Human recombinant (rh-) TGFB1, rh-activin A, recombinant mouse (rm)-follistatin 288 were purchased from R&D systems (Minneapolis, MN). Dulbecco's modified Eagle medium (DMEM), Lipofectamine/Plus, Lipofectamine 2000, gentamycin, and Trizol were from Invitrogen (Carlsbad, CA). Fetal bovine serum (FBS) was from JRH Biosciences (Lenexa, KS). The anti-Smad3 affinity purified rabbit polyclonal antibody was purchased from Zymed (South San Francisco, CA). Anti-Smad2/3 and phospho-Smad2 affinity purified rabbit polyclonal antibodies were purchased from Upstate Biotech (Waltham, MA). The phospho-Smad3 rabbit polyclonal antibody was a generous gift of Dr. Michael Reiss (Robert Wood Johnson Medical School). Protease inhibitor tablets (CompleteMini) were purchased from Roche (Indianapolis, IN). Deoxynucleotide triphosphates (dNTPs), MMLV reverse transcriptase, random primer hexamers, and Taq polymerase were from Promega (Madison, WI). The -1990/+1 mFshb-luc reporter and constitutively active HA-rat ALK4 (Acvr1b) were described previously 32. HA-human TGFBR1(T204D) was provided by Dr. Peter Scheiffele (Columbia University). The 3TP-luc reporter and HA-human TGFBR2 expression construct were gifts of Dr. Joan Massague (Memorial Sloan Kettering Cancer Center).

Pituitaries from mice containing the ovine FSHB-H2K^K transgene were used for gonadotrope purification. Young mice (8–10 weeks old; 10–18 mice) or older mice (1 year old; 12 mice) were killed and their pituitaries were dispersed and the gonadotropes were purified as reported 50. Cells that did not attach to the magnetic column were labeled "gonadotrope-depleted," while cells eluted from the column after removal of the magnetic field were labeled "gonadotropes." Cell counts were obtained for all cell types using a hemocytometer. Equal numbers of cells were cultured in medium 199 (Gibco) with 10% charcoal-treated sheep serum and antibiotics/antimycotics as reported 51. Gonadotropes and gonadotrope-depleted cells purified from younger mice were plated in triplicate at a density of 18,000 cells per well (first experiment) or 30,000 cells per well (second and third experiments). Cells isolated from older mice were plated in triplicate at a density of 50,000 cells per well in two separate experiments. For treatments with activin A or TGFB1, purified gonadotropes or whole pituitary cells were plated in triplicate at a density of 10,000 cells per well in three separate experiments. Gonadotropes, gonadotrope-depleted cells, and whole pituitary cells were cultured in 200 μl of media in 96 well Primaria culture plates (Becton Dickinson & Co, Franklin Lakes, NJ). Cells were incubated at 37° under 5% CO₂ for 48 hrs prior to RNA isolation. All mice were handled in accordance with the rules and regulations of the Institutional Animal Care and Use Committee of North Carolina State University.

Immortalized murine gonadotrope LβT2 cells were provided by Dr. Pamela Mellon (University of California, San Diego) and were cultured as described previously 32. Murine fibroblast NIH3T3 cells were obtained from Dr. Patricia Morris (Population Council) and were cultured in DMEM/10% FBS. Cells were plated in 6- or 24-well plates at densities of 1 × 10⁶ or 2 × 10⁵ cells per well, respectively, approximately 36 hr prior to transfection. Cells were transfected with Lipofectamine/Plus or Lipofectamine 2000 following the manufacturer's instructions. Reporter plasmids were transfected at 1 μg (6-well) or 450 ng (24-well) per well. Expression plasmids were introduced at 300 ng (24-well) per well. In all experiments, the total amount of DNA added was balanced across treatments with empty expression vector pcDNA3.0 (Invitrogen).

In reporter experiments including ligand treatment, activin A or TGFB1 were added at the indicated concentrations for approximately 24 hr. Cells were washed with 1× PBS and lysed in 1× Passive Lysis Buffer (Promega). Luciferase assays were performed on a Luminoskan Ascent luminometer (Thermo Labsystems, Franklin, MA) as described 32. All transfection conditions were performed in triplicate and each experiment performed 2–3 times.

