Molecular Endocrinology. First published ahead of print September 24, 2013 as doi:10.1210/me.2013-1185

ORIGINAL RESEARCH

Forkhead Box O1 Is a Repressor of Basal and GnRH- Induced Fshb Transcription in Gonadotropes

Danalea V. Skarra, David J. Arriola, Courtney A. Benson, and Varykina G. Thackray

Department of Reproductive Medicine and the Center for Reproductive Science and Medicine, University of California, San Diego, LA Jolla, CA 92093

Synthesis of the gonadotropin beta subunits is tightly controlled by a complex network of hor- monal signaling pathways that may be modulated by metabolic cues. Recently, we reported that regulates FOXO1 phosphorylation and cellular localization in pituitary gonadotropes and that FOXO1 overexpression inhibits Lhb transcription. In the current study, we investigated whether FOXO1 modulates Fshb synthesis. Here, we demonstrate that FOXO1 represses basal and GnRH-induced Fshb transcription in L␤T2 cells. Additionally, we show that PI3K inhibition, which increases FOXO1 nuclear localization, results in decreased Fshb mRNA levels in murine primary pituitary cells. FOXO1 also decreases transcription from the human FSHB promoter, suggesting that FOXO1 regulation of FSHB transcription may be conserved between rodents and humans. Although the FOXO1 DNA binding domain is necessary for suppression of Fshb, we do not observe direct binding of FOXO1 to the Fshb promoter, suggesting that FOXO1 exerts its effect through -protein interactions with transcription factors required for Fshb synthesis. FOXO1 sup- pression of basal Fshb transcription may involve PITX1 since PITX1 interacts with FOXO1, FOXO1 repression maps to the proximal Fshb promoter containing a PITX1 binding site, PITX1 induction of Fshb or a PITX1 binding element in CV-1 cells is decreased by FOXO1, and FOXO1 suppresses Pitx1 mRNA and protein levels. GnRH induction of an Fshb promoter containing a deletion at -50/-41 or -30/-21 is not repressed by FOXO1, suggesting that these two regions may be involved in FOXO1 suppression of GnRH-induced Fshb synthesis. In summary, our data demonstrate that FOXO1 can negatively regulate Fshb transcription, and suggest that FOXO1 may relay metabolic hormonal signals to modulate gonadotropin production.

SH is produced by gonadotrope cells in the anterior FSH is a heterodimeric composed of a Fpituitary and plays a key role in mammalian fertility. common alpha subunit, shared with LH and TSH, as well In females, it is required for ovarian folliculogenesis, as a unique beta subunit that confers biological specific- while in males it promotes spermatogenesis in conjunc- ity. Transcription of the Fshb is dynamically regu- tion with (1, 2). Female Fshb knockout mice lated during the estrous cycle. Alterations in Fshb mRNA exhibit an arrest in ovarian folliculogenesis prior to the levels are observed prior to changes in serum FSH levels, antral stage, while males are fertile but have impaired indicating that transcription of Fshb is a rate-limiting step reproductive function (3). In contrast, the human FSHB for production of the mature (5, 6). Proper reg- gene appears to be critical for reproductive function in ulation of FSH levels is important for fertility. Low FSH both genders. Nonsynonymous mutations result in absent levels are associated with defective follicular growth, or incomplete pubertal development and infertility in while high levels are associated with premature ovarian women, as well as azoospermia and infertility in men (4). failure (7).

ISSN Print 0888-8809 ISSN Online 1944-9917 Abbreviations: Abbreviations: AP1 activator protein 1; AP1BE AP1 binding element; ␤-gal Printed in U.S.A. ␤-galactosidase; DBD DNA binding domain; DMEM Dulbecco’s Modification of Eagles Copyright © 2013 by The Endocrine Society Medium; EGR1 early growth response protein 1; FBE forkhead binding element; FBS fetal Received June 19, 2013. Accepted September 18, 2013. bovine serum; GFP green fluorescent protein; HD homeodomain; HDBE HD binding ele- ment; HSD Honestly Significant Difference; LHX LIM gene; luc luciferase; NFY nuclear Y; PBS phosphate buffered saline; PITX paired-like homeodo- main transcription factor; PITXBE PITX binding element; SF1

doi: 10.1210/me.2013-1185 Mol Endocrinol mend.endojournals.org 1

Copyright (C) 2013 by The Endocrine Society 2 FOXO1 Suppresses Fshb Mol Endocrinol

