Rat Anterior Pituitary Folliculostellate Cells Are Targets of Interleukin-1ß and a Major Source of Intrapituitary Follistatin
Abstractk*wm, 百拇医药
Folliculostellate cells of the anterior pituitary are postulated to be an important source of factors, such as follistatin, that regulate pituitary function by intercellular communication. To gain further insight into the function of this cell type, folliculostellate cells were enriched from cultured rat anterior pituitary cells, and an immortalized cell line designated FS/D1h was established and characterized. These FS/D1h cells express S100 immunoreactivity and produce IL-6 but not pituitary hormones such as GH, ACTH, FSH, and LH. Importantly, FS/D1h cells express large amounts of follistatin mRNA and secrete the protein, as quantified indirectly by the amount of [125I]activin A immunoprecipitated with a follistatin antiserum. The FS/D1h cells also express {alpha} , ßA, and ßB inhibin/activin subunit mRNAs, but whether they produce the corresponding activins and inhibins has not been determined. The response of FS/D1h cells to agents thought to modulate folliculostellate cell function was evaluated. IL-1ß (0.005–5 nM) stimulated the secretion of follistatin and increased mRNA expression. In parallel, IL-6 secretion was stimulated. Dexamethasone, pituitary adenylate cyclase-activating polypeptide(1–27), and lipopolysaccharide but not testosterone, 12-O-tetradecanoylphorbol-13-acetate, or forskolin also increased follistatin secretion. Surprisingly, activin had no effect on follistatin mRNA levels, despite the fact that FS/D1h cells express ActRII, ActRIIB, and ALK-4 (ActRIB). Activin, on the other hand, induced Smad7 mRNA accumulation and exerted an antiproliferative effect on FS/D1h cells. Altogether, these observations support the possibility that follistatin originating from folliculostellate cells participates in mediating the effects of IL-1ß, glucocorticoids, and other agents on the response of pituitary cells to activins.
Introductiontf[u, 百拇医药
FOLLICULOSTELLATE CELLS OF the anterior pituitary comprise a group of S100-antigen-positive nonendocrine cells whose function remains poorly defined. They are postulated to play an important role as a source of paracrine factors that act locally to modulate pituitary responses to hypothalamic and peripheral signals (1, 2). Electrophysiological and morphological studies suggest that folliculostellate cells form a functional network of interconnecting cells that facilitate synchronization and long-range communication among different parts of the pituitary (3). These cells also have been viewed as a pool of progenitor cells capable of differentiating into specialized endocrine cells under certain conditions (4).tf[u, 百拇医药
Activins, along with their functional antagonists, the inhibins and follistatins, are an important group of autocrine/paracrine factors involved in the local modulation of the anterior pituitary (5, 6, 7, 8, 9, 10, 11). Activins and inhibins are members of the TGFß family of growth and differentiation factors, and as is characteristic of this family of proteins, they are expressed in most tissues and exert a broad spectrum of local or endocrine effects (5, 12). Activin A, B, and AB; homo/heterodimers of inhibin/activin ßA- and ßB-subunits; and inhibin A and B, heterodimers of the same ßA- or ßB- subunits with the structurally related inhibin/activin {alpha} -subunit are documented to exert pituitary effects and be expressed in this tissue (5, 13). Activin effects are mediated by the widely distributed activin-specific type II (ActRII and ActRIIB) and type I (ActRIB or ALK4) receptors (12, 14). By contrast, the antagonistic function of inhibin appears to be restricted to a subset of activin-responsive cell types that express inhibin-specific binding proteins (15, 16, 17, 18). Most anterior pituitary cell types are responsive to activins (7, 13), but gonadotropes seem to be the only known targets of inhibin within this tissue (5, 13, 16, 19).
Follistatins are widely distributed activin-binding glycoproteins that serve an important function by their ability to biologically inactivate the activins (6, 20, 21). The observation that follistatin-deficient mice suffer many abnormalities and die within a few hours of birth is presumed to reflect the absence of an important mechanism that modulates the actions of activins and possibly other TGFß family members also known to bind to follistatin (22). Two forms of follistatins differing in length at the C terminus, FS315 and FS288, are encoded by two alternatively spliced mRNA products of a single gene (23). The relative distribution of the two follistatins is not well characterized, but FS315 is likely to be the circulating form because truncated forms have a higher affinity for cell surface heparin sulfate proteoglycans and the mRNA encoding FS315 is more abundant than the shorter version (24, 25, 26). Consistently, follistatins present in human serum display characteristics of the longer FS315 form, whereas those in human follicular fluid are likely to be the FS288 splice variant probably involved in autocrine/paracrine functions (27). Follistatins and inhibin/activin subunits often exhibit an overlapping pattern of expression. Pituitary follistatin expression is strongly induced by activins (8, 28, 29). Follistatin mRNA is present in most cell types of the anterior pituitary, including folliculostellate cells. Various preparations of anterior pituitary cultures or enriched pituitary folliculostellate cells obtained from a number of different species have been reported to express the mRNA or secrete the protein (24, 30, 31, 32, 33).
