Premature Expression of FOXO1 in Developing Mouse Pituitary Results in Anterior Lobe Hypoplasia

Abstract

The process by which the somatotrope lineage emerges in the developing pituitary is regulated by the activity of specific signaling and transcription factors expressed during development. We set out to understand the contribution of FOXO1 to that process by using a mouse model in which FOXO1 is prematurely expressed in the pituitary primordium. Expression of FOXO1 in the oral ectoderm as early as embryonic day (e)9.5 resulted in pituitary gland hypoplasia and reduced expression of anterior lobe hormone transcripts at e18.5. Of note, the relative numbers of somatotropes and thyrotropes were also decreased at e18.5. LHX3 and PITX2, markers of pituitary identity, were present in a reduced number of cells during the formation of the Rathke pouch. Thus, premature expression of FOXO1 may affect adoption of pituitary identity during differentiation. Our results demonstrate that the timing of FOXO1 activation affects its role in pituitary gland organogenesis and somatotrope differentiation.

A mouse model with premature expression of FOXO1 in the developing pituitary gland demonstrates a reduction in the number of cells with pituitary identity, resulting in hypoplasia.

The endocrine pituitary gland forms during development as the result of highly specific spatial and temporal cues originating from both the neural and oral ectoderm. Many of these cues come from the ventral diencephalon, also referred to as the pituitary organizer (1). In mice, the oral ectoderm begins to invaginate by embryonic day (e) 10.5, and the invaginated region begins to separate from the oral ectoderm by apoptosis at e11.5 (2). At e12.5, the Rathke pouch is formed and exhibits a region of continued proliferation around the lumen, which is responsible for the lobe’s expansion. Aberrant expression of developmentally important signaling and transcription factors can change ultimate hormone production within the gland on the basis of their control of terminal differentiation of the different cell types and gland morphology (3). Therefore, it is important to understand the intricate expression patterns of key molecules in pituitary gland development. Foxo1 is one such molecule and is required after pituitary progenitor commitment but before terminal differentiation of somatotropes (4). In complement to our previous knockout mouse model investigation, this study aimed to ascertain the specific role of Foxo1 in pituitary gland development using a constitutively active approach.

Foxo1, which encodes a forkhead box transcription factor, was shown to be present in the embryonic pituitary during a transcriptome analysis of both wild-type (WT) and Prop1^df/df animals (5, 6). Forkhead box transcription factors are basic helix-turn-helix proteins with highly conserved DNA-binding or “forkhead” domains 100 amino acids in length (7). FOXO1 has a multiplicity of roles, including regulation of cellular differentiation, redox homeostasis, cell cycle control, and insulin signaling (7). FOXO1 exerts most of its biological control in the nucleus, where it acts alone or in complexes to control both histone remodeling and gene expression (8). Approximately 40% of somatotropes are FOXO1 positive, with smaller populations of the other hormone-producing cell types also expressing the forkhead box protein (9). Nuclear FOXO1 is first apparent in pituitary cells at e14.5, suggesting that FOXO1 becomes active in the pituitary shortly before somatotrope differentiation (9). Foxo1 was recently identified as necessary for the correct timing of somatotrope differentiation on the basis of an investigation of a conditional knockout mouse model (4). A role in differentiation was not unexpected owing to similar findings in other tissues, including adipose and pancreas tissues (10).

Here, we describe a gain-of-function mouse model for investigating the role of Foxo1 in pituitary development, hypothesizing that early gene expression would result in early somatotrope differentiation. The model uses Foxg1-cre to drive FOXO1 expression in the oral ectoderm as early as e9.5. This includes the precursors of the Rathke pouch and progenitors for all of the hormone-producing cell types, approximately 5 days before typical FOXO1 nuclear detection (9, 11). Surprisingly, premature expression of FOXO1 in the pituitary primordium led to reduced anterior lobe hormone transcripts and fewer somatotropes and thyrotropes at e18.5 as well as reduced gland size. This is most likely the result of reduced induction of the oral ectoderm to invaginate and form the Rathke pouch. Overall, we showed that Foxo1 plays a dynamic role in anterior pituitary development, depending on time of activation, and its premature expression results in pituitary hypoplasia.

Materials and Methods

Animals and genotyping

Foxo1^AAA/AAA mice were generously provided by Dr. Ming O. Li at Memorial Sloan Kettering Cancer Center (12). Homozygous females were maintained as breeders and genotyped using the following primers: WT allele 5′-TGGGAAGTCTTGTCCCTCCAAT-3′ and 5′-TCGTGATCTGCAACTCCAGTCTTTC-3′, mutant allele 5′-AAAGTCGCTCTGAGTTGTTAT-3′ and 5′-AAGACCGCGAAGAGTTTGTC-3′. Foxg1^cre/+ mice were purchased from Jackson Laboratories (stock no. 004337) (11, 13). Female Foxo1^AAA/AAA mice were crossed with male Foxg1^cre/+ to generate experimental offspring Foxo1^AAA/+;Foxg1^cre/+, referred to as CA-FOXO1, and Foxo1^AAA/+;Foxg1^+/+, serving as littermate controls and hereby referred to as WT. Presence of the Foxg1-cre allele was detected with the primers 5′-GCGGTCTGGCAGTAAAAACTATC-3′ and 5′-GTGAAACAGCATTGCTGTCACTT-3′.

