In vivo and in vitro effect of glucocorticoids on fibroblast growth factor (FGF)-2 and FGF receptor 1 expression.

In order to clarify the physiological function of fibroblast growth factor (FGF-2) in the adrenal medulla the regulation of FGF-2 and FGF receptor 1 (FGFR1) was studied in vitro and in vivo in response to glucocorticoids. To assess the effects of glucocorticoids, in vivo extracts of adrenal medulla and adrenal cortex were analyzed by RNase protection assay and Western blot analysis. PC12 cells were chosen as a model system to study the effects of glucocorticoids in vitro. In PC12 cells, dexamethasone (DEX) was found to stimulate dramatically the expression of both FGF-2 mRNA and protein. Western blot analysis revealed that exclusively the 21-kDa FGF-2 isoform was enhanced. In contrast to the FGF-2 mRNA level FGFR1 was not affected by treatment with glucocorticoids. In vivo FGF-2 mRNA level and 21-kDa FGF-2 isoform level are significantly enhanced in the adrenal medulla 24 h after DEX injection. In vivo application of DEX leads to an increase of the medullary and cortical FGFR1 transcript levels. Glucocorticoid effects on FGF-2 expression were not found in adrenal cortex, heart, skeletal muscle, and kidney, respectively, in vivo and in L6 rat myoblasts in vitro. In addition to adrenal medullary cells glucocorticoids elevated the FGF-2 mRNA and protein level also in vivo in the brain and in vitro in immortalized Schwann cells. The present results suggest that the 21-kDa FGF-2 isoform mediates a physiological function specific for neuronal tissue which is modulated by glucocorticoids.

In order to clarify the physiological function of fibroblast growth factor (FGF-2) in the adrenal medulla the regulation of FGF-2 and FGF receptor 1 (FGFR1) was studied in vitro and in vivo in response to glucocorticoids. To assess the effects of glucocorticoids, in vivo extracts of adrenal medulla and adrenal cortex were analyzed by RNase protection assay and Western blot analysis. PC12 cells were chosen as a model system to study the effects of glucocorticoids in vitro. In PC12 cells, dexamethasone (DEX) was found to stimulate dramatically the expression of both FGF-2 mRNA and protein. Western blot analysis revealed that exclusively the 21-kDa FGF-2 isoform was enhanced. In contrast to the FGF-2 mRNA level FGFR1 was not affected by treatment with glucocorticoids. In vivo FGF-2 mRNA level and 21-kDa FGF-2 isoform level are significantly enhanced in the adrenal medulla 24 h after DEX injection. In vivo application of DEX leads to an increase of the medullary and cortical FGFR1 transcript levels. Glucocorticoid effects on FGF-2 expression were not found in adrenal cortex, heart, skeletal muscle, and kidney, respectively, in vivo and in L6 rat myoblasts in vitro.
In addition to adrenal medullary cells glucocorticoids elevated the FGF-2 mRNA and protein level also in vivo in the brain and in vitro in immortalized Schwann cells. The present results suggest that the 21-kDa FGF-2 isoform mediates a physiological function specific for neuronal tissue which is modulated by glucocorticoids.
Basic fibroblast growth factor (FGF-2) 1 is a member of the fibroblast growth factor family currently comprising nine members (1). FGFs are heparin binding proteins with a widespread distribution in mesoderm-and neuroectoderm-derived tissues (2). In addition to the mitogenic function, FGFs are suggested to be involved in embryonic developmental events, like induction and differentiation, in angiogenesis, in promotion of chemotaxis, and in repair (3). For FGF-1 and FGF-2, the first family members which were characterized, the brain is known as one of the richest sources (4). FGF-2 mRNA and protein are present in several regions of the embryonic, postnatal, and adult central nervous system (5)(6)(7). In the peripheral nervous system, where FGF-2 is also found, the mRNA is up-regulated in spinal ganglia and in the peripheral nerve during regener-ation. 2 Although FGF-2 mediates a variety of effects on neurons in vivo and in vitro (4,8), the physiological function of this protein, however, is still not clear.