In addition to follistatins, preparations of rat, mouse, bovine, or primate primary pituitary folliculostellate cells or cell lines derived from them have been shown to also produce IL-6 (34, 35, 36, 37), lipocortin 1 (38), vascular endothelial growth factor (VEGF) (30, 39), leukemia inhibitory factor (40), TGFß (41, 42), and basic fibroblast growth factor (43, 44). Moreover, these cells have been reported to be responsive to a number of agents with known pituitary effects, including TGFß, activin, pituitary adenylate cyclase-activating polypeptide (PACAP), calcitonin, IL-1ß, IL-6, glucocorticoids, and lipopolysaccharide (LPS) (37, 39, 42, 44, 45, 46, 47). It has been postulated that these agents exert effects on the endocrine cells of the anterior pituitary, in part, by modulating the production of paracrine factors from folliculostellate cells (1, 2). Along these lines, it has been proposed that the reciprocal relationship between follistatin and FSHß mRNA levels seen in nonhuman primate anterior pituitary cell preparations reflects the importance of folliculostellate cells as a local source of follistatin (48).
Studies with rat anterior pituitary preparations have also suggested that follistatin derived from folliculostellate cells exerts a paracrine influence on the regulation of gonadotropes by activin (33). Similarly, the influence of PACAP on gonadotropes may reflect its ability to modulate follistatin production from folliculostellate cells (46). Finally, one of the proposed mechanisms by which IL-1ß modulates the hypothalamo-pituitary-adrenal or hypothalamo-pituitary-gonadal axis is by stimulating IL-6 and follistatin production from folliculostellate cells and thereby modifying ACTH or FSH secretion from the pituitary (49, 50). To gain further insight into their role, a previously described strategy was used to obtain enriched preparations of rat anterior pituitary folliculostellate cells (33). These cells were subjected to dilution cloning and independent lines were selected for their ability to proliferate, express S100, and secrete measurable follistatin. One of these lines, FS/D1h, was characterized further and shown to secrete follistatin and IL-6 in response to IL-1ß and be responsive to activin A, dexamethasone, PACAP, and LPS, confirming that a subset of S100-positive folliculostellate cells are targets of IL-1ß and a source of pituitary follistatin and IL-6. These studies with the FS/D1h cell line also provide further support for a role of rat anterior pituitary folliculostellate cells in mediating the suppressive effects of IL-1ß on activin-mediated FSH secretion.
Materials and Methods]0cnh, 百拇医药
Enrichment and selection of FS/D1h folliculostellate cells]0cnh, 百拇医药
Folliculostellate cells were enriched by a previously published method (33). Anterior pituitary cells from male Sprague Dawley rats (180–200 g) were digested with collagenase and the dispersed cells [rat anterior pituitary cells (RAP)] were cultured in ßPJ medium (51) supplemented with 15% fetal bovine serum (FBS) and 6 mM Gln (5 x 106 cells per 10-cm dish). The medium was changed every 2–3 d, and 2 wk later a population of flat adherent cells that covered the dish were visible. These adherent cells were trypsinized and replated at a dilution of 1:6 to 1:8 and allowed to grow to confluence. This process was repeated several times, and the cells that survived after the fourth passage were distributed into several 24-well plates at a low density. Cells that produced a factor with FSH inhibitory activity were removed by trypsinization, pooled, and subjected to dilution cloning by culturing them in two 48-well plates at a calculated density of less than 1cell/well. After approximately 2 wk, 8 of 96 wells contained visible colonies of proliferating cells, and medium conditioned by cells in one of these wells contained measurable follistatin levels. The cells recovered from this well were subjected to another round of dilution cloning, and one of the wells ultimately yielded the cell line designated FS/D1h.
Follistatin and IL-6 measurements.h9}, 百拇医药
Follistatin secretion was measured by an indirect RIA using a rabbit anti-rhFS288 serum (no. 5542) that binds either free follistatin or the follistatin:activin complex. Briefly, serial dilutions of conditioned medium or recombinant human (rh) FS288 were incubated for 18–20 h at 4 C with anti-rhFS288 (1:5,000) in the presence of approximately 100,000 cpm [125I]activin A (~ 3 x 106 cpm/pmol), 0.1% BSA, Aprotinin (1:500), and 0.01% Triton X-100 in a final volume of 0.5 ml. Activin A was labeled as described previously (52). The follistatin:[125I]activin A complexes bound to anti-FS288 were immunoprecipitated by first incubating the samples with sheep antirabbit IgG (1:100) and normal rabbit serum (1:1000) for 30 min at 4 C and then centrifuging them in the presence of 5% polyethylene glycol in a final volume of 1 ml. Nonspecific interactions were determined by substituting normal rabbit serum for anti-FS288. After discarding the supernatants, radioactivity remaining in the pellets was determined on a {gamma} -counter, and specifically immunoprecipitated [125I]activin A values were determined by subtracting nonspecific counts. Regression analysis was used to analyze the data with Prism 3.0a for MacIntosh (GraphPad Software, Inc., San Diego, CA). All samples and standard curves were measured in triplicate. The minimum detectable concentration of the assay was in the range of 0.05–0.1 ng/ml rhFS288, and the assay was linear over approximately two log units.
The secretion of IL-6 bioactivity was determined using a 7TD1 cell survival assay, essentially as described (53). Briefly, 7TD1 cells were cultured in RPMI 1640 containing 5% FBS, 50 µM 2-ß-mercaptoethanol, antibiotics, and 2 ng/ml rhIL-6. Samples or standards were serially diluted (1:2) in assay medium and added to cells cultured in 96-well plates and 72 h later, the plates were centrifuged (300 x g for 5 min) to remove the medium. Cell proliferation was measured colorimetrically using the tetrazolium salt, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide, and quantification at 570 nm test and 640 nm reference wavelengths using a microplate reader (MR700, Dynatech Corp. Laboratories, Chantilly, VA). Concentrations of unknown samples were determined by measuring the displacement of their dilution curves from the known standard (the detection limit of the assay was 0.08–0.25 international interim units/ml). All samples were assayed in triplicate (on separate plates), and IL-6 values are presented as international interim units, based on comparisons with the standard.