All mice were housed on a 12-hour light/dark cycle with feed (Formulab Diet 5008; Purina Mills) and water ad libitum. Animals were euthanized using CO₂ inhalation, and embryos were collected for analysis. Embryonic age was determined by designating the date of copulatory plug detection as e0.5. All procedures using mice were approved by the Southern Illinois University Animal Care and Use Committee. All experiments were conducted in accordance with the principles and procedures outlined in the National Institutes of Health Guidelines for the Care and Use of Experimental Animals.

Histology and volume measurement

Embryos and dissected pituitary gland tissue taken for immunohistochemistry (IHC) were fixed in 4% formalin in PBS then dehydrated in graded ethanol (50% then 80%) and stored at −20°C in 80% ethanol until paraffin processing.

Five-μm paraffin sections of processed tissue were stained using hematoxylin and eosin (H&E) to visualize general morphology. The volume of the e18.5 anterior lobe was measured using the Cavalieri method of segmentation with point counting (stereology.info/volume). Briefly, the entire pituitary of selected specimens was sectioned, including one CA-FOXO1 and one WT animal from each of three litters. Various sections were then measured for the number of points on a grid system with arbitrary spacing of 176.3 μm². The final volume of each anterior lobe was then calculated using the following formula:

$$\widehat{V}=A_{p}m\prime \overline{t}({\displaystyle \sum }_{i=0}^n{P}_i)$$

No shape factor was used.

Immunofluorescence

IHC analysis of paraffin sections was performed by deparaffinization in xylene and rehydration in graded ethanol (100% then 95%) then PBS. Slides were incubated in 1.5% H₂O₂ in PBS for 20 minutes to inactivate endogenous peroxidases. Antigen retrieval was performed by boiling slides in 10 mM of citric acid buffer (pH 6.0) for 5 to 10 minutes according to the protein being detected. Sections were then blocked for 60 minutes in Tyramide Signal Amplification (TSA) Kit Blocking Solution (TSB; Perkin Elmer). Primary antibody was diluted in TSB as detailed in Table 1 and incubated on tissue sections overnight at 4°C in a humidified chamber. Sections were then washed 3 × 5 minutes in 0.05% PBS-Tween (PBS-T). Secondary antibody was added for 60 minutes at room temperature followed by 3 × 5‒minute PBS-T washes. Depending on the protein being detected, sections were either washed and counterstained with 4′,6-diamidino-2-phenylindole (DAPI) for 5 minutes [GH, LH β (LHB)] or incubated with streptavidin–horseradish peroxidase (SA-HRP) for 30 minutes, washed, and then incubated with fluorophore [fluorescein isothiocyanate or tetramethylrhodamine isothiocyanate (TRITC) from TSA kit (Perkin Elmer)] for 10 minutes, followed by washes and then incubated with DAPI [HA, FOXO1, phospho-histone H3 (pHH3), fibroblast growth factor receptor (FGFR) 1, LHX3, PITX2, SHH, CTNNB1, TSH β (TSHB), ACTH, prolactin (PRL), POU1F1] (14–27). Sections were mounted using immunofluorescent mount [0.5 mM polyvinyl alcohol, 0.12 M Tris (pH 8.5), 0.3% w/v glycerol, 2.5% w/v 1,4-diazabicyclo[2.2.2]octane] and glass coverslips and imaged as described later.

Table 1.

Antibodies

Peptide/Protein Target	Name of Antibody	Manufacturer, Catalog No., or Name of Source	Species Raised in, Monoclonal or Polyclonal	Dilution Used	RRID	TSA Amplification
FoxO1 (C29H4)	FOXO1	Cell Signaling Technology, 2880	Rabbit, monoclonal	1:50	AB_2106495	Yes
HA	HA-tag	Genscript, A00168	Goat, polyclonal	1:1000	AB_2313847	Yes
BrdU	MoBU-1	Thermo Fisher Scientific, B35128	Mouse, monoclonal	1:100	AB_2536432	Yes
Activated β-catenin	Anti-ABC	Millipore, 05-665	Mouse, monoclonal	1:100	AB_309887	Yes
FGFR1	Anti-FGFR1	abcam, ab10646	Rabbit, polyclonal	1:100	AB_297367	Yes
SHH		R&D Systems, AF464	Goat, polyclonal	1:50	AB_355373	Yes
Mouse GH	GH	A.F. Parlow National Hormone and Peptide Program, AFP5672099	Rabbit, polyclonal	1:10,000	AB_2721132	No
Rat LHB	Anti-rat LHb	A.F. Parlow National Hormone and Peptide Program, AFP-22238790GPOLHB	Guinea pig, polyclonal	1:625	AB_2665565	No
Rat TSHB	Rat TSHb	A.F. Parlow National Hormone and Peptide Program, AFP-1274789	Rabbit, polyclonal	1:2000	AB_2665563	Yes
Rat ACTH	ACTH	A.F. Parlow National Hormone and Peptide Program, AFP-156102789	Rabbit, polyclonal	1:500	AB_2313902	Yes
Mouse PRL, recombinant	PRL	A.F. Parlow National Hormone and Peptide Program, AFP107120402	Rabbit, polyclonal	1:10,000	AB_2721133	Yes
Lim 3 (LHX3)	LHX3	DSHB, 67.4E12	Mouse, monoclonal	1:100	AB_2135805	Yes
PITX2	PITX2	Novus, H00005308-M01	Mouse, monoclonal	1:100	AB_547066	Yes
pHH3	pHH3 (Ser10)	Millipore, 06-570	Rabbit, polyclonal	1:100	AB_310177	Yes
POU1F1	Anti-PIT1	Simon J. Rhodes, lot 1603	Rabbit, polyclonal	1:2000	AB_2722652	Yes

Abbreviations: FGFR, fibroblast growth factor receptor; LHB, LH β; PRL, prolactin; TSA, tyramide signal amplification; TSHB, TSH β.