Basal Fshb transcription involves multiple transcrip- Recently, we reported that insulin signaling regulated tion factors, including LIM homeobox gene (LHX), nu- FOXO1 phosphorylation in a PI3K-dependent manner in clear transcription factor Y (NFY) and paired-like home- immortalized gonadotropes, similarly to what has been odomain transcription factor (PITX) (8). LHX3 was observed in other cells (26, 35, 36). Since we also discov- reported to bind the porcine Fshb promoter at several ered that overexpression of FOXO1 suppressed basal and binding sites, including one element which is highly con- GnRH-induced Lhb transcription, the purpose of this served among mammals (9). Additionally, NFY was study was to determine whether FOXO1 can regulate shown to bind to a site at –76/-72 in the murine Fshb Fshb synthesis. We show that overexpression of FOXO1 promoter and mutation of this site decreased basal Fshb in L␤T2 cells results in decreased basal and GnRH-in- synthesis by 30% (10). Furthermore, Lamba et al, showed duced transcription from murine and human FSHB-luc that PITX1/2 binds to a conserved element at –54/-49 of reporters as well as endogenous Fshb mRNA. In addition, the murine Fshb promoter and that mutation of this bind- we show that inhibition of PI3K, which increases FOXO1 ing element reduced basal Fshb gene expression (11). nuclear localization (35), results in decreased basal and Hormonal regulation of Fshb gene expression has been GnRH-induced Fshb mRNA in murine primary pituitary shown to occur via multiple signaling pathways (12). cells. We also demonstrate that the FOXO1 suppression GnRH increases Fshb transcription via PKC and MAPK maps to the proximal Fshb promoter. Moreover, our data signaling pathways through the induction of the activator suggest that the FOXO1 suppression occurs through pro- protein 1 (AP1) transcription factors, c-FOS and JUNB, tein-protein interactions with factors necessary for Fshb which bind and induce the Fshb promoter, along with transcription such as PITX1, since the FOXO1 DNA c-JUN and FOSB (13–18). AP1 also interacts with factors binding domain (DBD) is required for the suppression but involved in basal expression of Fshb, such as NFY, in the FOXO1 does not appear to bind directly to the proximal mouse and USF1, in the rat to integrate GnRH respon- Fshb promoter. This idea is further supported by the fact siveness (16, 19). Activin is also an inducer of the Fshb that PITX1 induction of Fshb or a PITX1 binding element gene (20). Activin signaling through SMAD and FOXL2 in CV-1 cells is suppressed by FOXO1, that PITX1 inter- also appears to be critical for transcriptional reg- acts with FOXO1, and that two regions of the proximal ulation of the Fshb gene (8, 21). Fshb promoter (-50/-41 and –30/-21) appear to be neces- In addition to reproductive , metabolic hor- sary for FOXO1 suppression of GnRH-induced Fshb mones may regulate FSH production at the level of the transcription. gonadotrope. While insulin, IGF1 and treatment of rat primary pituitary cells have been reported to increase FSH, as well as LH levels (22–25), it is not clear whether Materials and Methods this is due to an effect of these hormones on Fshb tran- scription vs production and/or secretion of the mature Murine Primary Pituitary Cell Culture hormone. One possibility is that metabolic hormones reg- Seven-week old, male C57 Black 6 mice (Harlan Laborato- ulate Fshb transcription in gonadotropes through the ac- ries, Indianapolis, IN) were housed in the UCSD vivarium for tivity of downstream effectors such as the FOXO subfam- one week under standard conditions. All animal procedures were conducted in accordance with the UCSD Institutional An- ily of forkhead transcription factors. This family is imal Care and Use Committee requirements. Eight mice were composed of FOXO1, 3, 4 and 6. The transcriptional sacrificed and their pituitaries were collected in ice-cold Dulbec- activity of FOXOs is controlled by posttranslational co’s A Phosphate-buffered saline (PBS). After a PBS rinse, the modifications in multiple cell and tissue types (26–28). pituitaries were minced on ice with fine scissors and then placed FOXO phosphorylation has been shown to be regulated in dissociation media containing phosphate buffered 0.25% by insulin, IGF1 and leptin (29–31). Insulin stimulation trypsin-EDTA (Gibco/Life Technologies, Grand Isle, NY) and 0.25% collagenase (Life Technologies). The pituitaries were induces PI3K/AKT activation, triggering AKT phosphor- shaken for 30 minutes at 37°C in a water bath, then an equal ylation of FOXO1 on residues Thr24, Ser256, and Ser319 volume of Dulbecco’s Modification of Eagles Medium (DMEM) (32). This results in FOXO1 export from the nucleus to (Mediatech Inc., Herndon, VA) containing 10% fetal bovine the cytoplasm, where it is sequestered by 14–3–3 pro- serum (FBS) (Omega Scientific, Inc., Tarzana, CA) was added teins, preventing FOXO1 transcriptional activity. While along with DNaseI (Worthington Biochemical, Lakewood, NJ) ␮ considerable progress has been made in understanding at a final concentration of 25 g/mL and incubated for another 15 minutes at 37°C. After removal of tissue debris, the cells were the role of FOXOs in reproductive tissues such as the pelleted by centrifugation and plated at a density of 6 ϫ 105/2 and uterus (33, 34), FOXO regulation of gonado- cm2 well in Primaria plates (BD Biosciences, San Jose, CA). tropin hormone production in pituitary gonadotrope cells After 24 hours in 10% FBS DMEM, the cells were changed to remains largely unexplored. serum-free media 18 hours before treatment. Cells were pre- doi: 10.1210/me.2013-1185 mend.endojournals.org 3 treated with DMSO or 50 ␮M LY294002 (EMD Biosciences, Plasmid Constructs San Diego, CA) for 1 hour, then treated with 0.1% BSA vehicle The pcDNA3 Flag human FOXO1 and pcDNA3 Flag or 30 nM GnRH (Sigma-Aldrich Co., St. Louis, MO) along with FOXO1-CA expression vectors were previously described (32). DMSO or LY294002 for an additional 5 hours, after which the We obtained the pcDNA3 FOXO1-⌬DBD (⌬208–220) from cells were lysed to obtain total RNA. Dr. William Sellers (Addgene plasmid 10694). The pcDNA3 murine PITX1 expression vector was described previously (41). Immortalized Cell Culture The construction of the –1000 murine Fshb luciferase (luc) re- Ј Cell culture was performed using the L␤T2 cell line which porter plasmid and 5 truncations were described previously has many characteristics of a mature, differentiated gonado- (16, 42). The 10 bp deletions of the –398 Fshb promoter from trope (37, 38). CV-1 cells lacking PITX1 were also used (39, 40). –70/-11 were generously provided by Dr. Djurdjica Coss. The ϩ The cells were maintained in 10 cm plates in DMEM with 10% –1028/ 7 human FSHB-luc reporter plasmid was provided by FBS and penicillin/streptomycin antibiotics (Gibco/Invitrogen, Dr. Daniel Bernard. The 4xFBE-luc was constructed by insert- Grand Island, NY) at 37°C and 5% CO . ing four repeats of a consensus forkhead binding element (FBE) 2 (CCGTAAACAACT) upstream of a minimal –81 Herpes sim- plex virus thymidine kinase promoter in pGL3 using KpnI and Adenoviral Infection NheI restriction enzyme sites while the 4xHDBE-luc contains Adenoviral vectors containing cDNA of green fluorescent four repeats of a consensus bicoid-related homeodomain bind- protein (Ad-GFP) and constitutively active FOXO1 (T24A/ ing element (HDBE) (ACTAATCCCT) (39). The elements are S256A/S319A) (Ad-FOXO1-CA) were generously provided by underlined. Sequences were confirmed by dideoxyribonucle- ␤ ϫ 6 Dr. Domenico Accili. L T2 cells were seeded at 2 10 cells/ otide sequencing. well on 6-well plates. The next morning, cells were transduced with a multiplicity of infection of 200 of Ad-GFP or Ad- Mutagenesis FOXO1-CA for 6 hours, then switched to serum-free DMEM. The QuikChange Site-Directed Mutagenesis Kit (Stratagene, 24 hours after adenoviral infection, cells were treated with ve- La Jolla, CA) was used to generate mutations in a plasmid con- hicle (0.1% BSA) or 10 nM GnRH for 1 or 6 hours as noted. taining the murine Fshb promoter. Mutagenesis was performed with the following mutations in the –54/-49 PITX1 and –72/-69 Quantitative RT-PCR AP1 binding elements, respectively: Total RNA was extracted from L␤T2 cells with TRIzol Re- 5Ј-GTGGATCGAC-3Ј and 5Ј-ATTGGTCCCG-3Ј (mutated agent (Life Technologies, Carlsbad, CA) following the manu- bases are indicated in bold). The sequences of the promoters facturer’s protocol. Contaminating DNA was removed with were confirmed by dideoxyribonucleotide sequencing. DNA-free reagent (Life Technologies). 2 ␮g of RNA was re- verse-transcribed using the iScript cDNA Synthesis Kit (Bio-Rad Transient Transfection Laboratories, Inc., Hercules, CA) according to the manufactur- L␤T2 cells were seeded at 4.5 ϫ 105 cells/well on 12-well er’s protocol. Quantitative real-time PCR was performed in an plates and transfected 18 hours later, using Polyjet (SignaGen iQ5 iCycler using iQ SYBR Green Supermix (Bio-Rad Labora- Laboratories, Rockville, MD) following the manufacturer’s in- tories, Inc.) and the following primers: Fshb forward, GC- structions. CV-1 cells were seeded at 1.5 ϫ 105 cells/well. For all CGTTTCTGCATAAGC; Fshb reverse, CAATCTTACG- experiments, the cells were transfected with 400 ng of the indi- GTCTCGTATACC; Pitx1 forward, cated luc reporter plasmid and 200 ng of a ␤-galactosidase (␤- GTCCGACGCTGATCTGCCA; Pitx1 reverse, TTGCCGC- gal) reporter plasmid driven by the Herpes Virus thymidine CGCTGTTTCTTCTT; c-Fos forward, GGCAAAGTAGAG- kinase promoter to control for transfection efficiency. The cells CAGCTATCTCCT; c-Fos reverse GGCAAAGTAGAGCAGC- were switched to serum-free DMEM containing 0.1% BSA, 5 TATCTCCT; c-Jun forward, mg/L transferrin, and 50 mM sodium selenite 6 hours after GTCCCCTATCGACATGGAGTCT; c-Jun reverse, transfection. After overnight incubation in serum-free media, GAGTTTTGCGCTTTCAAGGTTT; Gapdh forward, TG- the cells were treated with vehicle (0.1% BSA) or 10 nM GnRH CACCACCAACTGCTTAG; Gapdh reverse, GGATG- for 6 hours, as indicated. CAGGGATGATGTTC; under the following conditions: 95°C for 5 minutes, followed by 40 cycles at 95°C for 45 seconds, Luciferase and ␤-galactosidase Assays 56°C for 45 seconds, and 72°C for 45 seconds except for an To harvest the cells, they were washed with 1xPBS and lysed annealing temperature of 61°C for Pitx1. Each sample was as- with 0.1 M K-phosphate buffer pH 7.8 containing 0.2% Triton sayed in triplicate and the experiment was conducted at least X-100. Lysed cells were assayed for luc activity using a buffer three times. Standard curves with dilutions of a plasmid con- containing 100 mM Tris-HCl pH 7.8, 15 mM MgSO4,10mM taining Fshb, Pitx1, c-Fos, c-Jun, or Gapdh cDNA were gener- ATP, and 65 ␮M luciferin. ␤-Gal activity was assayed using the ated with the samples in each run. In each experiment, the Tropix Galacto-light assay (Applied Biosystems, Foster City, amount of Fshb, Pitx1, c-Fos, or c-Jun was calculated by com- CA), according to the manufacturer’s protocol. Both assays paring the threshold cycle obtained for each sample with the were measured using a Veritas Microplate Luminometer (Pro- standard curve generated in the same run. Replicates were av- mega, Madison, WI). eraged and divided by the mean value of Gapdh in the same sample. After each run, a melting curve analysis was performed Statistical Analyses to confirm that a single amplicon was generated for each primer Quantitative RT-PCR and transient transfection experi- pair. ments were performed in triplicate and each experiment was 4 FOXO1 Suppresses Fshb Gene Expression Mol Endocrinol