S100 immunocytochemistry(+@6o, http://www.100md.com
The FS/D1h cells were cultured in eight-chambered Falcon CultureSlides (Becton Dickinson and Co. Labware, Bedford, MA) at a density of 50,000 cells/chamber. Three days later, they were washed with PBS, fixed with 4% paraformaldehyde for 15 min on ice, rewashed, and permeabilized with 0.2% Triton X-100 for 10 min and washed again with PBS. Nonspecific sites were blocked by incubation for 30 min at ambient temperature with PBS containing 2% normal donkey serum, 0.1% BSA, 0.1% Tween 20 before adding affinity-purified rabbit anti-S100 IgG (1:250) (Sigma, St. Louis, MO) in the same buffer for 1 h. After several washes, Texas Red dye-conjugated affinity-purified donkey antirabbit IgG (1:300, Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was added and allowed to incubate for 1 h at ambient temperature before washing, applying ProLong Antifade reagent (Molecular Probes, Inc., Eugene, OR) and cover slipping. Specificity was determined by either excluding the primary antibody or substituting it with normal rabbit IgG. Immunostaining for glial fibrillary acidic protein (GFAP) was evaluated using guinea-pig antihuman GFAP (1:250) (Advanced ImmunoChemical Inc., Long Beach, CA) and Texas Red dye-conjugated affinity-purified donkey anti-guinea pig IgG (1:400; Jackson ImmunoResearch Laboratories, Inc.). Fluorescent signals were captured using an eclipse E600 microscope (Nikon, Melville, NY) fitted with a mercury epifluorescence mercury light source and a CoolSNAP monochromatic digital camera and analyzed using the Image-Pro Plus (version 4.5) image analysis software (Media Cybernetics, Inc., Silver Spring, MD).
Ribonuclease (RNase) protection$2z%{, http://www.100md.com
Total RNA was extracted using the RNeasy kit (QIAGEN, Hilden, Germany). Steady-state inhibin/activin {alpha} , ßA, ßB, follistatin, and Smad7 mRNA levels were measured by RNase protection, using rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an internal control as previously described (28, 54). For the analysis of primary follistatin transcripts, a template containing 203 bp of exon 3 of the rat follistatin gene and 260 bp of the adjacent third intron (in pBluescript IISK, Stratagene, La Jolla, CA) was digested with HindIII and used to measure unprocessed transcript and mature mRNA simultaneously. The antisense riboprobe was synthesized using T3 RNA polymerase and levels were measured using total RNA. The ability of this riboprobe to detect primary transcript was validated by analyzing nuclear and cytoplasmic RNA samples individually and by kinetic studies to establish an optimal time point for these measurements. Quantitative analysis was performed using the PhosphorImager system and the ImageQuant 4.0 software package (Molecular Dynamics, Inc., Sunnyvale, CA). Reported data reflect transcript levels normalized to the internal GAPDH control. Treatments were replicated at least three times, and the results are reported as means ± SEM.
Cell proliferationyi/fe-, 百拇医药
Cells were cultured in 96-well plates at a density of 103 cells/well in 0.1 ml complete medium and equilibrated overnight. Test substances were added into triplicate wells, and cell proliferation was assessed after 4 d using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide proliferation assay kit (Cell Titer 96, Promega Corp., Madison, WI).yi/fe-, 百拇医药
FSH measurements and statistical analysisyi/fe-, 百拇医药
For the measurement of FSH secretion, RAP cells were plated in 48-well plates (Costar, Cambridge, MA) (0.15 x 106 cells/well) and allowed to recover 3 d in ßPJ medium (51) supplemented with 2% FBS. The cells were washed, and test substances or conditioned medium was added to the wells and secretion was measured in a final volume of 0.5 ml. FSH was quantified using RIA reagents generously provided by Dr. Parlow through the National Pituitary and Hormone Distribution Program at NIDDK. All treatments were performed in triplicate, and data from at least three determinations were subjected to either the t test or ANOVA followed by Dunnett’s multiple comparison test, as indicated in the figure legends.
Reagentsx3x-, 百拇医药
The rh-activin-A was provided by Genentech, Inc. (San Francisco, CA); rhFS288 was obtained through the National Hormone and Pituitary Program of NIDDK; rhIL-1ß was a gift from Dr. Liu of Neurocrine Biosciences Inc. (La Jolla, CA); rhIL-6 was purchased from Endogen, Inc. (Woburn, MA), and LPS was from Sigma (Escherichia coli serotype 026:B6m code L3755, lot 20H4025).x3x-, 百拇医药
Resultsx3x-, 百拇医药
Selection and establishment of a rat anterior pituitary folliculostellate cell line, FS/D1hx3x-, 百拇医药
An indirect RIA was used for measuring follistatin secretion from rat anterior pituitary cells. As shown in two examples of a typical standard curve, the minimum concentration of free FS288 detectable by this indirect immunoassay was in the range of 0.05–0.1 ng/ml, and the assay was linear over approximately two log units (Fig. 1). Concentrations of activin A greater than 0.01 nM lowered the sensitivity of the assay because of competition with radiolabeled activin A, indicating that the assay provides an estimate of only free follistatin (Fig. 1A). This assay was used to measure relative changes in follistatin levels by comparing the linear portion of serially diluted media to the standard curve, as shown by a typical example of the analysis of a set of control and treated samples (Fig. 1B).
