosterone signaling we have observed here in RAW 264.7 macrophages is a novel paradigm for a nongenomic steroid effect with impact on gene expression and cell functioning, independent of the cognate intracellular steroid receptor. This nongenomic signaling is presumably of importance with respect to the long known immunosuppressive activities of testosterone that are not yet explainable by the classic iAR response. In particular, the testosterone-induced nongenomic decrease in NO production of macrophages is compatible with our previous findings that testosterone dramatically diminishes the ability of mice to eliminate blood stages of Plasmodium chabaudi malaria (56,57). Furthermore, testosterone not only prevents the development of protective immunity against P. chabaudi infections but even impairs the efficacy of protective vaccination against P. chabaudi malaria (58). These immunosuppressive effects of testosterone are mediated neither through iAR nor, after aromatization of testosterone to E 2 , through iER (59,60).