repeated at least three times. For the transient transfections, the GTGGGCAACTGTGACTAGTAGAACACGG-3Ј with the data were normalized for transfection efficiency by expressing mutations underlined. luc activity relative to ␤-gal and relative to the empty reporter plasmid to control for hormone effects on the vector DNA. The GST Interaction Assay data were analyzed by Student’s t test for independent samples GST-PITX1 was provided by Dr. Jacques Drouin and the or one-way analysis of variance (ANOVA) followed by post hoc GFP expression vector by Dr. Douglass Forbes. 35S-labeled pro- comparisons with the Tukey-Kramer Honestly Significant Dif- teins were produced using the TnT Coupled Reticulolysate Sys- ference (HSD) test using the statistical package JMP 10.0 (SAS, tem. Bacteria transformed with GST plasmids were grown to Cary, NC). Significant differences were designated as P Ͻ .05. OD of 0.6 and induced with IPTG overnight at 30°C (44). Bacterial pellets were sonicated in 0.1% Triton X-100, 5 mM Immunofluorescence EDTA in 1x PBS, centrifuged and the supernatant was bound to glutathione sepharose 4B resin (Amersham Pharmacia Biotech, L␤T2 cells were seeded onto poly L-lysine coverslips (BD Piscataway, NJ). Beads were washed 4x in PBS and in HND Biosciences, Franklin Lakes, NJ) at 4.5 ϫ 105 cells/well in 12- buffer (10 mg/ml BSA, 20 mM Hepes pH 7.8, 50 mM NaCl, 5 well plates and then transfected 18 hours later with pcDNA3 mM DTT, and 0.1% NP-40). For the interaction assay, 20 ␮lof Flag human FOXO1 or pcDNA3 Flag FOXO1-CA expression 35S-labeled in vitro transcribed and translated FOXO1, vectors. Six h after transection, the cells were switched to serum- FOXO1-⌬DBD or GFP was added to the beads with 400 ␮lof free DMEM containing 0.1% BSA, 5 mg/L transferrin, and 50 HND buffer. Beads were incubated overnight at 4°C, washed 2x mM sodium selenite for overnight incubation. The next day, the with HND buffer and 2x with 0.1% NP-40 in PBS. Thirty ␮lof cells were washed twice with PBS and fixed with 4% parafor- 2x Laemmli load buffer was added, the samples were boiled and maldehyde for 10 minutes. Cells were then washed in PBS twice electrophoresed on a 10% SDS-polyacrylamide gel. One fourth and permeabilized with Nonidet P-40 solution (PBS containing of the 35S-labeled in vitro transcribed-translated product was 0.2% Nonidet P-40, 20% goat serum, 1% BSA) for 1 hour at loaded onto the gel as input. room temperature. Cells were washed twice in PBS and then incubated with mouse anti-Flag (Sigma, 1:100, F3165) primary Western Blot Analysis antibody in blocking buffer (PBS containing 20% goat serum, Whole cell extracts from adenoviral-infected L␤T2 cells were 1% BSA) for 24 hours at 4°C. Cells were washed 3 times with analyzed by western for PITX1, c-JUN, c-FOS and FOXO1. PBS for 5 minutes and incubated with Alexa488-conjugated Cells were washed twice with ice cold PBS, then harvested by goat antimouse secondary antibody (Invitrogen, 1:300, incubating in lysis buffer [10 mM Tris-HCl, pH 7.4, 150 mM A11001) in blocking buffer for 1 hour at room temperature. NaCl, 1% Nonidet P-40, 1 mM EDTA, 1 mM phenylmethyl- Cells were washed 3 times with PBS for 5 minutes and incubated sulfonyl fluoride, complete protease inhibitor mixture pellet with 300 nM DAPI (Invitrogen) for 4 minutes, then washed 3 (Roche Molecular Biochemical, Indianapolis, IN) and phospha- times with PBS for 5 minutes each. Coverslips were mounted tase inhibitor mixture pellet (Roche)] for 10 minutes at 4°C. using Prolong Gold (Invitrogen) and cells were viewed using a Lysates were centrifuged at 20,000 x g for 30 minutes. The Nikon Eclipse TE 2000-U inverted fluorescence microscope. protein concentration was determined by Bradford assay. An Digital images were collected using a CoolSNAP EZ cooled equal amount of protein per sample was loaded on a 10% SDS- CCD camera (Roper Scientific, Trenton, NJ) and analyzed with PAGE gel. Proteins were resolved by electrophoresis and trans- the Version 2.3 NIS-elements image analysis system. ferred onto a polyvinylidene difluoride membrane (Millipore, Billerica, MA). Membranes were blocked overnight in 5% non- Electrophoretic Mobility Shift Assay (EMSA) fat milk, then incubated overnight at 4°C with goat antihuman PITX1 (1:500, sc-18922), rabbit antihuman c-JUN (1:1000; sc- Flag-FOXO1-CA was transcribed and translated using a 1694), rabbit antihuman c-FOS (1:1000; sc-52), rabbit antihu- TnT Coupled Reticulolysate System (Promega Corp., Madison, man FOXO1 (1:1000; sc-11350), or rabbit antihuman ␤-Tubu- WI). The oligonucleotides were end-labeled with T4 polynucle- lin (1:3000, sc-9104) primary antibodies (Santa Cruz otide kinase and [␥-32P] ATP. 4 ␮L of TnT lysate was incubated Biotechnology). Blots were incubated with an antirabbit or an- with 1 fmol of 32P-labeled oligo at 4°C for 30 minutes in a tigoat horseradish peroxidase-linked secondary antibody (1: DNA-binding buffer [10 mM Hepes pH 7.8, 50 mM KCl, 5 mM 10000; Santa Cruz Biotechnology) and bands were visualized ␮ MgCl2, 0.1% Nonidet P-40, 1 mM dithiothreitol, 2 g poly(dI- using the SuperSignal West Dura Substrate (Thermo Scientific, dC), and 10% glycerol]. After 30 minutes, the DNA binding Rockford, IL). Densitometric analysis of band intensity was per- reactions were run on a 5% polyacrylamide gel (30:1 acryl- formed with ImageJ software (National Institute of Health, amide: bisacrylamide) containing 2.5% glycerol in a 0.25x TBE Bethesda, MD). buffer. Murine anti-Flag M2 antibody was used to supershift Flag-FOXO1-CA; murine IgG was used as a control for non- specific binding; 100x cold oligo was used as a competitor. The Results following oligonucleotides were used for EMSA: –95/-61 5Ј- Ј CAGCAGGCTTTATGTTGGTATTGGTCATGTTAACA-3 , Overexpression of Constitutively Active FOXO1 –65/-31 5Ј-TAACACCCAGTAAATCCACAGGGTTTTA- and PI3K Inhibition Decreases Basal and GnRH- AGTTTGT-3Ј, –35/-1 5Ј-TTTGTATAAAAGATGAGGTGTA- ACTTGACTCAGTG-3Ј, the consensus FBE (43) 5Ј- Induced Fshb mRNA Levels CTAGATGGTAAACAACTGTGACTAGTAGAACACGG-3Ј To ascertain whether the FOXO1 transcription factor and the mutated consensus FBE 5Ј-CTAGATG- can suppress Fshb gene expression, we transduced L␤T2 doi: 10.1210/me.2013-1185 mend.endojournals.org 5 cells with adenovirus (Ad) expressing GFP or constitu- We then determined whether repression of Fshb tran- tively active FOXO1 (FOXO1-CA) and measured Fshb scription could be observed in murine primary pituitary mRNA levels using quantitative RT-PCR. FOXO1-CA is cells. Since there is no pharmacological activator of localized in the nucleus due to the inability of insulin/ FOXO1, we used a PI3K inhibitor to mimic constitutively growth factor signaling to phosphorylate the mutated res- active FOXO1 since inhibition of PI3K results in FOXO1 idues (Figure 1A). As observed previously (45), GnRH nuclear localization in gonadotropes (35). As shown in induced Fshb transcription by 3 fold (Figure 1B). We also Figure 1C, treatment of dispersed pituitary cells with 30 demonstrated that Fshb mRNA synthesis was signifi- nM GnRH resulted in a significant 1.7-fold induction of cantly decreased in cells infected with Ad-FOXO1-CA vs Fshb. Inhibition of PI3K with 50 ␮M LY294002 resulted Ad-GFP (Figure 1B). Basal Fshb mRNA levels were re- in a substantial decrease in basal and GnRH-induced duced by 67% while GnRH-induced Fshb transcription Fshb mRNA levels (79% and 95%, respectively). Al- was decreased by 84%. These results indicate that though pharmacological inhibition of PI3K may affect FOXO1 can suppress endogenous Fshb synthesis in the multiple PI3K/AKT targets (46), these results demon- context of the native chromatin. strate, for the first time, that inhibition of PI3K results in

A

B β

C

FIGURE 1. Constitutively Active FOXO1 and Inhibition of PI3K Reduces Basal Transcription and GnRH Induction of Fshb mRNA in Gonadotropes. A. Diagram illustrating wild-type FOXO1 and constitutively active FOXO1-CA (T24A/S256A/S319A). DBD, DNA binding domain; NLS, nuclear localization signal; NES, nuclear export signal. B. L␤T2 cells were transduced with Ad-GFP or Ad-FOXO1-CA for 6 hours, then switched to serum- free media. 24 hours after adenoviral infection, cells were treated with 0.1% BSA vehicle (Veh) or 10 nM GnRH for 6 hours, as indicated. The results represent the mean Ϯ SEM of three experiments performed in triplicate and are presented as amount of Fshb mRNA relative to Gapdh.* indicates that Fshb transcription is significantly repressed by FOXO1-CA compared to GFP using Student’s t test. C. After overnight incubation in serum-free media, dispersed murine primary pituitary cells were pretreated with DMSO or 50 ␮M LY294002 for 1 hour, then treated with Veh or 30 nM GnRH along with DMSO or LY294002 for an additional 5 hours. The results represent the mean Ϯ SEM of four experiments performed in triplicate and are presented as amount of Fshb mRNA relative to Gapdh. # indicates that GnRH significantly increased Fshb mRNA levels compared to Veh using Student’s t test. * indicates that Fshb transcription is significantly repressed by LY294002 compared to DMSO using Student’s t test. 6 FOXO1 Suppresses Fshb Gene Expression Mol Endocrinol decreased basal and GnRH-induced Fshb transcription in A murine primary pituitary cells, potentially through in- creased nuclear localization of FOXO1.

FOXO1 Suppresses Basal Transcription and GnRH Induction of Murine Fshb-luc Since FOXO1-CA suppressed endogenous Fshb mRNA levels, FOXO1 or FOXO1-CA were overex- pressed in L␤T2 cells to determine whether FOXO1 reg- ulates the Fshb promoter. Initially, we determined the cellular localization of pcDNA3 Flag human FOXO1 and pcDNA3 Flag human FOXO1-CA transiently transfected B in L␤T2 cells and incubated in serum-free media. As shown in Figure 2A, FOXO1 is predominantly localized in the cytoplasm with some nuclear localization while

FOXO1-CA is localized in the nucleus. We then tested β whether FOXO1 transfected into L␤T2 cells could induce transcription from a minimal promoter containing a char- acterized forkhead binding element (FBE). FOXO1 in- creased transcription of four copies of a consensus FBE linked to a luc reporter (4xFBE-luc) (Figure 2B). In con- C trast, when 200 ng of FOXO1 or FOXO1-CA were tran- siently transfected with the –1000 murine Fshb promoter linked to a luc reporter (mFshb-luc), we observed repres- sion of Fshb synthesis (Figure 2C). More specifically,

basal Fshb transcription was decreased by 50% and 99%, β respectively (Figure 2D) while GnRH-induced Fshb was reduced by 63% and 29%, respectively (Figure 2E).