fig.ommitteedn|h73h!, 百拇医药
Figure 1. Typical FS288 standard curves obtained from an indirect RIA in which follistatin concentrations were quantified based on follistatin:[125I]activin A complexes immunoprecipitated with anti-FS288 (no. 5542). Approximately 100,000 cpm [125I]activin A were incubated with the indicated concentrations of FS288, with and without cold activin A (A), or with serially diluted medium conditioned by FS/D1h cells treated as indicated (B). The samples were immunoprecipitated with anti-FS288 sheep antirabbit IgG and processed as described in Materials and Methods. The values reflect the mean ± SEM of triplicate determinations from typical experiments.n|h73h!, 百拇医药
After culturing rat anterior pituitary cells for 2–3 wk in high FBS (15%), medium conditioned by the surviving cells contained no measurable GH, FSH, LH, or ACTH (data not shown). However, the medium was observed to contain an activity that suppressed basal FSH secretion and attenuated the effect of submaximal concentrations of activin A (Fig. 2A). Subjecting this population of cells to multiple dilution cloning steps, as described in Materials and Methods, yielded several independent lines that await further analysis. One of these lines, herein referred to as FS/D1h cells, was determined to express considerably more follistatin mRNA, compared with the other lines or to the mixed population of rat anterior pituitary cells (data not shown), and secrete substantial amounts of follistatin (Fig. 2B). Conditioned medium from the FS/D1h line when applied to cultured rat anterior pituitary cells also suppressed basal FSH secretion, consistent with the expected action of follistatin on gonadotropes (Fig. 2C).
fig.ommitteedf, 百拇医药
Figure 2. The suppression of FSH secretion by medium conditioned by enriched rat anterior pituitary folliculostellate cells (A) and the evaluation of medium conditioned by several selected folliculostellate lines for follistatin secretion (B) and effects on FSH secretion (C). A, Control or conditioned medium (0.3 ml) from enriched folliculostellate cells in 10-cm dishes was added to rat anterior pituitary cells cultured in 48-well plates (Costar) at a final volume of 0.5 ml/well, with and without various concentrations of activin A. B, Follistatin levels in medium conditioned for 7 d with the initial pool of cells or several selected folliculostellate lines were quantified by an indirect RIA described in the text. C, 0.2 ml control or conditioned medium by the same lines shown in panel B were tested for effects on FSH secretion as described for panel A. The effect of FS288 on FSH secretion is shown at the far right. The data are means ± SEM of triplicate determinations of representative experiments. The data within each panel were subjected to ANOVA and Dunnett’s multiple comparison test. For A, statistical comparisons are relative to control medium for each x-axis value. For B and C, comparisons are relative to FSH or follistatin (FS) secretion from the pool: **, P < 0.001; #, P < 0.01; *, P < 0.05.
Immunocytochemical evaluation revealed that all FS/D1h cells expressed the S100 antigen (Fig. 3). This cell-associated staining was no longer visible when anti-S100 was omitted or substituted with normal rabbit IgG, verifying the specificity of the S100 immunoreactivity (Fig. 3B). The FS/D1h cells were also positive for GFAP immunostaining, another marker of the anterior pituitary cell population (1) (data not shown). The FS/D1h cells had a doubling time of approximately 48 h, formed a monolayer, and essentially stopped growing when they reached confluence. Light microscopy indicated the presence of large vacuolar structures in the cytoplasm (data not shown). The FS/D1h cells established cell-cell contact via long cytoplasmic processes, and during the logarithmic growth phase, they had a more stellate appearance but became more rounded at higher densities (data not shown). Because folliculostellate cells have been proposed to be the source of pituitary IL-6, the production of this cytokine from FS/D1h cells was evaluated and levels were compared with those of RAP cells. The results indicated that FS/D1h cells secrete considerably more IL-6, compared with RAP cells (Fig. 4A). Moreover, the cells responded to IL-1ß by an increase in the production of not only IL-6 but also follistatin (Fig. 4B). These observations indicate that FS/D1h cells possess at least some of the characteristics usually associated with pituitary folliculostellate cells. The cells were amenable to propagation through multiple passages (up to 35–40 passages tested) without loss of phenotypic characteristics such as follistatin and IL-6 production, morphological appearance, and growth pattern; S100 immunoreactivity was not evaluated beyond passage no. 19.d\, http://www.100md.com
fig.ommitteedd\, http://www.100md.com
Figure 3. Immunocytochemical analysis for S100 protein expression in FS/D1h cells (passage no. 5). Texas Red fluorescence was captured at x40 magnification. B, The signal obtained when the primary S100 antiserum was omitted is shown.d\, http://www.100md.com
fig.ommitteedd\, http://www.100md.com
Figure 4. The secretion of IL-6 and follistatin from FS/D1h cells (passage no. 10) and cultured RAP cells in response to 0.5 nM IL-1ß. The data are means ± SEM of triplicate determinations of a representative experiment. The t test was used to compare IL-6 or follistatin (FS) secretion from the indicated cell types incubated with or without IL-1ß. ND, Not detectable. **, P < 0.001; #, P < 0.01.(Louise M. Bilezikjian Angela M. O. Leal Amy L. Blount Anne Z. Corrigan Andrew V. Turnbull and Wylie )