Cells in S-phase were labeled by first injecting pregnant dams (IP) with 100 mg/kg of body weight 5-bromo-2′-deoxyuridine (BrdU; BP2508250; Fisher) 2 hours before collection (28). Embryos were collected, processed, sectioned, dewaxed, and rehydrated as explained previously. Samples were then soaked in 1.5% H₂O₂ in PBS for 20 minutes to remove potential interference of endogenous peroxidases. Antigen retrieval was performed by boiling sections in 10 mM of citric acid buffer (pH 6.0) for 10 minutes. After blocking for 60 minutes in TSA kit TSB Block (Perkin Elmer), sections were incubated overnight at 4°C with anti-BrdU antibody (B35128; clone ZBU at 1:100; Invitrogen). After 3 × 5‒minute washes in 0.05% PBS-T, sections were then treated with secondary antibody, biotinylated goat anti-mouse (BA-9200; Vector Laboratories), for 60 minutes at room temperature. To amplify the signal, SA-HRP and TRITC from the TSA kit were used as previously described, followed by a DAPI stain for nuclei.

Apoptosis was measured in sagittal tissue sections of mice at e10.5, e11.5, and e12.5, prepared as mentioned previously. The DeadEnd™ Fluorometric TUNEL System (G3250; Promega) was used according to the kit protocol. Tissue sections were stained with DAPI for 5 minutes to visualize nuclei and then mounted with immunofluorescent mount and glass coverslips before imaging.

All IHC and immunofluorescent images were taken using a Retiga 2000R digital camera attached to a Leica DM 5000B fluorescent microscope. Individual captures of fluorescein isothiocyanate, TRITC, and DAPI channels were merged using Adobe Photoshop CS3. Some images were brightened for illustrative purposes; however, the exact alterations were duplicated in both control and experimental images to maintain the ability to compare results. Anterior lobe area was measured using ImageJ version 1.49u (100).

Quantitative RT-PCR

Embryos collected for quantitative RT-PCR (RT-qPCR) were kept in PBS on ice until decapitation, and either whole pituitary (e15.5) or only adenohypophysis (e18.5) was dissected therefrom and placed immediately in lysis buffer for RNA isolation (RNAqueous Micro Kit; AM1931). RNA was then isolated following manufacturer’s instructions. All samples were treated with DNase I for 20 minutes using reagents from the same kit according to provided protocol. mRNA was reverse transcribed to cDNA using the Promega M-MLV kit according to kit instructions (M5313). cDNA from one WT and one CA-FOXO1 animal from a total of eight litters was then analyzed. Five ng of cDNA were used per reaction in 15 μL of total reaction volume with 1 μM of primer mix, 1X PowerUp SYBR Green Master Mix (A25743; Applied Biosystems), and water. All primer sequences can be found in Table 2. Samples were run in duplicate and analyzed using the ΔΔCt method using the Bio-Rad CFX Connect Real-Time PCR Detection System. Two normalization genes were included: transferrin receptor (Tfrc) and RNA polymerase II subunit b (Polr2b). Data presented were normalized to Polr2b for ΔCt, then calculated relative to Tfrc for ΔΔCt. Polr2b and Tfrc have been tested and found to be expressed at consistent levels between genotypes and ages.

Table 2.

Primer Sequences

Gene	Forward	Reverse
Gh1	TGG CTG CTG ACA CCT ACA AAG AGT	GCG AGA AGC GAA GCA ATT CCA T
Fshb	ACA AGG TTA GGC TGC TCA ACT CCT	TCC TGT CTG CCT GGA ACA ATC CAT
Lhb	ATG AGT TCT GCC CAG TCT GCA TCA	AGG GCT ACA GGA AAG GAG ACT ATG
Prl	AGG CCT ATC CTG AAG CCA AAG GAA	TTG GCA CCT CAG GAC CTT GAG AAA
Pomc	AAG AGC AGT GAC TAA GAG AGG CCA	TGC AGA GGC AAA CA GAT TGG AGG
Tshb	TTG GCA CCT CAG GAC CTT GAG AAA	TGC AGT AGT TGG TTC TGA CAG CCT
Nr5a1	TCC TGA ACA ACC ACA GCC TCG TAA	TTG CCC AAA TGC TTG TGG TAC AGG
Tbx19	ATG GAG GTG GGC AGA TAA TGC	GCC TTG GCG AAA GGG TTG TA
Gnrhr	AGC TCG GCT GAG AAC GCT AAA GAT	ATT CAC TGG CTC TGA CAC CCT GTT
Pou1f1	CTG ACG CTT CTG CTG CCC	ATG AAG TCC TGT CGC TGT GG