FOXO1 Suppresses Transcription of Human FSHB-luc Given that the proximal murine and human FSHB pro- DE moters are highly conserved (8), we hypothesized that FOXO1 would have a similar repressive effect on human FSHB gene expression as on murine Fshb. When 200 ng of FOXO1 was transiently transfected with the –1028/ϩ7 human FSHB promoter linked to a luciferase reporter (hFSHB-luc), we observed repression of basal and GnRH- induced FSHB synthesis (Figure 3A-C). Basal FSHB tran- scription was reduced 38% by FOXO1 while GnRH- induced FSHB gene expression was reduced by 52%.

FOXO1 Suppression Maps to the Proximal Fshb Promoter FIGURE 2. FOXO1 Suppresses Basal Transcription and GnRH Induction of the Murine Fshb-luc in L␤T2 Cells. A. The pcDNA3 Flag Since the proximal Fshb promoter contains PITX1, human FOXO1 and the pcDNA3 Flag human FOXO1-CA expression NFY and AP-1 binding elements necessary for basal and vectors were transfected into L␤T2 cells and then incubated overnight GnRH induction of Fshb (Figure 4A), we determined in serum-free media. Immunofluorescence was performed with a mouse anti-Flag primary antibody and an Alexa488-conjugated goat which regions of the Fshb promoter were required for antimouse secondary antibody. DAPI was used as a nuclear marker. FOXO1 suppression using 5Ј truncation analysis of the Representative images were obtained using a Nikon Eclipse TE2000-U promoter. Basal transcription and GnRH induction were inverted fluorescence microscope at 60x magnification. B. The 4xFBE- luc reporter was transiently transfected into L␤T2 cells along 200 ng of measured using –1000, –500, –304, –95, and –64 Fshb- pcDNA3 empty vector (EV) or FOXO1 expression vectors, as indicated. luc reporter plasmids (Figure 4B). FOXO1 suppression of After overnight incubation in serum-free media, cells were treated for doi: 10.1210/me.2013-1185 mend.endojournals.org 7 basal Fshb gene expression and GnRH induction still oc- The DNA-Binding Domain of FOXO1 Is Required curred with the –64 Fshb-luc (Figure 4C-D), suggesting for Suppression of Fshb Synthesis that elements present in the 64 base pairs upstream of the To determine whether FOXO1 repression requires the transcription start site are sufficient for FOXO1 repres- FOXO1 DBD, we tested whether a DNA-binding defi- sion of basal and GnRH-induced Fshb transcription. cient FOXO1, termed FOXO1- ⌬DBD (⌬208–220) (Fig- ure 5A), could elicit a repressive response. As a control for the level of protein expression, we showed that FOXO1 and FOXO1-⌬DBD are stably expressed when trans- Legend to Figure 2 Continued. . . fected into L␤T2 cells (Figure 5B). We then demonstrated 6 hours with 0.1% BSA vehicle (Veh) or 10 nM GnRH. The results that while overexpression of FOXO1 reduced Fshb gene Ϯ represent the mean SEM of three experiments performed in expression, the FOXO1-⌬DBD did not alter basal Fshb triplicate and are presented as luc/␤-gal. * indicates that the induction by FOXO1 is significantly different from the empty vector using synthesis (Figure 5C) or suppress GnRH-induced Fshb Student’s t test. B-D. The –1000 murine Fshb-luc reporter was gene expression (Figure 5D). These results indicate that transfected into L␤T2 cells along with EV, FOXO1 or FOXO1-CA (CA), as indicated. After overnight incubation in serum-free media, cells were treated for 6 hours with Veh or 10 nM GnRH. The results A represent the mean Ϯ SEM of three experiments performed in triplicate and are presented as luc/␤-gal (C), basal transcription relative to EV (D) or fold GnRH induction relative to the vehicle control (E). The different lowercase letters indicate that Fshb-luc transcription is B significantly repressed by FOXO1 or FOXO1-CA compared to EV using one-way ANOVA followed by Tukey’s HSD post hoc test.

A β

β C

B C D

FIGURE 4. FOXO1 Suppression Maps to the Proximal Fshb Promoter. A. Diagram illustrating the location of binding elements for NFY, AP1 FIGURE 3. FOXO1 Suppresses Basal Transcription and GnRH and PITX1 transcription factors on the murine Fshb Promoter. B-D. The Induction of Human FSHB-luc. A-C. The –1028/ϩ7 human FSHB-luc –1000, –500, –304, –95, and –64 murine Fshb-luc reporters were reporter was transfected into L␤T2 cells along with pcDNA3 empty transiently transfected into L␤T2 cells along with pcDNA3 empty vector vector (EV) or FOXO1, as indicated. After overnight incubation in (EV) or FOXO1, as indicated. After overnight incubation in serum-free serum-free media, cells were treated for 6 hours with 0.1% BSA media, cells were treated for 6 hours with 0.1% BSA vehicle (Veh) or vehicle (Veh) or 10 nM GnRH. The results represent the mean Ϯ SEM 10 nM GnRH. The results represent the mean Ϯ SEM of three of three experiments performed in triplicate and are presented as luc/ experiments performed in triplicate and are presented as luc/␤-gal (B), ␤-gal (A), basal transcription relative to empty vector (B) or fold GnRH basal transcription relative to empty vector (C) or fold GnRH induction induction relative to the vehicle control (C). * indicates that FSHB-luc relative to the vehicle control (D). * indicates that Fshb-luc transcription transcription is significantly repressed by FOXO1 compared to EV using is significantly repressed by FOXO1 compared to EV using one-way Student’s t test. ANOVA followed by Tukey’s HSD post hoc test. 8 FOXO1 Suppresses Fshb Gene Expression Mol Endocrinol the DBD of FOXO1 is necessary for FOXO1 repression FOXO1 Does Not Bind to the Proximal Fshb of Fshb basal transcription and GnRH induction on the Promoter via a High Affinity Binding Site Fshb promoter and suggest that FOXO1 exerts an effect Since the FOXO1 repression mapped to the proximal by either binding to the Fshb promoter or indirectly Fshb promoter and required the FOXO1 DBD, we per- through FOXO1 DBD interaction with other factors. formed EMSA to determine whether FOXO1 could bind to the proximal promoter in vitro. Three 35-mer oligonu- cleotide probes were designed to span the –95/-1 region A ∆ ∆ relative to the transcription start site. Flag-FOXO1-CA, synthesized with TnT rabbit reticulocyte lysate, bound to an oligonucleotide probe containing a consensus FBE (Figure 6A, lane 1) but not to probes encompassing the –95/-1 region of the Fshb promoter (Figure 6A, lanes 4, 7, B and 10). TnT Flag-FOXO1 also bound to a consensus ∆ FBE (data not shown). To identify which complex con- tained the Flag-FOXO1-CA bound to the FBE, we super- shifted the complex with an anti-Flag antibody (Figure β 6A, lane 3) but not with control IgG (Figure 6A, lane 2). This complex was not present when rabbit reticulocyte C lysate containing the pcDNA3 empty vector was used (data not shown). This complex also showed evidence of self-competition (Figure 6B, lane 14) but did not compete with a mutated consensus FBE (Figure 6B, lane 15). Incu- bation with oligos encompassing the –95/-1 region of the Fshb promoter did not result in competition (Figure 6B, lanes 16–18). These results suggest that, in contrast to the consensus FBE, FOXO1 does not bind to the proximal ∆ Fshb promoter via a high affinity binding site.

D AB

∆ FIGURE 5. DNA Binding Domain of FOXO1 Is Required to Suppress Basal and GnRH-induced Fshb Gene Expression. A. Diagram illustrating FOXO1-⌬DBD (⌬208–220). B. L␤T2 cells were transfected with pcDNA3 empty vector (EV), pcDNA3-FOXO1 (WT) or pcDNA3-FOXO1- ⌬DBD for 6 hours, then switched to serum-free media. 24 hours after transfection, the cells were harvested. Western blot analysis was performed on whole cell extracts using FOXO1 and ␤-Tubulin primary antibodies and a horseradish peroxidase-linked secondary antibody. A representative image is shown. C-D. The –1000 murine Fshb-luc FIGURE 6. FOXO1 Does Not Bind Directly to the Proximal Fshb reporter was transfected into L␤T2 cells along with EV, FOXO1 or Promoter. TnT Flag-FOXO1-CA was incubated with a consensus FBE, FOXO1-⌬DBD, as indicated. After overnight incubation in serum-free –95/-61, –65/31, or –35/-1 Fshb probes and tested for complex media, cells were treated for 6 hours with 0.1% BSA or 10 nM GnRH. formation in EMSA. A. FOXO1-CA-DNA complex on the FBE is shown The results represent the mean Ϯ SEM of three experiments performed in lane 1, while anti-Flag supershift is shown in lane 3 and IgG control in triplicate and are presented as basal transcription relative to empty in lane 2. The FOXO1-CA-DNA complex (arrow), antibody supershift vector (C) or fold GnRH induction relative to the vehicle control (D). * (ss) and nonspecific binding of proteins (ns) are indicated on the left of indicates that Fshb-luc transcription is significantly repressed by FOXO1 the gel. B. Self-competition with excess cold FBE is shown in lane 14, compared to EV or FOXO1-⌬DBD using one-way ANOVA followed by lack of competition with mutant FBE, –95/-61, –65/31, or –35/-1 Fshb Tukey’s HSD post hoc test. oligos in lanes 15–18. doi: 10.1210/me.2013-1185 mend.endojournals.org 9