LβT2 and NIH3T3 cells were seeded at 7 or 4 × 10⁵ cells per well, respectively, in 6-well plates. After 24–48 hr., cells were washed with serum-free DMEM and then incubated in the same medium overnight. The following day, cells were treated with the indicated concentrations of activin A or TGFB1 in fresh serum-free DMEM for 1 hr. After a wash with PBS, whole cell lysates prepared in RIPA buffer containing protease inhibitors. Equivalent amounts of protein were separated by 8% Tris-glycine SDS-PAGE and transferred to Protran (Schleicher & Schuell, Keene, NH). Filters were probed with anti-phospho-Smad2, anti-phospho-Smad3, anti-Smad2/3, or anti-Smad3 using previously described methods 32.

Total RNA was extracted from adult female CD-1 murine pituitaries and LβT2 cells using Trizol following the manufacturer's instructions. Four μg of total RNA were reverse transcribed (RT) into cDNA using 100 ng random hexamer primers and 100 U MMLV-RT. A second set of samples was processed similarly, except the RT enzyme was omitted (no RT) as a control for contaminating genomic DNA in the RNA samples. One-tenth of each RT or RT- reaction was used as template in PCRs for Tgfbr1 (503 bp) and Tgfbr2 (536 bp). PCR was run using the following conditions for 35 cycles: 94C for 30 sec, 53C for 30 sec, and 72C for 30 sec. Reactions contained 0.4 pmol of each primer, 200 μM dNTPs, 1.5 mM MgCl₂, 1× PCR buffer, and 2.5 U Taq polymerase. Following a final 7 min. extension step at 72C, one-fifth of each reaction was resolved on a 1 % agarose gel containing ethidium bromide. Gels were photo-documented using a digital camera interfaced with an IBM ThinkPad computer running the Kodak Digital Science 1D software (v.2.0.2) software. Reactions with no template (H₂O only) were used to confirm the absence of contaminating DNA in the reagents. The primer sets for Tgfbr1 and Tgfbr2 were as follows: Tgfbr1, (forward) AACCTGTTGTATTGCAGACTT and (reverse) GAGCAGAGTTCCCACGGTGT; Tgfbr2, (forward) TTGCCTGTGTGACTTCGGGCT and (reverse) CTATTTGGTAGTGTTCAGCGA.

Total RNA from primary and LβT2 cells was isolated and converted to cDNA as reported 51. Oligonucleotides for Taqman real-time PCR were designed for murine cDNA using software from Integrated DNA Technologies, Inc (Coralville, IA) for Tgfbr1, Tgfbr2, Fshb and prolactin (Table 1). Using the same oligonucleotides as described previously 50, murine 18s ribosomal RNA served as the endogenous control. All Taqman probes were 5' -labeled with FAM and real-time PCR of all cDNA samples was performed at the same time. Real-time PCR was performed in duplicate on triplicate cDNA samples from both gonadotropes and gonadotrope-depleted cells using an iCycler (Bio-Rad, Inc). Samples were incubated at 95°C for 3 min, and then for 40 complete cycles (95°C for 30 sec, 55°C for 30 sec, and 72°C for 30 sec). There was a final extension step of 72°C for 3 min. Threshold cycle (C_T) values were determined with Bio-Rad software and used for relative quantitation with the 2^-ΔΔCt method 52.

The data from replicate luciferase assay experiments were highly similar and were pooled (n = 6 or 9 per treatment) for statistical analyses. Data are presented as fold-change from the control condition in each experiment. Differences between means were compared using one- or two-way analyses of variance followed by post-hoc Scheffe or Bonferroni tests (Systat 10.2, Richmond, CA). Comparisons of relative receptor mRNA expression in gonadotrope and gonadotrope-depleted cells in different age groups were performed with two-way ANOVAs of log-transformed data. Fshb mRNA levels were compared in one-way ANOVAs of log-transformed data. In all cases, significance was assessed relative to p < 0.05.

LβT2 cells were transfected with a murine -1990/+1 Fshb luciferase promoter-reporter construct (-1990/+1 mFSHB-luc) 32 and were treated with different concentrations of activin A or TGFB1 for approximately 24 hr. Whereas activin A dose-dependently stimulated reporter activity, TGFB1 had no effect at concentrations up to 5 nM (Fig. 1A). In addition, activin A, but not TGFB1, stimulated Smad 2 and 3 phosphorylation in these cells (Fig. 1B). In contrast, the same lot of TGFB1 at lower concentrations (4–400 pM) dose-dependently stimulated the activin/TGFB responsive promoter of 3TP-luc 53, and Smad2/3 phosphorylation in murine NIH3T3 fibroblast cells (Figs. 2A and 2B). Thus, the TGFB1 ligand was biologically active, but LβT2 cells were somehow insensitive to it.