In situ hybridization

In situ hybridization was performed as previously described (29). Briefly, slides were deparaffinized, washed, permeabilized, digested with proteinase K, acetylated, and incubated in hybridization solution. The Prop1 probe was linearized with appropriate enzymes and transcribed with polymerase in the presence of digoxigenin-labeled nucleotides. Labeled probe was denatured for 10 minutes and incubated overnight at hybridization temperature. Slides were then washed in 0.5× formamide solution then 0.5× sodium citrate and blocked in in situ hybridization-blocking solution (10% heat-inactivated sheep serum, 2% bovine serum albumin, and 0.1% Triton-X 100 in Tris-buffered saline). After blocking, the slides were incubated with antidigoxigenin antibody (1:500; Roche) diluted in in situ hybridization block. Slides were then washed in Tris-buffered saline (pH 7.5 and then pH 9.5) and incubated for 18 hours in 4-nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate, 4-toluidine salt developing solution (1:50; Roche).

Statistical analysis

Data were analyzed using the Student t test for comparison between genotypes, except in the case of e18.5 qRT-PCR, which used the Mann-Whitney test. Outliers were identified using the ROUT method. All data were analyzed using Prism 7 (GraphPad Software, Inc). Asterisks indicate significant findings, where (*) refers to P < 0.05, (**) refers to P < 0.01, (***) refers to P < 0.001, and (****) refers to P < 0.0001.

Results

Nuclear FOXO1 and HA were detected in CA-FOXO1 adenohypophysis during development

To further our understanding of the role of FOXO1 in pituitary gland development, an animal model was generated using cre/loxP technology. Foxo1^AAA/AAA mice possess a recombinant human FOXO1 gene on the Rosa26 locus, with point mutations at Thr24Ala, Ser256Ala, and Ser319Ala (12). These amino acids are typically phosphorylated downstream of AKT signaling, which then results in the nuclear expulsion of FOXO1 and its eventual degradation (30). By mutating them to alanine, AKT-mediated regulation is abrogated and FOXO1 remains active in the nucleus. The expression of FOXO1^AAA is controlled by a stop cassette upstream of the gene flanked by two loxP sites. Crossing these mice with Foxg1^cre/+ animals results in premature expression of FOXO1 in the oral ectoderm as early as e9.5, as well as in the ventral telencephalon, and ectopic expression in the retina and foregut. The resultant experimental model is hereby referred to as CA-FOXO1. This cre-driver was chosen because is not expressed in primordial hypothalamic tissue and therefore should not exert its effect in the ventral diencephalon, or pituitary organizer (11). Some advantages of using the Foxg1-cre are its early action, specificity to the oral ectoderm, and target of all anterior lobe progenitors.

To verify the localization of FOXO1 in the CA-FOXO1 mouse model, embryos were collected at e10.5, e12.5, and e18.5. CA-FOXO1 pups die either during or shortly after birth; therefore, all data presented are from embryonic stages. IHC was performed using both anti-FOXO1 and anti-HA antibodies on distinct but adjacent tissue sections [Fig. 1(a)]. The anti-HA antibody was used to detect the protein fusion tag, which is expressed only where cre-mediated recombination has occurred, thereby distinguishing ectopic FOXO1 from endogenous protein. Per the expected expression pattern based on Foxg1 gene expression, HA was evident in the ventral telencephalon in a more expanded region than endogenous FOXO1, oral ectoderm (where FOXO1 is normally detected a few days later in development), foregut, and retina (retina not pictured), but not in the ventral diencephalon [Fig. 1(a) and 1(b); data not shown]. Cellular localization was also nuclear, matching the predicted antiexpulsion design from the mutated amino acids. The detection of both nuclear FOXO1 and HA persisted through e18.5 [Fig. 1(b)]. Overall, there were fewer cells positive for nuclear HA or FOXO1 than expected in mutant animals at the later time point (e18.5). However, from e10.5 to e12.5, robust detection of both nuclear HA and FOXO1 was evident. Therefore, the pituitary-related changes in the CA-FOXO1 model are expected to be the result of premature FOXO1 protein activation.

Figure 1.

Ectopic expression of FOXO1 was detected by both anti-HA and anti-FOXO1 antibodies. (a) Midsagittal sections from e10.5 embryos comparing FOXO1 (green) and HA (red) with nuclei in blue via IHC. Endogenous FOXO1 was detected throughout the ventral telencephalon and rostral of the ventral diencephalon in the anterior hypothalamus in both WT and CA-FOXO1 animals. CA-FOXO1 animals showed an expanded region of nuclear FOXO1-positive cells within the telencephalon and in the oral ectoderm, particularly notable when HA expression was compared with endogenous protein. Additional ectopic expression was also observed in the spinal cord and tegmentum. (b) CA-FOXO1 mice were collected at e10.5, e12.5, and e18.5, and IHC using TSA amplification was performed against HA (red) and FOXO1 (green). A minimum of three sections per animal were analyzed, and CA-FOXO1 mice are shown with WT controls (insets). Midsagittal sections were used for e10.5 and e12.5 and coronal sections for e18.5. HA protein was evident in all CA-FOXO1 sections but not in control tissue. Nuclear FOXO1 was detected in CA-FOXO1 pituitary gland tissue but not in WT glands except at e18.5. (c) H&E staining of CA-FOXO1 and WT embryos to determine pituitary gland morphology during embryonic development. CA-FOXO1 adenohypophyseal tissue appeared reduced compared with that of controls. (d) Comparison of pituitary volume (anterior lobe) between CA-FOXO1 and WT animals using the point-count stereological technique. CA-FOXO1 anterior lobe at e18.5 was approximately 12% the size of WT controls. Data are represented as mean ± SEM. Significance was determined using the Student t test. **P < 0.01. All scale bars indicate 100 µm. oe, oral ectoderm; poa, preoptic area; tg, tegmentum; vd, ventral diencephalon; vt, ventral telencephalon.