FOXO1 Suppression of Basal Fshb Transcription ments in CV-1 cells lacking PITX1 because L␤T2 cells May Involve PITX1 contain PITX1 (39, 40). We first tested whether a multi- Since the –64 Fshb-luc containing the PITX1 binding mer of a bicoid-related homeodomain binding element site was sufficient to elicit FOXO1 suppression, we tested (HDBE) was sufficient to elicit the FOXO1 suppressive whether Fshb transcription due to PITX1 overexpression effect. As shown in Figure 7A, PITX1 induction of the was reduced by FOXO1. We performed these experi- 4xHDBE was significantly decreased by FOXO1. We then showed that overexpression of PITX1 resulted in a 14-fold induc- A B tion of Fshb gene expression which was reduced 73% by FOXO1 (Fig- ure 7C). In addition, we tested whether the PITX1 binding element in the proximal Fshb promoter shown in Figure 7B was necessary for the PITX1 induction. Our results demonstrated that the PITX1 bind- ing element was absolutely required C D β for PITX1 induction of Fshb gene expression in CV-1 cells (Figure 7C). We also tested whether the PITX1 binding element was required in L␤T2 cells. As previously reported

β (11), mutation of the PITX1 binding element reduced basal Fshb tran- scription (Figure 7D) but had no ef- fect on GnRH induction (data not shown). Interestingly, overexpres- sion of FOXO1 reduced transcrip- E tion of Fshb from the mutated pro- moter by 49% compared to 64% for the wild-type promoter (Figure 7D), suggesting that the PITX1 binding element is not necessary for basal FOXO1 suppression of Fshb.Inad- ∆ dition, mutation of the PITX1 bind- ing element did not prevent FOXO1 suppression of GnRH-induced Fshb transcription (data not shown). FIGURE 7. FOXO1 Suppression of Basal Fshb Transcription May Involve PITX1. A. The 4xHDBE- luc reporter was transiently transfected into CV-1 cells along with pcDNA3 empty vector (EV), FOXO1 Interacts with PITX1 FOXO1 and PITX1, as indicated. The results represent the mean Ϯ SEM of three experiments performed in triplicate and are presented as fold PITX1 induction relative to pcDNA3 empty Since FOXO1 can suppress vector. * indicates that transcription is significantly repressed by FOXO1 compared to EV using PITX1 induction of Fshb and the Student’s t test. B. Diagram illustrating the mutations in the PITX1 binding element (PITXBE) in 4xHDBE in CV-1 cells, we investi- the murine Fshb promoter. C-D. The –1000 Fshb-luc reporter and the Fshb PITX1 mutant (mut) gated whether FOXO1 can physi- were transiently transfected into CV-1 (C) or L␤T2 (D) cells along with EV, FOXO1 and PITX1, as indicated. The results represent the mean Ϯ SEM of three experiments performed in triplicate cally interact with PITX1. We tested and are presented as fold PITX1 induction relative to pcDNA empty vector (C) or luc/␤-gal (D). * whether FOXO1 or DNA-binding indicates that Fshb-luc transcription is significantly repressed by FOXO1 compared to EV using deficient FOXO1 interacts with Student’s t test while # indicates that the PITX1 mut was significantly repressed compared to the wild-type Fshb promoter. E. GST interaction assays were performed using bacterially expressed PITX1 by incubating a GST-PITX1 GST-fusion proteins (indicated above each lane) and 35S-labeled in vitro transcribed and fusion protein with in vitro-tran- translated FOXO1, FOXO1-⌬DBD and GFP (indicated on the left of the panels). GFP was used as scribed and translated 35S-labeled a negative control. The GST-fusion proteins included GST alone and GST-PITX1. One quarter of ⌬ the protein used in the interaction assay was loaded in the lane marked input. The experiment FOXO1 or FOXO1- DBD in pull- was repeated several times with the same results and a representative experiment is shown. down experiments. As shown in Fig- 10 FOXO1 Suppresses Fshb Gene Expression Mol Endocrinol ure 7E, there was minimal interaction between the GST- 35 PITX1 fusion protein and the negative control ( S-GFP) A or with GST alone incubated with FOXO1 or FOXO1- ⌬DBD. In addition, FOXO1 bound to PITX1 while there was no detectable interaction between FOXO1-⌬DBD and PITX1, indicating that the interaction between FOXO1 and PITX1 requires the FOXO1 DBD.

FOXO1 Reduces Pitx1 mRNA and Protein Levels Since the FOXO1 DBD is required for suppression of Fshb but FOXO1 does not appear to bind to the proximal Fshb promoter, we hypothesized that FOXO1 may mod- ulate Fshb transcription indirectly by regulating the amount of PITX1 in the cells. In these experiments, L␤T2 cells were transduced with Ad-GFP or Ad-FOXO1-CA B and Pitx1 mRNA levels were measured using quantitative RT-PCR (Figure 8A). FOXO1-CA significantly decreased Pitx1 mRNA levels by 57%. We also tested whether the decrease in Pitx1 mRNA levels translated into lower PITX1 protein levels. As shown in Figure 8B-C, FOXO1-CA transduction of L␤T2 cells resulted in a 40% reduction in PITX1 protein levels. These results suggest that FOXO1 may suppress Fshb transcription indirectly through regulation of Pitx1 mRNA and protein levels.

FOXO1 Does Not Regulate c-Fos or c-Jun mRNA β and Protein Levels Since GnRH induction of the murine Fshb promoter C involves the intermediate early , c-FOS and c-JUN that comprise the transcriptional AP1 complex, we deter- mined whether FOXO1-CA overexpression alters pro- duction of these two genes. L␤T2 cells were transduced with Ad-GFP or Ad-FOXO1-CA and c-fos and c-jun

mRNA levels were measured using quantitative RT-PCR β (Figure 9A and C). Similar to previous reports (47, 48), 1 hour of GnRH treatment substantially increased c-fos and c-jun mRNA levels (Figure 9B and D). However, FOXO1-CA overexpression had little or no effect on c-fos and c-jun mRNA. We also demonstrated that there was no effect of FOXO1-CA overexpression on c-FOS and FIGURE 8. FOXO1 Suppression of Basal Fshb Transcription May c-JUN protein levels (Figure 9E). Involve FOXO1 Regulation of Pitx1 mRNA and Protein Levels. L␤T2 cells were transduced with Ad-GFP or Ad-FOXO1-CA for 6 hours, then FOXO1 Suppression of GnRH-Induced Fshb switched to serum-free media. 24 hours after adenoviral infection, cells Transcription Involves Two Regions of the were harvested for total mRNA (A) or protein (B-C). A. The results represent the mean Ϯ SEM of seven experiments performed in Proximal Fshb Promoter at –50/-41 and –30/-21 triplicate and are presented as amount of Pitx1 mRNA relative to Since GnRH induction involves an AP1 binding ele- Gapdh. * indicates that Pitx1 transcription is significantly repressed by ment in the proximal Fshb promoter, we tested whether FOXO1-CA compared to GFP using Student’s t test. B-C. Western blot analysis was performed on whole cell extracts using PITX1, FOXO1 and this site was necessary for FOXO1 suppression of Fshb ␤-Tubulin primary antibodies and a horseradish peroxidase-linked transcription. Mutation of the AP1 binding element secondary antibody. A representative image is shown (B) or graph Ϯ shown in Figure 10A reduced GnRH induced Fshb tran- representing the mean SEM of four experiments presented as the amount of PITX1 protein relative to ␤-Tubulin (C). * indicates that scription (Figure 10B), as previously reported (16). How- PITX1 protein levels are significantly decreased by FOXO1-CA ever, overexpression of FOXO1 reduced transcription of compared to GFP using Student’s t test. doi: 10.1210/me.2013-1185 mend.endojournals.org 11

Fshb from the mutated promoter by 50% compared to promoter ranging from –70/-61 to –20/-11 within the 59% for the wild-type promoter (Figure 10B), indicating –398 Fshb-luc were compared to wild-type –398 Fshb- that the AP1 binding element is not required for FOXO1 luc. Basal transcription of Fshb from either wild-type or suppression. Since FOXO1 suppression of GnRH-in- mutated promoters was suppressed by FOXO1 (data not duced Fshb transcription occurred within the –64 Fshb shown), suggesting that none of the 10 bp regions indi- promoter lacking the AP1 binding element, we sought to vidually were required for FOXO1 suppression of basal map which regions of the proximal promoter were re- Fshb synthesis. Interestingly, GnRH induction of Fshb quired for the suppression. Ten bp deletions of the Fshb was reduced with the mutated promoters compared to the wild-type promoter (Figure 10C), ABindicating that GnRH induction in- volves other regions of the Fshb pro- moter in addition to the AP1 binding element. Moreover, FOXO1 was unable to repress GnRH induction of Fshb from promoters containing deletions at –50/-41 or –30/-21 (Fig- ure 10C), indicating that these re- gions may play a role in FOXO1 suppression of Fshb gene expression induced by GnRH. C D