In the adrenal gland FGF-2 is found in the cortex and medulla (9,10). Chromaffin cells of the adrenal medulla that are modified sympathetic neurons contain FGF-2 protein and its transcript (9 -11). In vivo, immunoreactivity for FGF-2 shows a spatial and temporal pattern in postnatal and adult chromaffin cells (10). FGF-2 immunoreactivity becomes detectable at postnatal day 8 and increases to adulthood (10). The noradrenergic chromaffin cell subpopulation but not the adrenergic subpopulation expresses FGF-2 (10). In cultured bovine chromaffin cells, the FGF-2 mRNA is enhanced after stimulation of neurotransmitter and hormonal receptors (11).
FGFs are known to mediate cellular effects via cell surface receptor tyrosine kinases (FGF receptors, FGFRs). At present, the cDNAs for four high affinity receptors (FGFR1-4), encoded by distinct genes, have been cloned (12,13,14). In addition, low affinity receptors have been characterized as cell surface heparan sulfate proteoglycans, e.g. the integral membrane proteoglycan syndecan (15,16). The mRNAs for FGFR1, FGFR2, and FGFR4 are found in extracts of the adrenal gland (17,18). In the adrenal cortex of the embryonic mouse the FGFR4 mRNA could be demonstrated by in situ hybridization (19). In the rat adrenal gland FGFR1 transcript is present in preparations of the cortex and medulla (20). The developmental expression pattern of the medullary FGFR1 mRNA matches that of the expression of the FGF-2 immunoreactivity in developing chromaffin cells (20).
PC12 cells are a catecholamine-secreting pheochromocytoma cell line cloned from the rat adrenal medulla. Since a great number of cells are available, they are a useful model system for detailed studies. PC12 cells express both FGF-2 and FGFR1 transcript (20). Nerve growth factor (NGF) induces a significant increase of FGFR1 mRNA in PC12 cells under serumcontaining culture conditions suggesting that NGF enhances the responsivity to FGF (20).
With regard to the putative physiological function(s) of the chromaffin cell-derived FGF-2, several possibilities have been proposed. From in vivo studies it was suggested that the medullary FGF-2 could function as a neurotrophic factor for preganglionic sympathetic neurons (21). Exogenously applied FGF-2 was shown to prevent the lesion-induced death of the preganglionic sympathetic neurons of the spinal cord after adrenomedullectomy (21). In line with this result was the finding that preganglionic sympathetic neurons possess FGFR1 and are able to specifically and retrogradely transport iodinated FGF-2 which was injected into the medulla (22). 3 On the other hand, since several FGF-2-mediated effects on cultured chromaffin cells and on PC12 cells were demonstrated, an autocrine/paracrine role of FGF-2 for chromaffin cells has been discussed. FGF-2 stimulates proliferation and neurite outgrowth of embryonic chromaffin precursor cells (23), increases chromaffin cell proliferation of neonatal and adult rats synergistically with insulin-like growth factors (24), promotes catecholamine storage and synthesis of early postnatal chromaffin cells (25), induces the expression of the tyrosine hydroxylase and proenkephalin genes of adult chromaffin cells (26), and enhances in combination with ciliary neurotrophic factor the ability of NGF to induce transdifferentiation of postnatal and adult chromaffin cells into neurons (27). The responses of PC12 cells to FGF-2 include differentiation into sympathetic-like neurons and development of electrical excitability (28 -33).
Glucocorticoids (GCs), which are synthesized in the adrenal cortex and are present at high concentrations in the medulla, are implicated in the development of medullary chromaffin cells (34). It is suggested that GCs are essential and sufficient to trigger the differentiation of noradrenergic sympathoadrenal precursor cells to adrenergic chromaffin cells (35). A recent study on glucocorticoid-deficient mice by gene targeting suggests the existence of two different cell populations in the adrenal medulla, an adrenergic population which requires glucocorticoids for the development and a noradrenergic population which develops in the absence of the glucocorticoid signal (36). Postnatal chromaffin cells in vitro transdifferentiate into sympathetic neurons after exposure to NGF (37,38). This NGF-induced transdifferentiation can be antagonized by GCs which maintain the endocrine status (23, 35, 38 -40). The mechanism, however, which causes the effect of GCs on chromaffin cells is not yet clear.