CA-FOXO1 pituitary gland was hypoplastic

Histological analysis of CA-FOXO1 sections using H&E staining revealed a significant change in both pituitary gland size and morphology compared with that of WT animals. Midsagittal sections of e10.5 embryos demonstrated a narrower oral ectoderm that is not always in contact with the neural ectoderm as in normally developing embryos [Fig. 1(c)]. The oral ectoderm had not closed off in most CA-FOXO1 sections analyzed at e11.5 as in WT controls, and a smaller portion appeared to be invaginating. On the basis of what seems to be reduced induction at e10.5 to e11.5, the resultant Rathke pouch was smaller at e12.5. At e18.5, adenohypophysis was consistent in morphology and significantly smaller than in control tissue. This was confirmed by performing volumetric analysis of three whole pituitary glands from both mutant and control tissues wherein the CA-FOXO1 pituitary volume was 12% the size of the WT tissue [Fig. 1(d)]. Some morphological abnormalities were also observed in the retina, parts of the forebrain, and the palate of the CA-FOXO1 embryos, with varying severity (data not shown).

Apoptosis was increased in CA-FOXO1 animals in the oral ectoderm at e11.5

To investigate whether premature expression of FOXO1 was causing hypoplasia via apoptosis, TUNEL assays were performed. During typical murine development, there is little to no apoptosis in the region of the oral ectoderm at e10.5. This was consistent in both WT and CA-FOXO1 animals [Fig. 2(a)]. However, the anterior hypothalamus in CA-FOXO1 animals presented with significantly more apoptotic cells than in controls, specifically at e10.5 [Fig. 2(b)]. Although this region was not directly undergoing cre-mediated recombination, it is possible changes in surrounding tissue were affecting cell viability. In addition, at e11.5 ectopic apoptosis was detected in the CA-FOXO1 embryo in the caudal portion of the invaginating pouch but never in controls. At e12.5, the number and localization of cells undergoing programmed death appeared similar in both mutants and controls [Fig. 2(b)]. Therefore, the apoptosis presented in Fig. 2 may have contributed to the hypoplasia observed in the resultant pituitary gland in CA-FOX1 animals but is most likely not the primary cause.

Figure 2.

Apoptosis was evaluated using the TUNEL assay. Three sections per animal were analyzed. (a) CA-FOXO1 and control embryos were compared by evaluating the number and localization of apoptotic cells at e10.5 to e12.5. Cells undergoing apoptosis are shown in green. DAPI was used as a stain for nuclei (blue). (b) The number of apoptotic cells was counted and compared between WT and control animals according to localization. No apoptotic cells were present in the Rathke pouch in WT controls, and significantly more cells underwent apoptosis in the anterior hypothalamus at e10.5 in CA-FOXO1 embryos than in control embryos, but not at other time points. Data are represented as mean ± SEM and were compared using the Student t test. *P < 0.05. All scale bars indicate 100 µm. ah, anterior hypothalamus; mes, mesenchyme; ND, not detected; oe, oral ectoderm; poa, preoptic area; rp, Rathke pouch; vd, ventral diencephalon.

To ascertain whether FOXO1 was responsible for hypoplasia as the result of disrupting cellular proliferation and glandular expansion, BrdU incorporation and pHH3 protein detection were analyzed [Fig. 3(a)]. BrdU is incorporated into newly synthesized DNA during cellular replication, whereas histone H3 is phosphorylated during phases G2 and M of the cell cycle (31, 32). Total cells stained for BrdU incorporation were counted and compared with total DAPI-stained cells in each section. IHC for PITX2 was used to identify the region of the oral ectoderm being induced to form the Rathke pouch in CA-FOXO1 tissue at e10.5 (see Fig. 4). BrdU analysis showed no significant difference in the total number of proliferating cells at e10.5 and e11.5 [Fig. 3(b)]. A marked increase in proliferating cells was apparent in CA-FOXO1 animals at e12.5 directly surrounding the lumen of the Rathke pouch. Embryonic samples at the aforementioned time points were also analyzed for pHH3 using IHC; however, no difference was seen between WT and CA-FOXO1 embryos [Fig. 3(c) and 3(d)]. This indicates proliferation was not disrupted in CA-FOXO1 animals at e10.5 to e12.5. Therefore, changes in proliferation from e10.5 to e12.5 do not appear to be the primary cause of pituitary hypoplasia in the CA-FOXO1 mouse model.

Figure 3.

Cellular proliferation was evaluated using the BrdU incorporation assay and IHC against pHH3. Two to three sections per animal were analyzed for a minimum of three (pHH3) or five (BrdU) litters. (a) CA-FOXO1 and control embryos were compared by evaluating the number and localization of proliferating cells at e10.5 to e12.5. Proliferation was assessed by injecting pregnant dams with BrdU and collecting pups 2 hours later. Midsagittal sections were evaluated for BrdU incorporation using IHC, with BrdU shown in red and nuclei shown in blue. (c) Anti-pHH3 antibody was used to detect cells in G2 or M phases of the cell cycle (red). All scale bars indicate 100 µm. (b, d) The numbers of proliferating cells were counted and compared with the total number of DAPI-stained cells in the anterior lobe and measured for significance using the Student t test. Data are represented as mean ± SEM. A significant increase was seen in the BrdU assay at e12.5. ***P < 0.001.