Discussion

Our studies suggest that the FOXO1 transcription factor may regulate fertility through modulation of go- nadotropin beta subunit gene ex- pression in pituitary gonadotrope cells. Overexpression of wild-type or constitutively active FOXO1 greatly diminished basal and GnRH-induced Fshb mRNA levels E and Fshb-luc expression in L␤T2 cells (Figures 1 and 2), similarly to the repressive effect of FOXO1 on Lhb gene expression (35). Although wild-type FOXO1 is predominantly localized in the cytoplasm under se- rum-free media conditions, the level of nuclear FOXO1 appears to be sufficient to suppress Fshb tran- scription (Figure 2) (35). Our results FIGURE 9. FOXO1 Suppression of GnRH-Induced Fshb Transcription Does Not Involve demonstrating that inhibition of Regulation of c-Fos or c-Jun mRNA or Protein Levels. L␤T2 cells were transduced with Ad-GFP or Ad-FOXO1-CA for 6 hours, then switched to serum-free media. 24 hours after adenoviral PI3K results in increased nuclear lo- infection, cells were treated with 0.1% BSA vehicle (Veh) or 10 nM GnRH for 1 hour, as calization of FOXO1 (35) and de- indicated. Cells were harvested for total mRNA (A-D) or protein (E). A-D. The results represent creased basal and GnRH induction Ϯ the mean SEM of three experiments performed in triplicate and are presented as amount of of Fshb mRNA in primary pituitary c-Fos mRNA relative to Gapdh (A), fold GnRH induction of c-Fos relative to Veh (B), c-Jun mRNA relative to Gapdh (C), fold GnRH induction of c-Jun relative to Veh (D). * indicates that cells (Figure 1) provides support for transcription is significantly repressed by FOXO1-CA compared to GFP using Student’s t test. E. the idea that FOXO1 is a repressor Western blot analysis was performed on whole cell extracts using c-FOS, c-JUN and FOXO1 of Fshb transcription in a more primary antibodies and a horseradish peroxidase-linked secondary antibody. A representative image is shown. physiological context. The suppres- 12 FOXO1 Suppresses Fshb Gene Expression Mol Endocrinol sion of Fshb and Lhb transcription by FOXO1 is not due occur in humans since FOXO1 suppressed transcription to a global repressive mechanism since reporter plasmids of a reporter plasmid containing either the human FSHB containing the insulin response element from the IGFBP1 or LHB promoter in L␤T2 cells (Figure 3) (35). gene (35) or a consensus FBE (Figure 2) are induced by So, how does FOXO1 regulate gonadotropin gene ex- FOXO1 in L␤T2 cells while endogenous c-fos and c-jun pression via transcriptional repression? Although there mRNA levels are not altered by adenoviral transduction are several similarities between FOXO1 repressive effects of L␤T2 cells with FOXO1-CA (Figure 9). In addition, on the Fshb and Lhb promoters, there are promoter-spe- FOXO1 repression of transcription is not an artifact of cific mechanisms that are also required for FOXO1 sup- the reporter plasmids since substantial repression of en- pression of the individual promoters. First, the DBD of dogenous Fshb and Lhb mRNA was observed in L␤T2 FOXO1 is required for the repression of both genes (Fig- cells transduced with constitutively active FOXO1 (Fig- ure 5) (35), although the specific residues involved have ure 1) (35). not yet been mapped and it is unknown whether residues At this time, it remains to be determined whether required for the repression differ from those necessary for FOXO1 regulates transcription of the gonadotropin beta binding to DNA. Second, there is no evidence that subunits in human gonadotropes. There is one report that FOXO1 binds directly to either the Fshb or Lhb proximal FOXO1 colocalized with Lhb-staining gonadotropes in a promoters (Figure 6) (35). It is possible that FOXO1 human male pituitary section (49). Our results also sug- binds to these promoters with such low affinity that it gest that FOXO1 regulation of gonadotropin levels may cannot be detected in a gel-shift assay, although FOXO1 binding to a consensus FBE is not A B altered by 100-fold excess of cold oligonucleotides from the proximal Fshb or Lhb promoters. Thus, our data support the idea that FOXO1 elicits a repressive effect on the go- nadotropin beta subunit promoters through indirect mechanisms in- volving the FOXO1 DBD. This idea is not altogether unexpected since C FOXO1 has been shown to interact with other proteins such as SMADS through its DBD (50). Given that FOXO1 has been re- ported to physically interact with other proteins to regulate transcrip- tion of specific target genes (51), it is plausible that FOXO1 represses basal transcription and GnRH in- duction of the Fshb and Lhb pro- ∆ ∆ ∆ ∆ ∆ ∆ moters through specific protein-pro- tein interactions with transcription FIGURE 10. FOXO1 Suppression of GnRH-Induced Fshb Transcription Involves Two Elements in the Proximal Fshb Promoter. A. Diagram illustrating the mutations in the AP1 binding element factor(s) or cofactors necessary for (AP1BE) in the murine Fshb promoter. B. The –398 Fshb-luc reporter and the Fshb AP1 mutation gonadotropin synthesis. Although (mut) were transiently transfected into L␤T2 cells along with pcDNA3 empty vector (EV) or basal transcription of Fshb involves FOXO1, as indicated. After overnight incubation in serum-free media, cells were treated for 6 hours with 0.1% BSA vehicle or 10 nM GnRH. The results represent the mean Ϯ SEM of three several factors, including LHX3 and experiments performed in triplicate and are presented as fold GnRH induction relative to the NFY, FOXO1 suppression of basal vehicle control. * indicates that Fshb-luc transcription is significantly repressed by FOXO1 Fshb gene expression mapped to the compared to EV using Student’s t test while # indicates that the AP1 mut was significantly –64/-1 region of the Fshb promoter repressed compared to the wild-type Fshb promoter. C. The –398 Fshb-luc reporter and 10 bp deletions ranging from –70/-61 to –20/-11 were transiently transfected into L␤T2 cells along with (Figure 4), suggesting that FOXO1 EV or FOXO1, as indicated. After overnight incubation in serum-free media, cells were treated for interaction with factors binding to Ϯ 6 hours with 0.1% BSA or 10 nM GnRH. The results represent the mean SEM of three this region eg, PITX1 may play a experiments performed in triplicate and are presented as fold GnRH induction relative to the vehicle control. * indicates that Fshb-luc transcription is significantly repressed by FOXO1 role in the repression of transcrip- compared to EV using one-way ANOVA followed by Tukey’s HSD post hoc test. tion. In comparison, FOXO1 sup- doi: 10.1210/me.2013-1185 mend.endojournals.org 13 pression of basal Lhb transcription mapped to the Interestingly, little is known about the factors that regu- –150/-87 region of the Lhb promoter, which contains late PITX1/2 expression during pituitary development PITX1 and SF1 binding sites (35). Since the PITX1 bind- and in the adult (59). Since overexpression of FOXO1 in ing site in the proximal FSHB promoter is conserved L␤T2 cells resulted in a significant reduction of Pitx1 among mammals and mutation of this site in the murine mRNA and PITX1 protein levels (Figure 8), it is possible and human promoters reduced FSHB transcription (11), that FOXO1 may repress Pitx1 synthesis through regula- it is interesting to speculate that FOXO1 suppression of tion of the Pitx1 promoter. basal FSHB transcription via PITX1 may also be GnRH responsiveness of the Fshb and Lhb promoters conserved. occurs due to the induction and binding of gene-specific The hypothesis that PITX1 plays a role in FOXO1 transcription factors such as AP1 to the Fshb promoter suppression of the Fshb promoter is supported by our and early growth response protein 1 (EGR1) to the Lhb results demonstrating that PITX1 induction of the Fshb promoter. Our results suggest that FOXO1 suppression promoter in CV-1 cells is repressed by FOXO1, that of GnRH-induced Fshb gene expression does not occur PITX1 induction of a consensus HDBE is also decreased through regulation of AP1. Transduction of L␤T2 cells by FOXO1 and that PITX1 physically interacts with with FOXO1-CA did not alter c-Fos and c-Jun mRNA or FOXO1 (Figure 7). The fact that the FOXO1 DBD is protein levels (Figure 9). In addition, FOXO1 suppression required for the interaction between FOXO1 and PITX1 of the GnRH induction of Fshb still occurred on the –64 is also in agreement with our data showing that the Fshb promoter lacking the characterized AP1 binding el- FOXO1 DBD is necessary for FOXO1 repression of Fshb. ement (Figure 4). We also demonstrated that the AP1 Although this is the first report of a physical interaction binding element was not necessary for FOXO1 to elicit a between PITX1 and FOXO1, PITX2 interacts with an- repressive effect on Fshb transcription (Figure 10). These other member of the forkhead transcription factor family, results are not that surprising, given that GnRH respon- FOXC1 via the C-terminal activation domain of FOXC1 siveness on the murine Fshb promoter appears to involve and the homeodomain of PITX2 (52). Interestingly, the additional regions outside the AP1 binding element (16). PITX1 binding element in the proximal Fshb promoter This is also illustrated by the fact that the –64 Fshb pro- was not required for FOXO1 suppression in L␤T2 