Folliculostellate cells of the anterior pituitary are postulated to be an important source of factors, such as follistatin, that regulate pituitary function by intercellular communication. To gain further insight into the function of this cell type, folliculostellate cells were enriched from cultured rat anterior pituitary cells, and an immortalized cell line designated FS/D1h was established and characterized. These FS/D1h cells express S100 immunoreactivity and produce IL-6 but not pituitary hormones such as GH, ACTH, FSH, and LH. Importantly, FS/D1h cells express large amounts of follistatin mRNA and secrete the protein, as quantified indirectly by the amount of [125I]activin A immunoprecipitated with a follistatin antiserum. The FS/D1h cells also express {alpha} , ßA, and ßB inhibin/activin subunit mRNAs, but whether they produce the corresponding activins and inhibins has not been determined. The response of FS/D1h cells to agents thought to modulate folliculostellate cell function was evaluated. IL-1ß (0.005–5 nM) stimulated the secretion of follistatin and increased mRNA expression. In parallel, IL-6 secretion was stimulated. Dexamethasone, pituitary adenylate cyclase-activating polypeptide(1–27), and lipopolysaccharide but not testosterone, 12-O-tetradecanoylphorbol-13-acetate, or forskolin also increased follistatin secretion. Surprisingly, activin had no effect on follistatin mRNA levels, despite the fact that FS/D1h cells express ActRII, ActRIIB, and ALK-4 (ActRIB). Activin, on the other hand, induced Smad7 mRNA accumulation and exerted an antiproliferative effect on FS/D1h cells. Altogether, these observations support the possibility that follistatin originating from folliculostellate cells participates in mediating the effects of IL-1ß, glucocorticoids, and other agents on the response of pituitary cells to activins.
Introductiontf[u, 百拇医药
FOLLICULOSTELLATE CELLS OF the anterior pituitary comprise a group of S100-antigen-positive nonendocrine cells whose function remains poorly defined. They are postulated to play an important role as a source of paracrine factors that act locally to modulate pituitary responses to hypothalamic and peripheral signals (1, 2). Electrophysiological and morphological studies suggest that folliculostellate cells form a functional network of interconnecting cells that facilitate synchronization and long-range communication among different parts of the pituitary (3). These cells also have been viewed as a pool of progenitor cells capable of differentiating into specialized endocrine cells under certain conditions (4).tf[u, 百拇医药
Activins, along with their functional antagonists, the inhibins and follistatins, are an important group of autocrine/paracrine factors involved in the local modulation of the anterior pituitary (5, 6, 7, 8, 9, 10, 11). Activins and inhibins are members of the TGFß family of growth and differentiation factors, and as is characteristic of this family of proteins, they are expressed in most tissues and exert a broad spectrum of local or endocrine effects (5, 12). Activin A, B, and AB; homo/heterodimers of inhibin/activin ßA- and ßB-subunits; and inhibin A and B, heterodimers of the same ßA- or ßB- subunits with the structurally related inhibin/activin {alpha} -subunit are documented to exert pituitary effects and be expressed in this tissue (5, 13). Activin effects are mediated by the widely distributed activin-specific type II (ActRII and ActRIIB) and type I (ActRIB or ALK4) receptors (12, 14). By contrast, the antagonistic function of inhibin appears to be restricted to a subset of activin-responsive cell types that express inhibin-specific binding proteins (15, 16, 17, 18). Most anterior pituitary cell types are responsive to activins (7, 13), but gonadotropes seem to be the only known targets of inhibin within this tissue (5, 13, 16, 19).
Follistatins are widely distributed activin-binding glycoproteins that serve an important function by their ability to biologically inactivate the activins (6, 20, 21). The observation that follistatin-deficient mice suffer many abnormalities and die within a few hours of birth is presumed to reflect the absence of an important mechanism that modulates the actions of activins and possibly other TGFß family members also known to bind to follistatin (22). Two forms of follistatins differing in length at the C terminus, FS315 and FS288, are encoded by two alternatively spliced mRNA products of a single gene (23). The relative distribution of the two follistatins is not well characterized, but FS315 is likely to be the circulating form because truncated forms have a higher affinity for cell surface heparin sulfate proteoglycans and the mRNA encoding FS315 is more abundant than the shorter version (24, 25, 26). Consistently, follistatins present in human serum display characteristics of the longer FS315 form, whereas those in human follicular fluid are likely to be the FS288 splice variant probably involved in autocrine/paracrine functions (27). Follistatins and inhibin/activin subunits often exhibit an overlapping pattern of expression. Pituitary follistatin expression is strongly induced by activins (8, 28, 29). Follistatin mRNA is present in most cell types of the anterior pituitary, including folliculostellate cells. Various preparations of anterior pituitary cultures or enriched pituitary folliculostellate cells obtained from a number of different species have been reported to express the mRNA or secrete the protein (24, 30, 31, 32, 33).