Figure 4.

(a) Immunofluorescent detection of important early signaling and transcription factors in pituitary development at e10.5. All signals were amplified using TSA signal amplification. Adjacent sections of control and CA-FOXO1 tissues were compared for protein localization of PITX2 and LHX3. Additional midsagittal sections were evaluated for FGFR1, β-catenin, and SHH. Three to five sections per animal were analyzed. White arrowheads delineate the region of ventral diencephalon positive for SHH. All scale bars indicate 100 µm. (b) The number of cells positive for either PITX2 or LHX3 were counted in WT and CA-FOXO1 oral ectoderm. There were significantly fewer cells presenting with the transcription factors in the CA-FOXO1 embryos, indicating fewer cells with pituitary identity. Data are represented as mean ± SEM. The Student t test was used to determine significance. ***P < 0.001; ****P < 0.0001. (c) Prop1 was detected using in situ hybridization. Both WT and CA-FOXO1 animals expressed Prop1 in the Rathke pouch at e12.5, although it appeared reduced in mutant animals. Scale bars indicate 100 µm.

Decreased pituitary identity was observed at e10.5 in CA-FOXO1 mice

To determine whether premature activation of FOXO1 was potentially disrupting proper pouch induction, IHC was performed at e10.5 against proteins used to mark early pituitary development, including PITX2 and LHX3. Staining for PITX2 was observed throughout the dysmorphic oral ectoderm, although it did not appear to extend as far ventrally in reference to the diencephalon as in controls [Fig. 4(a)]. LHX3 appeared in only a few cells in the region thereof when adjacent tissue sections were compared [Fig. 4(a)]. The numbers of both PITX2-positive and LHX3-positive cells were significantly reduced in CA-FOXO1 pituitary glands analyzed at e10.5 [Fig. 4(b)]. This demonstrates a reduced pituitary identity among cells differentiating from the oral ectoderm to form the Rathke pouch. This result is most likely independent of the proximity of the oral ectoderm to the ventral diencephalon on the basis of the analysis of multiple sections per animal. In addition, in situ hybridization was performed to investigate whether Prop1 was normally expressed at e12.5. CA-FOXO1 animals analyzed appear to have reduced Prop1 in the Rathke pouch [Fig. 4(c)].

Protein localization of signaling factors in oral ectoderm and ventral diencephalon was primarily intact in the CA-FOXO1 mouse model

SHH, β-catenin, and FGFR1 were also investigated to evaluate some of the pathways most important for proper pituitary gland development. It is possible the reduction in LHX3-positive cells is the result of decreased induction of the oral ectoderm from signaling factors from the ventral diencephalon. Each of these signaling pathways must be functional in order for proper pituitary gland formation and the emergence of the different hormone-producing cell types to occur during development (1). SHH was detected in the ventral diencephalon and appeared more dorsally extended in CA-FOXO1 mice than in controls [Fig. 4(a), arrows]. Activated β-catenin (CTNNB1) and FGFR1 staining appeared similar to that of WT controls, albeit in dysmorphic structures [Fig. 4(a)]. As stated previously, this did not appear to depend on the nearness of the oral and neural ectoderm according to the examination of multiple sections per mouse.

CA-FOXO1 pituitary demonstrated a significant decrease in the proportion of somatotropes and thyrotropes

Foxo1 deletion from the developing pituitary gland resulted in delayed somatotrope differentiation (4). In accordance with this finding, it was hypothesized that premature FOXO1 activation would cause early differentiation of the same. Gh1 transcripts were analyzed at e15.5 to detect whether premature FOXO1 was sufficient to drive an early increase in gene transcription. Interestingly, a significant decrease was observed in experimental animals, with a reduction in Pou1f1 transcript also observed [Fig. 5(b)]. In accordance with this finding, no early detection of GH protein was observed by IHC at e15.5 in CA-FOXO1 animals [Fig. 5(a)]. CA-FOXO1 pituitary gland analyzed at e15.5 did present with POU1F1-positive cells, indicating the initial differentiation of somatotropes was intact, albeit to a lesser degree [Fig. 5(a)].

Figure 5.

(a) Immunofluorescent detection of GH and POU1F1 in midsagittal sections from e15.5 animals. Neither WT nor CA-FOXO1 mice displayed positive immunostaining for GH, whereas POU1F1 was detected in both. Representative sections are shown from three and four litters (POU1F1 and GH, respectively), and scale bars indicate 100 µm. (b) qRT-PCR analysis of various early signaling factors at e15.5. Litters were analyzed by combining both male and female samples. Expression is presented relative to Tfrc. Data are expressed as mean ± SEM of six animals of each genotype. Statistics were performed using the Student t test for significance. **P < 0.01; ****P < 0.0001.