We previously showed that a constitutively active form of rat Acvr1b (T206D), which can stimulate Smad phosphorylation in the absence of activins and the type II receptors 54, stimulated murine Fshb promoter-reporter activity 32. Here, we asked whether a constitutively active form of TGFBR1 (T204D; 55) could similarly stimulate Fshb transcription in LβT2 cells. As shown in Figure 3, both rat Acvr1b-TD and human TGFBR1-TD potently stimulated -1990/+1mFshb-luc. These data indicate that events downstream of TGFBR1 (whether Smad-dependent or Smad-independent; 3056) seem to be present in LβT2 cells and therefore that the cells' insensitivity to TGFB1 likely derives from a deficiency at the receptor level.

We used RT-PCR to examine Tgfbr2 and Tgfbr1 mRNA levels in LβT2 cells compared to adult murine pituitary glands. Whereas both LβT2 cells and pituitary glands expressed Tgfbr1 mRNA, only the latter expressed Tgfbr2 (Fig. 4).

RT-PCR analysis indicated that LβT2 cells do not express Tgfbr2 mRNA. Therefore, these cells may not respond to TGFB1 because of a deficiency in this receptor. It is possible, however, that additional mechanisms contribute to TGFB1 insensitivity. To address this issue, we transfected LβT2 cells with a human TGFBR2 expression construct and examined TGFB1-stimulated Fshb transcription. Over-expression of the receptor alone had no effect on basal transcription, but made it possible for TGFB1 to stimulate -1990/+1 mFshb-luc activity (Fig. 5). Therefore, a deficiency in Tgfbr2 expression appeared to account for the inability of LβT2 cells to respond to TGFB1.

LβT2 cells were derived from a pituitary tumor in a female transgenic mouse 57. Whereas these cells show many of the features of fully differentiated gonadotropes, they are transformed cells and exhibit clear differences from gonadotropes in vivo. For example, basal Fshb expression is substantially lower in LβT2 cells than in gonadotropes (personal observations). Therefore, it is possible that the Tgfbr2-deficiency observed in LβT2 cells may not accurately reflect receptor expression in gonadotropes in vivo, though previous analyses in rats indicated that within the pituitary, Tgfbr2 expression is most abundant in lactotropes 5859. In order to examine receptor expression in murine gonadotropes, we purified this cell type from male mice, aged eight to ten weeks (young) or 1 year of age (old), using a recently described transgenic model 50. Using real-time RT-PCR, prolactin (Prl) expression was examined to determine the level of purification of the gonadotropes from mixed primary pituitary cultures as described earlier 50. The level of purity ranged from 97 % to 99% (data not shown).

We then measured Tgfbr1 and Tgfbr2 mRNAs using real-time RT-PCR. Tgfbr1 mRNA was significantly higher in older than younger animals (p < 0.001), but did not differ significantly between gonadotropes and gonadotrope-depleted cells, nor was there a significant interaction between these two variables (Fig. 6A). In contrast to LβT2 cell data, Tgfbr2 mRNA was detected in gonadotropes but it was greater in older than younger animals (p < 0.001). Tgfbr2 mRNA was also higher in gonadotrope-depleted cells than pure gonadotropes across both age groups (p < 0.007) (Fig. 6B). In young mice, gonadotrope-depleted cells expressed Tgfbr2 6.5-fold higher than purified gonadotropes, whereas in the old mice the difference was reduced to 2.2-fold, but the interaction between cell type and age was not statistically significant. In the same assays, Tgfbr2 was undetectable in LβT2 cells, and Tgfbr1 in LβT2 cells was expressed at roughly 20% of Tgfbr1 in purified gonadotropes in young animals (data not shown).

Because purified gonadotropes express Tgfbr2, we used real-time RT-PCR to examine the effects of activin A and TGFB1 on endogenous Fshb mRNA levels. Treatment of purified gonadotropes from younger mice with activin A resulted in a 31-fold stimulation of Fshb mRNA (Fig. 7A). Surprisingly, treatment with TGFB1 resulted in a significant 95 %

reduction