cells, moter was still responsive to GnRH, that mutation of the suggesting that the repression can occur without PITX1 AP1 element reduced, but did not abolish, the GnRH binding to DNA. It is noteworthy that PITX1 was re- induction, and that 10 bp deletions of the proximal Fshb ported to activate the rat or human Fshb promoter even promoter from –70/-61 to –30/-21 all resulted in reduc- when the PITX1 binding element was mutated, suggest- tion of GnRH responsiveness. Although other factors in- ing that PITX1 activation can occur through protein-pro- volved in the GnRH induction of murine Fshb have not tein interactions with factors such as SF1, in addition to yet been identified, our studies highlight two regions be- direct DNA binding (11, 53–55). Altogether, our studies tween –50/-41 and –30/-21 which are required for imply that FOXO1 interaction with PITX1 may be im- FOXO1 suppression (Figure 10) and suggest that factors portant for suppression of basal Fshb and possibly Lhb other than AP1 are targeted by FOXO1. It should be transcription. noted that the –50/-41 region partially overlaps the In addition to the protein-protein interaction between PITX1 binding element while the –30/-21 region overlaps FOXO1 and PITX1 which may regulate Fshb transcrip- the TATA box. tion, it is possible that FOXO1 modulates Fshb mRNA In contrast to the Fshb promoter, FOXO1 suppression levels through regulation of Pitx1 mRNA and protein of GnRH responsiveness on the Lhb promoter mapped to levels. Both PITX1/2 are expressed in Rathke’s pouch at the –87/-1 region which contains an EGR1 binding ele- embryonic day 10.5 and in the developing anterior lobe at ment (35). Our previous studies also demonstrated that e11.5 (56, 57). PITX1/2 are also highly expressed in adult FOXO1 suppressed Lhb transcription induced by EGR1 gonadotrope cells and have been reported to regulate overexpression in L␤T2 cells. Overexpression of FOXO1 transcription of the alpha subunit, Fshb, Lhb and the with PITX1 or SF1 also suppressed EGR1 induction of GnRH (39, 57). Although knockout of PITX1 Lhb in CV-1 cells. These results provide support for the had little effect on development, lack of idea that an interaction between FOXO1 and EGR1 may PITX1/2 resulted in a more severe defect in pituitary be responsible for FOXO1 suppression of GnRH-induced growth and differentiation than the PITX2 knockout Lhb transcription while FOXO1 interactions with as yet alone (58). These studies indicate that PITX1/2 may have unknown factors are responsible for FOXO1 suppression distinct and overlapping functions in pituitary develop- of GnRH-induced Fshb gene expression. ment and gonadotropin hormone production in the adult. In summary, we demonstrate that FOXO1 suppresses 14 FOXO1 Suppresses Fshb Gene Expression Mol Endocrinol basal transcription and GnRH induction of the Fshb gene, gram in Reproduction and Infertility Research and by a UCSD potentially through protein-protein interactions between Academic Senate Health Sciences Research Grant. FOXO1 and transcription factors recruited to the proxi- mal Fshb promoter such as PITX1 as well as FOXO1 regulation of Pitx1 mRNA and protein levels. If FOXO1 References is regulated by metabolic hormone signaling in gonado- 1. Apter D. Development of the hypothalamic-pituitary-ovarian axis. tropes and represses Fshb and Lhb transcription, as our Ann N Y Acad Sci. 1997;816:9–21. studies suggest, then it may prove to be an important 2. Plant TM, Marshall GR. The functional significance of FSH in spermatogenesis and the control of its secretion in male primates. factor in the regulation of gonadotropin production in Endocr Rev. 2001;22(6):764–786. situations of caloric insufficiency or excess. Further stud- 3. Kumar TR, Wang Y, Lu N, Matzuk MM. Follicle stimulating hor- ies are necessary to comprehend how interactions among mone is required for ovarian follicle maturation but not male fer- FOXO1, PITX1, EGR1, and other factors contribute to tility. Nat Genet. 1997;15(2):201–204. 4. Lamminen T, Jokinen P, Jiang M, Pakarinen P, Simonsen H, Huhta- the regulation of Fshb and Lhb transcription in pituitary niemi I. Human FSH beta subunit gene is highly conserved. Mol gonadotropes. Investigation of the regulation of PITX1 Hum Reprod. 2005;11(8):601–605. synthesis may also provide insight into the mechanisms of 5. Ortolano GA, Haisenleder DJ, Dalkin AC, Iliff-Sizemore SA, Lan- defeld TD, Maurer RA, Marshall JC. Follicle-stimulating hormone FOXO1 repression. Additional studies will also help elu- beta subunit messenger ribonucleic acid concentrations during the cidate how FOXO1 integrates input from multiple hor- rat estrous cycle. Endocrinology. 1988;123(6):2946–2948. monal signaling pathways to regulate reproduction under 6. Halvorson LM, Weiss J, Bauer-Dantoin AC, Jameson JL. Dynamic favorable or adverse environmental conditions. regulation of pituitary follistatin messenger ribonucleic acids during the rat estrous cycle. Endocrinology. 1994;134(3):1247–1253. 7. Chand AL, Harrison CA, Shelling AN. Inhibin and premature ovar- ian failure. Hum Reprod Update. 2010;16(1):39–50. Acknowledgments 8. Bernard DJ, Fortin J, Wang Y, Lamba P. Mechanisms of FSH syn- thesis: what we know, what we don’t, and why you should care. Fertil Steril. 2010;93(8):2465–2485. The authors thank Djurdjica Coss, Pamela Mellon, Kellie Breen, 9. West BE, Parker GE, Savage JJ, Kiratipranon P, Toomey KS, Beach and Scott Kelley for helpful discussions and critical reading of LR, Colvin SC, Sloop KW, Rhodes SJ. Regulation of the follicle- the manuscript. stimulating hormone beta gene by the LHX3 LIM-homeodomain transcription factor. Endocrinology. 2004;145(11):4866–4879. Received June 19, 2013. Accepted September 18, 2013. 10. Jacobs SB, Coss D, McGillivray SM, Mellon PL. Nuclear factor Y Address all correspondence and requests for reprints to: and steroidogenic factor 1 physically and functionally interact to Varykina G. Thackray, Ph.D., Department of Reproductive contribute to cell-specific expression of the mouse Follicle-stimu- Medicine, University of California, San Diego, 9500 Gilman lating hormone-beta gene. Mol Endocrinol. 2003;17(8):1470– 1483. Drive, MC 0674, LA Jolla, CA 92093, USA, Tel.: (858) 822– 11. Lamba P, Khivansara V, D’Alessio AC, Santos MM, Bernard DJ. 7693; Email:[email protected] Paired-like homeodomain transcription factors 1 and 2 regulate The authors have nothing to disclose. follicle-stimulating hormone beta-subunit transcription through a The Endocrine Society NIH statement: “This is an uncopy- conserved cis-element. Endocrinology. 2008;149(6):3095–3108. edited author manuscript copyrighted by The Endocrine Soci- 12. Thackray VG, Mellon PL, Coss D. Hormones in synergy: Regula- ety. This may not be duplicated or reproduced, other than for tion of the pituitary gonadotropin genes. Molecular and Cellular personal use or within the rule of “Fair Use of Copyrighted Endocrinology. 2010;314(2):192–203. Materials” (section 107, Title 17, U.S. Code) without permis- 13. Dalkin AC, Burger LL, Aylor KW, Haisenleder DJ, Workman LJ, sion of the copyright owner, The Endocrine Society. From the Cho S, Marshall JC. Regulation of gonadotropin subunit gene tran- time of acceptance following peer review, the full text of this scription by gonadotropin- releasing hormone: measurement of pri- mary transcript ribonucleic acids by quantitative reverse transcrip- manuscript is made freely available by The Endocrine Society at tion-polymerase chain reaction assays. Endocrinology. 2001; http://www.endojournals.org/. The final copy edited article can 142(1):139–146. be found at http://www.endojournals.org/. The Endocrine Soci- 14. Vasilyev VV, Pernasetti F, Rosenberg SB, Barsoum MJ, Austin DA, ety disclaims any responsibility or liability for errors or omis- Webster NJ, Mellon PL. Transcriptional activation of the ovine sions in this version of the manuscript or in any version derived follicle-stimulating hormone-beta gene by gonadotropin-releasing from it by the National Institutes of Health or other parties. The hormone involves multiple signal transduction pathways. Endocri- citation of this article must include the following information: nology. 2002;143(5):1651–1659. author(s), article title, journal title, year of publication and 15. Liu F, Austin DA, Mellon PL, Olefsky JM, Webster NJ. GnRH DOI.” activates ERK1/2 leading to the induction of c-fos and LHbeta This work was supported by This work was supported by protein expression in LbetaT2 cells. Mol Endocrinol. 2002;16(3): 419–434. NIH grants K01 DK080467 and R01 HD067448 to V.G.T. as 16. Coss D, Jacobs SB, Bender CE, Mellon PL. A novel AP-1 site is well as T32 HD007203 and F32 HD074414 to D.V.S. This critical for maximal induction of the follicle-stimulating hormone work was also supported by a pilot and feasibility grant from the beta gene by gonadotropin-releasing hormone. The Journal of Bi- UCSD/UCLA Diabetes Research Center (P30 DK063491), by ological Chemistry. 2004;279:152–162. NICHD/NIH through a cooperative agreement (U54 17. Liu F, Ruiz MS, Austin DA, Webster NJ. Constitutively active Gq HD012303) as part of the Specialized Cooperative Centers Pro- impairs gonadotropin-releasing hormone-induced intracellular sig- doi: 10.1210/me.2013-1185 mend.endojournals.org 15