With regard to the central nervous system, it was shown that GCs regulate the expression of neurotrophins and FGF-2 (41)(42)(43)(44)(45). The aim of the present study was to examine whether GCs which are important for chromaffin cells could be involved in the regulation of FGF-2 expression. Therefore, we have analyzed the in vivo and in vitro effects of GCs on the expression of FGF-2 and FGFR1 in the adrenal medulla and PC12 cells. In addition, we compared the effect of GCs on FGF-2 expression in other neuronal tissues (brain, Schwann cells) with non-neuronal tissues (adrenal cortex, heart, skeletal muscle, kidney, L6 myoblasts).
The present results show that GCs enhance FGF-2 protein and mRNA in PC12 cells in vitro and in adrenal medullary cells in vivo. Since the FGF-2 transcript is also elevated in the brain but not in non-neuronal tissue, we suggest that the GC-mediated increase of FGF-2 mRNA is specific for neuronal tissue.

EXPERIMENTAL PROCEDURES
Cell Cultures-PC12 cells (kindly provided by Dr. R. Westermann, Marburg, Germany), L6 rat myoblasts (kindly provided by Dr. T. Braun, Braunschweig, Germany), and immortalized Schwann cells (46) (kindly provided by Dr. B. Seilheimer, Basel, Switzerland), respectively, were grown to confluency in 75-cm 2 plastic tissue culture flasks (Falcon) in Dulbecco's modified Eagle's medium containing 5% fetal calf serum and 10% horse serum at 37°C in a humidified 6% CO 2 incubator. Cells were routinely split in a 1:3 ratio, and the medium was changed twice a week. For experiments cells (2 ϫ 10 6 ) were seeded on 75-cm 2 flasks in serum-containing (see above) or serum-free (see below) medium. After 1 day medium was changed and cells received the glucocorticoids dissolved in serum-containing or serum-free medium. Serum-free medium was supplemented with N1 additives (47) and 0.25% bovine serum albumin. Hydrocortisone, progesterone, testosterone, and estradiol were presented at 10 M and dexamethasone at 10 M and 1 M. After various culture periods (1-120 h) cells were dislodged by 0.1% trypsin, 20 nM EDTA for RNA isolation, frozen in liquid nitrogen, and stored at Ϫ80°C until used. The number and morphology of cells was checked under microscopic control in sister cultures in the presence of the respective glucocorticoids by counting the cells in suspension, after gentle detachment from the flasks at the end of the experiments.
Treatment of Animals-Female Hanover Wistar rats weighing 180 -200 g were used. Animals received dexamethasone (5 mg/kg intraperitoneal) or vehicle injections and were killed by decapitation 24 h after injection. Tissues were rapidly dissected, frozen in liquid nitrogen, and stored at Ϫ80°C. For the analysis of the adrenal cortex and medulla, adrenal glands of rats from each group were quickly removed. Cortices and medullae were carefully dissected on ice under binocular control, collected in liquid nitrogen, and stored at Ϫ80°C.
Isolation and Analysis of RNA-Total RNA was isolated according to the method of Chomczynski and Sacchi (48) and quantified by absorption at 260 nm. Ribonuclease protection assay was performed as described previously (20). Total RNA was dissolved and heated to 95°C for 10 min in hybridization solution (80% formamide, 40 mM PIPES, pH 6.4, 400 mM NaCl, 1 mM EDTA) containing 100,000 cpm of 33 P-labeled cRNA probe. After hybridization overnight at 49°C the solution was diluted in RNase digestion buffer (10 mM Tris/HCl, pH 7.4, 300 mM NaCl, 5 mM EDTA) containing RNases A and T1 and incubated for 1 h at 30°C. After adding 6 l of proteinase K (20 mg/ml) and 20 l of SDS (10%) and incubation for 25 min at 35°C, phenol/chloroform extraction and ethanol/glycogen precipitation were done. The resulting pellets were suspended in loading buffer (80% formamide, 10 mM EDTA, 0.1% bromphenol blue), heated for 5 min at 85°C, and separated on a 4% polyacrylamide/urea sequencing gel. After fixation and drying, gels were exposed for several time intervals to Kodak BioMax film. tRNA was used as negative control. 18 S was used as an internal standard for the amount and integrity of RNA preparation. Experiments were performed 3-6 times.