In addition to follistatins, preparations of rat, mouse, bovine, or primate primary pituitary folliculostellate cells or cell lines derived from them have been shown to also produce IL-6 (34, 35, 36, 37), lipocortin 1 (38), vascular endothelial growth factor (VEGF) (30, 39), leukemia inhibitory factor (40), TGFß (41, 42), and basic fibroblast growth factor (43, 44). Moreover, these cells have been reported to be responsive to a number of agents with known pituitary effects, including TGFß, activin, pituitary adenylate cyclase-activating polypeptide (PACAP), calcitonin, IL-1ß, IL-6, glucocorticoids, and lipopolysaccharide (LPS) (37, 39, 42, 44, 45, 46, 47). It has been postulated that these agents exert effects on the endocrine cells of the anterior pituitary, in part, by modulating the production of paracrine factors from folliculostellate cells (1, 2). Along these lines, it has been proposed that the reciprocal relationship between follistatin and FSHß mRNA levels seen in nonhuman primate anterior pituitary cell preparations reflects the importance of folliculostellate cells as a local source of follistatin (48).
Studies with rat anterior pituitary preparations have also suggested that follistatin derived from folliculostellate cells exerts a paracrine influence on the regulation of gonadotropes by activin (33). Similarly, the influence of PACAP on gonadotropes may reflect its ability to modulate follistatin production from folliculostellate cells (46). Finally, one of the proposed mechanisms by which IL-1ß modulates the hypothalamo-pituitary-adrenal or hypothalamo-pituitary-gonadal axis is by stimulating IL-6 and follistatin production from folliculostellate cells and thereby modifying ACTH or FSH secretion from the pituitary (49, 50). To gain further insight into their role, a previously described strategy was used to obtain enriched preparations of rat anterior pituitary folliculostellate cells (33). These cells were subjected to dilution cloning and independent lines were selected for their ability to proliferate, express S100, and secrete measurable follistatin. One of these lines, FS/D1h, was characterized further and shown to secrete follistatin and IL-6 in response to IL-1ß and be responsive to activin A, dexamethasone, PACAP, and LPS, confirming that a subset of S100-positive folliculostellate cells are targets of IL-1ß and a source of pituitary follistatin and IL-6. These studies with the FS/D1h cell line also provide further support for a role of rat anterior pituitary folliculostellate cells in mediating the suppressive effects of IL-1ß on activin-mediated FSH secretion.
Materials and Methods]0cnh, 百拇医药
Enrichment and selection of FS/D1h folliculostellate cells]0cnh, 百拇医药
Folliculostellate cells were enriched by a previously published method (33). Anterior pituitary cells from male Sprague Dawley rats (180–200 g) were digested with collagenase and the dispersed cells [rat anterior pituitary cells (RAP)] were cultured in ßPJ medium (51) supplemented with 15% fetal bovine serum (FBS) and 6 mM Gln (5 x 106 cells per 10-cm dish). The medium was changed every 2–3 d, and 2 wk later a population of flat adherent cells that covered the dish were visible. These adherent cells were trypsinized and replated at a dilution of 1:6 to 1:8 and allowed to grow to confluence. This process was repeated several times, and the cells that survived after the fourth passage were distributed into several 24-well plates at a low density. Cells that produced a factor with FSH inhibitory activity were removed by trypsinization, pooled, and subjected to dilution cloning by culturing them in two 48-well plates at a calculated density of less than 1cell/well. After approximately 2 wk, 8 of 96 wells contained visible colonies of proliferating cells, and medium conditioned by cells in one of these wells contained measurable follistatin levels. The cells recovered from this well were subjected to another round of dilution cloning, and one of the wells ultimately yielded the cell line designated FS/D1h.
Follistatin and IL-6 measurements.h9}, 百拇医药
Follistatin secretion was measured by an indirect RIA using a rabbit anti-rhFS288 serum (no. 5542) that binds either free follistatin or the follistatin:activin complex. Briefly, serial dilutions of conditioned medium or recombinant human (rh) FS288 were incubated for 18–20 h at 4 C with anti-rhFS288 (1:5,000) in the presence of approximately 100,000 cpm [125I]activin A (~ 3 x 106 cpm/pmol), 0.1% BSA, Aprotinin (1:500), and 0.01% Triton X-100 in a final volume of 0.5 ml. Activin A was labeled as described previously (52). The follistatin:[125I]activin A complexes bound to anti-FS288 were immunoprecipitated by first incubating the samples with sheep antirabbit IgG (1:100) and normal rabbit serum (1:1000) for 30 min at 4 C and then centrifuging them in the presence of 5% polyethylene glycol in a final volume of 1 ml. Nonspecific interactions were determined by substituting normal rabbit serum for anti-FS288. After discarding the supernatants, radioactivity remaining in the pellets was determined on a {gamma} -counter, and specifically immunoprecipitated [125I]activin A values were determined by subtracting nonspecific counts. Regression analysis was used to analyze the data with Prism 3.0a for MacIntosh (GraphPad Software, Inc., San Diego, CA). All samples and standard curves were measured in triplicate. The minimum detectable concentration of the assay was in the range of 0.05–0.1 ng/ml rhFS288, and the assay was linear over approximately two log units.
The secretion of IL-6 bioactivity was determined using a 7TD1 cell survival assay, essentially as described (53). Briefly, 7TD1 cells were cultured in RPMI 1640 containing 5% FBS, 50 µM 2-ß-mercaptoethanol, antibiotics, and 2 ng/ml rhIL-6. Samples or standards were serially diluted (1:2) in assay medium and added to cells cultured in 96-well plates and 72 h later, the plates were centrifuged (300 x g for 5 min) to remove the medium. Cell proliferation was measured colorimetrically using the tetrazolium salt, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide, and quantification at 570 nm test and 640 nm reference wavelengths using a microplate reader (MR700, Dynatech Corp. Laboratories, Chantilly, VA). Concentrations of unknown samples were determined by measuring the displacement of their dilution curves from the known standard (the detection limit of the assay was 0.08–0.25 international interim units/ml). All samples were assayed in triplicate (on separate plates), and IL-6 values are presented as international interim units, based on comparisons with the standard.