In addition to the investigation of Gh1, other factors important for the proper differentiation of the hormone-producing cell types of the anterior lobe were analyzed via qRT-PCR. Tbx19 is an early factor in the adoption of the corticotrope cell fate and Nr5a1 and Gnrhr in the case of gonadotropes, and both Pomc and Tshb can be detected at e15.5 in the Rathke pouch, with the latter severely reduced (3). All factors with the exception of Gnrhr were significantly reduced in CA-FOXO1 animals at e15.5 [Fig. 5(b); Gnrhr not shown].

To determine whether the hormone-producing cell types were functionally differentiating in the CA-FOXO1 mouse model, expression of the different anterior lobe hormones at both the mRNA and protein levels was analyzed. By e18.5, Gh1 mRNA was still reduced compared with controls, and all other anterior lobe hormone transcripts were significantly lower than those of WT adenohypophyseal samples (Fig. 6). Only the anterior lobe was analyzed to prevent posterior lobe RNA from diluting the contents of the former in the case of the hypoplastic CA-FOXO1 samples. IHC was then used to evaluate whether hormone protein was generated in the hypoplastic pituitary. GH, TSHB, PRL, LHB, and ACTH were all present, indicating that terminal differentiation of the different hormone-producing cell types occurred in spite of aberrant FOXO1 activity [Fig. 7(a)]. Hormone-producing cell numbers were significantly decreased for somatotropes, thyrotropes, and corticotropes, but not for gonadotropes [Fig. 7(b)]. Hormone-positive cells were then compared with the total DAPI-stained area of the anterior lobe. This analysis revealed relatively similar populations of corticotropes and gonadotropes in CA-FOXO1 glands compared with control glands. Interestingly, both somatotropes and thyrotropes were reduced in relative cell number [Fig. 7(c)]. Therefore, despite a decrease in hormone transcripts, the relative number of differentiated cell types was similar in CA-FOXO1 animals except in the case of somatotropes and thyrotropes. Lactotropes were not counted because very few are present during embryonic development (33). This implicates FOXO1 having a role in regulating the POU1F1 lineage because both somatotropes and thyrotropes differentiate from a POU1F1-positive precursor (34).

Figure 6.

qRT-PCR analysis of hormone transcripts in WT and CA-FOXO1 anterior lobe samples at e18.5. Litters were analyzed by combining both male and female samples. Expression is presented relative to Tfrc. Data are expressed as mean ± SEM of eight animals of each genotype. Statistics were performed using the Mann-Whitney test for significance. **P < 0.01; ***P < 0.001.

Figure 7.

(a) Immunofluorescent detection of anterior lobe hormones in coronal sections of e18.5 animals. Three or four sections per animal were analyzed. Scale bars indicate 100 μm. Top panels are WT and bottom panels are CA-FOXO1 representative samples. (b) The total number of immune-positive cells were counted per section. Data are represented as mean ± SEM. (c) Immune-positive cells were counted and normalized according to total anterior lobe area by measuring DAPI staining with ImageJ. Data are expressed as mean ± SEM of three animals of each genotype. The number of GH-positive cells and TSH-positive cells were significantly reduced in CA-FOXO1 mice. The Student t test was used to determine significance. ***P < 0.001; ****P < 0.0001.

Discussion

We previously reported that deletion of Foxo1 delayed somatotrope differentiation (4). The current study utilizing a CA-FOXO1 gain-of-function model provides additional insight into the potentially multifaceted role of FOXO1 in both development and GH production. In CA-FOXO1 mice, nuclear FOXO1 was present in the oral ectoderm by e9.5. The region of ectopic CA-FOXO1 expression is important to clarify because of the influence of the ventral diencephalon on the oral ectoderm, which dictates the morphology of the eventual gland. No HA-positive or FOXO1-positive cells were detected via IHC in the ventral diencephalon. In addition, the nuclear localization of FOXO1 persisted, albeit to a reduced degree, through e18.5. Therefore, the mouse model accomplished its design: to produce constitutively nuclear FOXO1 prematurely in the developing anterior pituitary.

CA-FOXO1 animals present with pituitary hypoplasia as early as e10.5. To investigate the potential mechanism thereof, TUNEL staining was performed at e10.5 to e12.5. At e11.5, CA-FOXO1 mice demonstrated cell death in the invaginating region of the oral ectoderm, whereas no apoptotic cells were detected in the pouch of WT animals. The apoptotic cells observed may be CA-FOXO1 cells undergoing extrusion because of FOXO1-mediated apoptosis or potentially their inability to adapt a pituitary fate. This latter phenomenon was presented in an elegant study in Drosophila by Bielmeier et al. (35), wherein small clonal populations of cells with aberrant expression of the Foxa2 orthologous gene fkh were present alongside WT cells in imaginal discs. Smaller clone sizes of fkh cells were more likely to undergo apoptosis in their model, which they described as “size-dependent elimination of misspecified clusters,” in reference to the change in transcription factor expression in those groups (35). Therefore, in addition to FOXO1-mediated apoptosis, a potential mechanism by which FOXO1 cells are eliminated from the developing pituitary could be under the direction of the surrounding epithelium as a consequence of their aberrant gene expression. It is doubtful whether these isolated events at e11.5 were sufficient to cause the severe hypoplasia observed; however, they should be considered for their contribution.