Western Blot Analysis-PC12 cells and Schwann cells were lysed and homogenized in distilled water. A crude cytoplasmic fraction was received after centrifugation (45 min, 14,000 ϫ g). Protein concentrations were determined using Bio-Rad protein assay. After SDS-polyacrylamide gel electrophoresis and semi-dry blotting onto polyvinylidene difluoride membranes (Bio-Rad), immunological detection of FGF-2 was performed using monoclonal FGF-2 antibody (Transduction Laboratories, Lexington, KY) and the enhanced chemiluminescence system from Amersham (Braunschweig, Germany). Experiments were performed 2-3 times.  1, 12, 24, 48, and 120 h after DEX application (10 M). Protection assay was performed on 50 g of total RNA. Autoradiographic film was exposed for 7 days. point (Fig. 1). Further characterization of the steroid hormones which affected the FGF-2 mRNA expression revealed that hydrocortisone elevated the FGF-2 transcript level to the same extent as DEX (Fig. 2). No effects were found after treatment with progesterone, testosterone, and estradiol (Fig. 2). In addition to PC12 cells FGF-2 transcript level was also stimulated in immortalized Schwann cells in the presence of DEX (10 M) (not shown).

Effects of Corticosteroids on FGF-2 and FGFR1 Expression in
Western blot analysis of PC12 and Schwann cells using a monoclonal FGF-2 antibody revealed immunoreactive bands at 18, 21, and 22.5 kDa, which is in good agreement with the previously described rat FGF-2 isoforms. Analyzing the FGF-2 expression under DEX treatment, a dramatic increase of the FGF-2 protein level compared with control culture conditions was found in both PC12 and Schwann cells. Interestingly, this increase was confined to the 21-kDa FGF-2 isoform (Fig. 3). The weakly expressed 18-and 22.5-kDa isoforms remained unchanged (Fig. 3).
The FGFR1 mRNA level was not affected by the steroid hormones tested (not shown).

In Vivo Effect of Dexamethasone on FGF-2 mRNA and Protein Expression and on FGFR1 mRNA Expression in Neuronal
Tissue-To verify that this regulatory mechanism observed in PC12 cells is also operative in vivo, DEX (5 mg/kg) was systemically injected intraperitoneally. In preparations of the adrenal medulla the FGF-2 mRNA level was dramatically enhanced 24 h after injection of DEX (Fig. 4). Western blot analysis of adrenal medullary extracts revealed a significant increase of the 21-kDa FGF-2 isoform level after DEX treatment (Fig. 5). In addition, in vivo treatment with DEX also revealed an increase of the medullary FGFR1 transcript level (Fig. 6).
In Vivo and in Vitro Effect of Dexamethasone on FGF-2 and FGFR1 mRNA Expression in Non-neuronal Tissue-To clarify whether the DEX-induced increase of the FGF-2 mRNA level is a general effect on FGF-2 expressing tissues, non-neuronal tissues were analyzed. In vivo, treatment with DEX produced no elevation of the FGF-2 transcript level in adrenal cortex, heart, kidney, and skeletal muscle, respectively (Fig. 7). Furthermore, Western blot analysis of adrenal cortical extracts revealed no increase of the FGF-2 protein level after DEX injection (Fig. 9). Since these tissues were prepared from the same rats that were also used for the preparations of the adrenal medulla and brain, where the FGF-2 mRNA was increased, the DEX treatment was successful and effective. With regard to the FGFR1, this transcript level was enhanced in the adrenal cortex after DEX treatment (Fig. 10). In the other tissues the FGFR1 mRNA expression was unchanged (Fig. 8).