S100 immunocytochemistry(+@6o, http://www.100md.com
The FS/D1h cells were cultured in eight-chambered Falcon CultureSlides (Becton Dickinson and Co. Labware, Bedford, MA) at a density of 50,000 cells/chamber. Three days later, they were washed with PBS, fixed with 4% paraformaldehyde for 15 min on ice, rewashed, and permeabilized with 0.2% Triton X-100 for 10 min and washed again with PBS. Nonspecific sites were blocked by incubation for 30 min at ambient temperature with PBS containing 2% normal donkey serum, 0.1% BSA, 0.1% Tween 20 before adding affinity-purified rabbit anti-S100 IgG (1:250) (Sigma, St. Louis, MO) in the same buffer for 1 h. After several washes, Texas Red dye-conjugated affinity-purified donkey antirabbit IgG (1:300, Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was added and allowed to incubate for 1 h at ambient temperature before washing, applying ProLong Antifade reagent (Molecular Probes, Inc., Eugene, OR) and cover slipping. Specificity was determined by either excluding the primary antibody or substituting it with normal rabbit IgG. Immunostaining for glial fibrillary acidic protein (GFAP) was evaluated using guinea-pig antihuman GFAP (1:250) (Advanced ImmunoChemical Inc., Long Beach, CA) and Texas Red dye-conjugated affinity-purified donkey anti-guinea pig IgG (1:400; Jackson ImmunoResearch Laboratories, Inc.). Fluorescent signals were captured using an eclipse E600 microscope (Nikon, Melville, NY) fitted with a mercury epifluorescence mercury light source and a CoolSNAP monochromatic digital camera and analyzed using the Image-Pro Plus (version 4.5) image analysis software (Media Cybernetics, Inc., Silver Spring, MD).
Ribonuclease (RNase) protection$2z%{, http://www.100md.com
Total RNA was extracted using the RNeasy kit (QIAGEN, Hilden, Germany). Steady-state inhibin/activin {alpha} , ßA, ßB, follistatin, and Smad7 mRNA levels were measured by RNase protection, using rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an internal control as previously described (28, 54). For the analysis of primary follistatin transcripts, a template containing 203 bp of exon 3 of the rat follistatin gene and 260 bp of the adjacent third intron (in pBluescript IISK, Stratagene, La Jolla, CA) was digested with HindIII and used to measure unprocessed transcript and mature mRNA simultaneously. The antisense riboprobe was synthesized using T3 RNA polymerase and levels were measured using total RNA. The ability of this riboprobe to detect primary transcript was validated by analyzing nuclear and cytoplasmic RNA samples individually and by kinetic studies to establish an optimal time point for these measurements. Quantitative analysis was performed using the PhosphorImager system and the ImageQuant 4.0 software package (Molecular Dynamics, Inc., Sunnyvale, CA). Reported data reflect transcript levels normalized to the internal GAPDH control. Treatments were replicated at least three times, and the results are reported as means ± SEM.
Cell proliferationyi/fe-, 百拇医药
Cells were cultured in 96-well plates at a density of 103 cells/well in 0.1 ml complete medium and equilibrated overnight. Test substances were added into triplicate wells, and cell proliferation was assessed after 4 d using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide proliferation assay kit (Cell Titer 96, Promega Corp., Madison, WI).yi/fe-, 百拇医药
FSH measurements and statistical analysisyi/fe-, 百拇医药
For the measurement of FSH secretion, RAP cells were plated in 48-well plates (Costar, Cambridge, MA) (0.15 x 106 cells/well) and allowed to recover 3 d in ßPJ medium (51) supplemented with 2% FBS. The cells were washed, and test substances or conditioned medium was added to the wells and secretion was measured in a final volume of 0.5 ml. FSH was quantified using RIA reagents generously provided by Dr. Parlow through the National Pituitary and Hormone Distribution Program at NIDDK. All treatments were performed in triplicate, and data from at least three determinations were subjected to either the t test or ANOVA followed by Dunnett’s multiple comparison test, as indicated in the figure legends.
Reagentsx3x-, 百拇医药
The rh-activin-A was provided by Genentech, Inc. (San Francisco, CA); rhFS288 was obtained through the National Hormone and Pituitary Program of NIDDK; rhIL-1ß was a gift from Dr. Liu of Neurocrine Biosciences Inc. (La Jolla, CA); rhIL-6 was purchased from Endogen, Inc. (Woburn, MA), and LPS was from Sigma (Escherichia coli serotype 026:B6m code L3755, lot 20H4025).x3x-, 百拇医药
Resultsx3x-, 百拇医药
Selection and establishment of a rat anterior pituitary folliculostellate cell line, FS/D1hx3x-, 百拇医药
An indirect RIA was used for measuring follistatin secretion from rat anterior pituitary cells. As shown in two examples of a typical standard curve, the minimum concentration of free FS288 detectable by this indirect immunoassay was in the range of 0.05–0.1 ng/ml, and the assay was linear over approximately two log units (Fig. 1). Concentrations of activin A greater than 0.01 nM lowered the sensitivity of the assay because of competition with radiolabeled activin A, indicating that the assay provides an estimate of only free follistatin (Fig. 1A). This assay was used to measure relative changes in follistatin levels by comparing the linear portion of serially diluted media to the standard curve, as shown by a typical example of the analysis of a set of control and treated samples (Fig. 1B).