naling and secretion in LbetaT2 cells. Mol follicle stimulating hormone (FSH) in response to activin A. J En- Endocrinol. 2005;19(8):2074–2085. docrinol. 1999;162:R1–R5. 18. Wang Y, Fortin J, Lamba P, Bonomi M, Persani L, Roberson MS, 38. Pernasetti F, Vasilyev VV, Rosenberg SB, Bailey JS, Huang H-J, Bernard DJ. Activator protein-1 and proteins synergistically Miller WL, Mellon PL. Cell-specific transcriptional regulation of regulate human follicle-stimulating hormone beta-promoter activ- FSH␤ by activin and GnRH in the L␤T2 pituitary gonadotrope cell ity. Endocrinology. 2008;149(11):5577–5591. model. Endocrinology. 2001;142(6):2284–2295. 19. Ciccone NA, Lacza CT, Hou MY, Gregory SJ, Kam KY, Xu S, 39. Tremblay JJ, Lanctôt C, Drouin J. The pan-pituitary activator of Kaiser UB. A composite element that binds basic helix loop helix transcription, Ptx1 (pituitary homeobox 1), acts in synergy with and basic zipper transcription factors is important for go- SF-1 and Pit1 and is an upstream regulator of the Lim-homeodo- nadotropin-releasing hormone regulation of the follicle-stimulating main gene Lim3/Lhx3. Mol Endocrinol. 1998;12(3):428–441. hormone beta gene. Mol Endocrinol. 2008;22(8):1908–1923. 40. Jiang Q, Jeong KH, Horton CD, Halvorson LM. Pituitary homeo- 20. Weiss J, Guendner MJ, Halvorson LM, Jameson JL. Transcrip- box 1 (Pitx1) stimulates rat LHbeta gene expression via two func- tional activation of the follicle-stimulating hormone beta-subunit tional DNA-regulatory regions. Journal of Molecular Endocrinol- gene by activin. Endocrinology. 1995;136(5):1885–1891. ogy. 2005;35(1):145–158. 21. Coss D, Mellon PL, Thackray VG. A FoxL in the Smad house: 41. Rosenberg SB, Mellon PL. An Otx-related homeodomain protein activin regulation of FSH. Trends in Endocrinology and Metabo- binds an LHb promoter element important for activation during lism: TEM. 2010;21:562–568. gonadotrope maturation. Mol Endocrinol. 2002;16:1280–1298. 22. Adashi EY, Hsueh AJ, Yen SS. Insulin enhancement of luteinizing 42. Thackray VG, McGillivray SM, Mellon PL. Androgens, progestins ␤ hormone and follicle-stimulating hormone release by cultured pi- and glucocorticoids induce follicle-stimulating hormone -subunit tuitary cells. Endocrinology. 1981;108(4):1441–1449. gene expression at the level of the gonadotrope. Mol Endocrinol. 23. Kanematsu T, Irahara M, Miyake T, Shitsukawa K, Aono T. Effect 2006;20(9):2062–2079. of insulin-like growth factor I on gonadotropin release from the 43. Zhang X, Gan L, Pan H, Guo S, He X, Olson ST, Mesecar A, Adam -pituitary axis in vitro. Acta Endocrinol. 1991; S, Unterman TG. Phosphorylation of serine 256 suppresses trans- 125(2):227–233. activation by FKHR (FOXO1) by multiple mechanisms. Direct and 24. Ogura K, Irahara M, Kiyokawa M, Tezuka M, Matsuzaki T, Yasui indirect effects on nuclear/cytoplasmic shuttling and DNA binding. T, Kamada M, Aono T. Effects of leptin on secretion of LH and The Journal of Biological Chemistry. 2002;277(47):45276– 45284. FSH from primary cultured female rat pituitary cells. Eur J Endo- 44. Zappavigna V, Sartori D, Mavilio F. Specificity of HOX protein crinol. 2001;144(6):653–658. function depends on DNA-protein and protein- protein interac- 25. Tezuka M, Irahara M, Ogura K, Kiyokawa M, Tamura T, Matsu- tions, both mediated by the homeo domain. Genes Dev. 1994;8(6): zaki T, Yasui T, Aono T. Effects of leptin on gonadotropin secre- 732–744. tion in juvenile female rat pituitary cells. Eur J Endocrinol. 2002; 45. Coss D, Hand CM, Yaphockun KK, Ely HA, Mellon PL. p38 mi- 146(2):261–266. togen-activated protein kinase is critical for synergistic induction of 26. Van Der Heide LP, Hoekman MF, Smidt MP. The ins and outs of the FSH(beta) gene by gonadotropin-releasing hormone and activin FoxO shuttling: mechanisms of FoxO translocation and transcrip- through augmentation of c-Fos induction and Smad phosphoryla- tional regulation. Biochem J. . 2004;380(Pt 2):297–309. tion. Mol Endocrinol. 2007;21(12):3071–3086. 27. Jacobs FM, van der Heide LP, Wijchers PJ, Burbach JP, Hoekman 46. Hemmings BA, Restuccia DF. PI3K-PKB/Akt pathway. Cold MF, Smidt MP. FoxO6, a novel member of the FoxO class of Spring Harbor Perspectives in Biology. 2012;4(9):a011189. transcription factors with distinct shuttling dynamics. The Journal 47. Ely HA, Mellon PL, Coss D. GnRH Induces the c-Fos gene via of Biological Chemistry. 2003;278(38):35959–35967. phosphorylation of SRF by the calcium/calmodulin kinase II path- 28. Anderson MJ, Viars CS, Czekay S, Cavenee WK, Arden KC. Clon- way. Mol Endocrinol. 2011;25(4):669–680. ing and characterization of three human forkhead genes that com- 48. Lindaman LL, Yeh DM, Xie C, Breen KM, Coss D. Phosphoryla- prise an FKHR-like gene subfamily. Genomics. 1998;47(2):187– tion of ATF2 and interaction with NFY induces c-Jun in the go- 199. nadotrope. Mol Cell Endocrinol. 2013;365(2):316–326. 29. Kenyon CJ. The genetics of ageing. Nature. 2010;464(7288):504– 49. Stavrou E. Regulation of FOXO Transcription Factors by Gonad- 512. otropin-Releasing Hormone. University of Edinburgh; 2011. 30. Salih DA, Brunet A. FoxO transcription factors in the maintenance 50. Seoane J, Le HV, Shen L, Anderson SA, Massague J. Integration of of cellular homeostasis during aging. Curr Opin Cell Biol. 2008; Smad and forkhead pathways in the control of neuroepithelial and 20(2):126–136. glioblastoma cell proliferation. Cell. 2004;117(2):211–223. 31. Varela L, Horvath TL. Leptin and insulin pathways in POMC and 51. Monsalve M, Olmos Y. The complex biology of FOXO. Current AgRP neurons that modulate energy balance and glucose homeo- Drug Targets. 2011;12(9):1322–1350. stasis. EMBO Rep. 2012;13(12):1079–1086. 52. Berry FB, Lines MA, Oas JM, Footz T, Underhill DA, Gage PJ, 32. Tang ED, Nunez G, Barr FG, Guan KL. Negative regulation of the Walter MA. Functional interactions between FOXC1 and PITX2 forkhead transcription factor FKHR by Akt. The Journal of Bio- underlie the sensitivity to FOXC1 gene dose in Axenfeld-Rieger logical Chemistry. 1999;274(24):16741–16746. syndrome and anterior segment dysgenesis. Human Molecular Ge- 33. Richards JS, Pangas SA. The ovary: basic biology and clinical im- netics. 2006;15(6):905–919. plications. J Clin Invest. 2010;120(4):963–972. 53. Suszko MI, Antenos M, Balkin DM, Woodruff TK. Smad3 and 34. Brosens JJ, Wilson MS, Lam EW. FOXO transcription factors: Pitx2 cooperate in stimulation of FSHbeta gene transcription. Mo- from cell fate decisions to regulation of human female reproduc- lecular and Cellular Endocrinology. 2008;281(1–2):27–36. tion. Adv Exp Med Biol. 2009;665:227–241. 54. Tremblay JJ, Marcil A, Gauthier Y, Drouin J. Ptx1 regulates SF-1 35. Arriola DJ, Mayo SL, Skarra DV, Benson CA, Thackray VG. activity by an interaction that mimics the role of the ligand-binding FOXO1 transcription factor inhibits luteinizing hormone beta gene domain. EMBO J. 1999;18(12):3431–3441. expression in pituitary gonadotrope cells. The Journal of Biological 55. Zakaria MM, Jeong KH, Lacza C, Kaiser UB. Pituitary homeobox Chemistry. 2012;287(40):33424–33435. 1 activates the rat FSHbeta (rFSHbeta) gene through both direct 36. Calnan DR, Brunet A. The FoxO code. Oncogene. 2008;27(16): and indirect interactions with the rFSHbeta gene promoter. Mol 2276–2288. Endocrinol. 2002;16(8):1840–1852. 37. Graham KE, Nusser KD, Low MJ. LbT2 gonadotroph cells secrete 56. Lanctôt C, Gauthier Y, Drouin J. Pituitary homeobox 1 (Ptx1) is 16 FOXO1 Suppresses Fshb Gene Expression Mol Endocrinol

differentially expressed during pituitary development. Endocrinol- stages of pituitary organogenesis: pituitary primordium formation ogy. 1999;140(3):1416–1422. and cell specification. Development. 2002;129(2):329–337. 57. Charles MA, Suh H, Hjalt TA, Drouin J, Camper SA, Gage PJ. 59. Goodyer CG, Tremblay JJ, Paradis FW, Marcil A, Lanctot C, Gau- PITX genes are required for cell survival and Lhx3 activation. Mol thier Y, Drouin J. Pitx1 in vivo promoter activity and mechanisms Endocrinol. 2005;19(7):1893–1903. of positive autoregulation. Neuroendocrinology. 2003;78(3):129– 58. Suh H, Gage PJ, Drouin J, Camper SA. Pitx2 is required at multiple 137.