In vitro, application of DEX to cultured L6 myoblasts revealed no increase of the FGF-2 mRNA and protein level (Figs. 11 and 12). DISCUSSION In the present study we demonstrate for the first time that glucocorticoids mediate an increase of the FGF-2 mRNA and protein level in PC12 cells and in the adrenal medulla. In
FIG. 4. Ribonuclease protection analysis of FGF-2 mRNA expression in adrenal medullae of untreated rats (؊) and of DEXtreated rats (؉). Animals were treated with DEX (5 mg/kg intraperitoneal) for 24 h. Protection assay was performed on 50 g of total RNA. Autoradiographic film was exposed for 5 days.

FIG. 5. Western blot analysis of FGF-2 in adrenal medullary extracts of untreated (؊) and of DEX-treated rats (؉).
Animals were treated with DEX (5 mg/kg intraperitoneal) for 24 h. 100 g of protein was loaded. Human recombinant 18-kDa FGF-2 (200 pg) was used as standard.
FIG. 6. Ribonuclease protection analysis of FGFR1 mRNA expression in adrenal medullae of untreated (؊) and of DEXtreated rats (؉). Animals were treated with DEX (5 mg/kg intraperitoneal) for 24 h. Protection assay was performed on 20 g of total RNA. Autoradiographic film was exposed for 1 day.
FIG. 2. Ribonuclease protection analysis of FGF-2 mRNA expression in PC12 cells under serum-free conditions in the presence of dexamethasone (dex), hydrocortisone (cort), progesterone (prog), estradiol (est), and testosterone (test) after 24-h culture period. Control, cont. Protection assay was performed on 50 g of total RNA. Autoradiographic film was exposed for 5 days.
addition we show that the stimulatory effect of glucocorticoids on FGF-2 expression is confined to neuronal tissues.
From in vitro studies it is known that FGF-2 is involved in development and maintenance of chromaffin cells. In response to FGF-2 embryonic, neonatal, and/or adult, chromaffin cells show proliferation (23,24,49), enhanced catecholamine storage and synthesis (25), and increased tyrosine hydroxylase and proenkephalin expression (26). These FGF-2-mediated effects could be modulated by glucocorticoids. More complicated, however, is the interpretation of our results with regard to the known cooperation of exogenously administered FGF-2 in the differentiation of chromaffin and PC12 cells into sympatheticlike neurons (23, 27-33). Since glucocorticoids maintain the endocrine status of chromaffin cells (23, 35, 38 -40), it must be assumed that there is a difference between the reported effects of exogenously applied FGF-2 and the effect of elevated intracellular FGF-2 induced by glucocorticoids. In the present study we show that the previously reported FGF-2 isoforms (5, 11, 50 -56) are differentially regulated in PC12 cells and in the adrenal medulla after DEX stimulation. Exclusively, the expression of the 21-kDa isoform level is dramatically enhanced, whereas the other ones remain unchanged. One possible explanation could be that the individual FGF-2 proteins are differentially translated. Although the biological significance of the function of the different isoforms is still unclear, it could be suggested that exogenously applied 18-kDa FGF-2 mediates effects other than the intracellular 21-kDa isoform. In accordance with this suggestion is the finding that the overexpression of 18-kDa FGF-2 has a different effect on cell growth and tumorigenic potential than the overexpression of the higher FIG. 7. Ribonuclease protection analysis of FGF-2 mRNA expression in the brain, heart, and kidney of untreated (؊) and of DEX-treated rats (؉). Animals were treated with DEX (5 mg/kg intraperitoneal) for 24 h. Protection assay was performed on 50 g of total RNA. Autoradiographic film was exposed for 2 days.
FIG. 8. Ribonuclease protection analysis of FGFR1 mRNA expression in the brain, heart, and kidney of untreated (؊) and of DEX-treated rats (؉). Animals were treated with DEX (5 mg/kg intraperitoneal) for 24 h. Protection assay was performed on 50 g of total RNA. Autoradiographic film was exposed for 2 days. molecular weight isoforms (57)(58)(59). The present results could help to elucidate the mechanism that causes the effect of glucocorticoids on the establishment of the endocrine status of chromaffin and PC12 cells. Possibly the 21-kDa isoform could be involved in modulating the glucocorticoid-mediated effect. Further studies, however, are necessary to verify this hypothesis. The expression of FGF-2 isoforms during rat brain development, an 18-and 21-kDa isoform is present in the embryo whereas in the adult brain a 22-kDa form is found in addition to 18 and 21 kDa, also suggests that the different FGF-2 isoforms are involved in different functions (5).