fig.ommitteedn|h73h!, 百拇医药
Figure 1. Typical FS288 standard curves obtained from an indirect RIA in which follistatin concentrations were quantified based on follistatin:[125I]activin A complexes immunoprecipitated with anti-FS288 (no. 5542). Approximately 100,000 cpm [125I]activin A were incubated with the indicated concentrations of FS288, with and without cold activin A (A), or with serially diluted medium conditioned by FS/D1h cells treated as indicated (B). The samples were immunoprecipitated with anti-FS288 sheep antirabbit IgG and processed as described in Materials and Methods. The values reflect the mean ± SEM of triplicate determinations from typical experiments.n|h73h!, 百拇医药
After culturing rat anterior pituitary cells for 2–3 wk in high FBS (15%), medium conditioned by the surviving cells contained no measurable GH, FSH, LH, or ACTH (data not shown). However, the medium was observed to contain an activity that suppressed basal FSH secretion and attenuated the effect of submaximal concentrations of activin A (Fig. 2A). Subjecting this population of cells to multiple dilution cloning steps, as described in Materials and Methods, yielded several independent lines that await further analysis. One of these lines, herein referred to as FS/D1h cells, was determined to express considerably more follistatin mRNA, compared with the other lines or to the mixed population of rat anterior pituitary cells (data not shown), and secrete substantial amounts of follistatin (Fig. 2B). Conditioned medium from the FS/D1h line when applied to cultured rat anterior pituitary cells also suppressed basal FSH secretion, consistent with the expected action of follistatin on gonadotropes (Fig. 2C).
fig.ommitteedf, 百拇医药
Figure 2. The suppression of FSH secretion by medium conditioned by enriched rat anterior pituitary folliculostellate cells (A) and the evaluation of medium conditioned by several selected folliculostellate lines for follistatin secretion (B) and effects on FSH secretion (C). A, Control or conditioned medium (0.3 ml) from enriched folliculostellate cells in 10-cm dishes was added to rat anterior pituitary cells cultured in 48-well plates (Costar) at a final volume of 0.5 ml/well, with and without various concentrations of activin A. B, Follistatin levels in medium conditioned for 7 d with the initial pool of cells or several selected folliculostellate lines were quantified by an indirect RIA described in the text. C, 0.2 ml control or conditioned medium by the same lines shown in panel B were tested for effects on FSH secretion as described for panel A. The effect of FS288 on FSH secretion is shown at the far right. The data are means ± SEM of triplicate determinations of representative experiments. The data within each panel were subjected to ANOVA and Dunnett’s multiple comparison test. For A, statistical comparisons are relative to control medium for each x-axis value. For B and C, comparisons are relative to FSH or follistatin (FS) secretion from the pool: **, P < 0.001; #, P < 0.01; *, P < 0.05.
Immunocytochemical evaluation revealed that all FS/D1h cells expressed the S100 antigen (Fig. 3). This cell-associated staining was no longer visible when anti-S100 was omitted or substituted with normal rabbit IgG, verifying the specificity of the S100 immunoreactivity (Fig. 3B). The FS/D1h cells were also positive for GFAP immunostaining, another marker of the anterior pituitary cell population (1) (data not shown). The FS/D1h cells had a doubling time of approximately 48 h, formed a monolayer, and essentially stopped growing when they reached confluence. Light microscopy indicated the presence of large vacuolar structures in the cytoplasm (data not shown). The FS/D1h cells established cell-cell contact via long cytoplasmic processes, and during the logarithmic growth phase, they had a more stellate appearance but became more rounded at higher densities (data not shown). Because folliculostellate cells have been proposed to be the source of pituitary IL-6, the production of this cytokine from FS/D1h cells was evaluated and levels were compared with those of RAP cells. The results indicated that FS/D1h cells secrete considerably more IL-6, compared with RAP cells (Fig. 4A). Moreover, the cells responded to IL-1ß by an increase in the production of not only IL-6 but also follistatin (Fig. 4B). These observations indicate that FS/D1h cells possess at least some of the characteristics usually associated with pituitary folliculostellate cells. The cells were amenable to propagation through multiple passages (up to 35–40 passages tested) without loss of phenotypic characteristics such as follistatin and IL-6 production, morphological appearance, and growth pattern; S100 immunoreactivity was not evaluated beyond passage no. 19.d\, http://www.100md.com
fig.ommitteedd\, http://www.100md.com
Figure 3. Immunocytochemical analysis for S100 protein expression in FS/D1h cells (passage no. 5). Texas Red fluorescence was captured at x40 magnification. B, The signal obtained when the primary S100 antiserum was omitted is shown.d\, http://www.100md.com
fig.ommitteedd\, http://www.100md.com
Figure 4. The secretion of IL-6 and follistatin from FS/D1h cells (passage no. 10) and cultured RAP cells in response to 0.5 nM IL-1ß. The data are means ± SEM of triplicate determinations of a representative experiment. The t test was used to compare IL-6 or follistatin (FS) secretion from the indicated cell types incubated with or without IL-1ß. ND, Not detectable. **, P < 0.001; #, P < 0.01.(Louise M. Bilezikjian Angela M. O. Leal Amy L. Blount Anne Z. Corrigan Andrew V. Turnbull and Wylie )