With regard to the role of FGF-2 in the brain and in Schwann cells where FGF-2 is also dramatically enhanced after DEX treatment, several possibilities are discussed. Within the central nervous system FGF-2 appears to be involved in the development and de-/regeneration of neurons (4,8). Since at certain steps of development glucocorticoids play a major role in neuronal birth and death (60), it is likely that the synthesis of FGF-2 is modulated by glucocorticoids during this phase. FGF-2 protein and mRNA are expressed in the embryonic and newborn nervous system (5,6,61). Glucocorticoids are also important for adult central neurons. The specific degeneration in the hippocampal dentate gyrus induced by adrenalectomy (62), for example, could be the result of a reduced trophic factor production after depletion of glucocorticoids. In the postnatal and adult brain FGF-2 protein and mRNA are found in astrocytes and in distinct neuronal populations (63)(64)(65). Although in vitro studies have shown that the FGF-2 increase after DEX treatment is induced in astrocytes (44), it cannot be ruled out that the DEX-mediated FGF-2 increase in vivo also takes place in neurons. In addition to FGF-2 the biosynthesis of other trophic factors such as neurotrophins is regulated in the brain and in cultured neurons and astrocytes by glucocorticoids (41,42,45).
For Schwann cells several physiological functions of steroids are proposed. In vitro studies showed that progesterone that is synthesized in Schwann cells increases the myelination of axons (66). Glucocorticoids are potent co-mitogens for proliferation of Schwann cells of the postnatal rat (67). In addition to the FGF-2 protein shown in this study, the FGF-2 transcript is also found in immortalized and in clonal Schwann cells. 2 Exogenously applied 18-kDa FGF-2 stimulates its own expression in immortalized Schwann cells 2 but is not acting as a mitogen for isolated adult Schwann cells (68). Po, a myelin-specific protein, is regulated by exogenously administered 18-kDa FGF-2 (69). Whether the 21-kDa FGF-2 isoform which is increased after DEX treatment in immortalized Schwann cells mediates functions other than the 18-kDa form remains to be elucidated.
An additional important result from this study is the specificity of the glucocorticoid effect. Although non-neuronal tissues such as adrenal cortex, kidney, heart, skeletal muscle, and L6 myoblasts express FGF-2 and possess glucocorticoid receptors (70), the FGF-2 expression is not regulated after DEX treatment. In untreated and DEX-stimulated non-neuronal tissues (L6 myoblasts, adrenal cortex), the 21-kDa FGF-2 isoform displays a high expression. In chromaffin cells and Schwann cells the 21-kDa isoform is weakly expressed and increases dramatically after DEX treatment. This result could suggest that a stress-or damage-induced high glucocorticoid concentration increases FGF-2 in neuronal tissue. Further studies have to elucidate whether the 21-kDa FGF-2 isoform mediates an effect different from those mediated by the 18-kDa form.
The regulation of the FGFR1 seems to be more complex since in vitro the transcript level remains unchanged in PC12 cells after DEX stimulation, whereas in vivo DEX induced a dramatic increase of the FGFR1 mRNA level both in the cortex and the medulla of the adrenal gland. Previous studies also suggested that several molecules are acting synergistically on the FGFR1 expression. Under serum-containing but not under serum-free culture conditions NGF induced an increase of the FGFR1 mRNA level in PC12 cells (20).
In conclusion, we show here that glucocorticoids enhance FGF-2 protein and mRNA in PC12 cells and in the adrenal medulla in vivo. This effect is specific for neuronal tissue. Moreover, the individual FGF-2 isoforms appear to be regulated independently. Exclusively the 21-kDa FGF-2 form is increased after DEX treatment. This may account for a specific function of the 21-kDa form in neuronal tissue.