To ascertain whether FOXO1 was responsible for hypoplasia as the result of disrupting cellular proliferation and glandular expansion, BrdU incorporation and pHH3 staining were analyzed. BrdU is incorporated into newly synthesized DNA during cellular replication (31). Total cells positive for BrdU were counted and compared with total DAPI-stained cells in each section. IHC for PITX2 was used to identify the region of the oral ectoderm being induced to form the Rathke pouch in CA-FOXO1 tissue at e10.5 (see Fig. 4). BrdU analysis showed no significant difference in the total number of proliferating cells at e10.5 and e11.5 [Fig. 3(b)]. A marked increase in proliferating cells was apparent in CA-FOXO1 animals at e12.5 directly surrounding the lumen of the Rathke pouch. Typically, this region is characterized with a stem cell‒like population as cells migrate from the luminal area as undifferentiated progenitors to other regions of the gland during differentiation (36, 37). They then migrate ventrally, stop proliferating, and differentiate into the various cell types of the anterior lobe (2). Because the majority of differentiating, quiescent cells were absent in the CA-FOXO1 anterior lobe, the ratio of proliferating cells to total gland was increased compared with that of WT mice. pHH3 staining was not different between CA-FOXO1 mice and littermate controls. Therefore, changes in proliferation from e10.5 to e12.5 did not appear to be the primary cause of pituitary hypoplasia in the CA-FOXO1 mouse model.

The oral ectoderm contains fewer cells that have adopted a pituitary identity in CA-FOXO1 animals. During pituitary gland organogenesis, cells from the oral ectoderm respond to signals from the ventral diencephalon and mesenchyme to adopt pituitary identity and can be distinguished by their expression of transcription factors, including PITX2 and LHX3 (1, 38). In the CA-FOXO1 embryo, PITX2 was readily detected in the dysmorphic oral ectoderm at e10.5, confirming the tissue’s identity while highlighting the irregular shape. Compared with WT controls, the PITX2-positive cells did not extend as far ventrally or caudally through the invaginating region. In the CA-FOXO1 model, LHX3 comprised a smaller region of the oral ectoderm than PITX2, with less consistent expression throughout the gland. This indicates premature FOXO1 activity promoted a dysmorphic and hypoplastic gland that also possessed reduced expression of specific early transcription factors. The importance of Lhx3 has been shown using a null mouse model in which Lhx3 deletion resulted in an increase in apoptosis in the ventral region of the invaginating oral ectoderm and ultimately led to pituitary gland hypoplasia (39, 40). FGFs have been shown both in vivo and in vitro to be vital for the proliferation and specification of stem cells to become and remain LHX3 positive (40–42). The reduced pituitary identity in CA-FOXO1 mice did not appear to be the result of a loss of FGFR1 (one of the receptors through which FGF8/10 exert their function) because IHC for the same was similar between mutants and controls (43). Nor does it seem to be through the canonical β-catenin pathway, as all cells in the region were positive for activated β-catenin in control and CA-FOXO1 animals at e10.5.

A close analysis of hypothalamic morphology demonstrated a less pronounced preoptic area at e10.5 in CA-FOXO1 animals. However, this did not persist through e11.5 and e12.5. The preoptic area is also not known to be directly involved in adenohypophyseal patterning. Moreover, the posterior lobe of e18.5 animals did not appear smaller when CA-FOXO1 and WT animals were compared. These observations reinforce the hypothesis that changes in the oral ectoderm and a reduction in pituitary identity are responsible for the hypoplasia observed in CA-FOXO1 animals as opposed to a defect in signaling from the pituitary organizer.

The only other signaling factor observed by IHC to be irregular between CA-FOXO1 and WT mice is SHH. Normally at e10.5, SHH is active in the ventral diencephalon and in the oral ectoderm on either side of the invaginating pouch and is tightly regulated as part of a gradient with BMP2/4 (44). In the CA-FOXO1 model, SHH appeared to extend more dorsally than in control embryos. In other mouse models, extension of SHH signaling leads to disruption of the BMP4 signaling gradient from the ventral diencephalon and can result in glandular hyperplasia, the opposite of what was seen in the CA-FOXO1 mice (38, 41). Because the change was so slight in CA-FOXO1 animals, it may not be sufficient to induce this phenotype. In addition, it brings into question whether the CA-FOXO1 mutation was directly causing the change in SHH or if it occurred through indirect means. As described, FOXO1 was clearly nuclear in the oral ectoderm and not in the ventral diencephalon, where the SHH changes occurred in the CA-FOXO1 embryo.

Owing to the finding that conditional Foxo1 deletion delayed somatotrope differentiation in the Foxo1 pituitary knockout mouse model, it was hypothesized that premature expression of FOXO1 would result in early detection of GH. GH is first observed in the mouse pituitary at e16.5 by IHC. Various studies have demonstrated in vivo and in vitro that premature somatotrope differentiation and population expansion can be induced by treatment with dexamethasone and/or corticosterone (45). In the CA-FOXO1 mouse model, significantly fewer Gh1 transcripts and no GH-positive cells were evident at e15.5. Upstream regulators of the somatotrope lineage, including Prop1 and POU1F1, were detected at e12.5 and e15.5, respectively. Although qualitative, the in situ results show reduced Prop1 in the Rathke pouch, potentially from fewer differentiating progenitors or a delay in Prop1 transcription. By e15.5, POU1F1 was present, indicating that the early steps of progenitor commitment and adoption of the POU1F1 lineage had taken place by this time. Either additional unknown machinery that is necessary for GH production was not yet in place, or aberrant expression of FOXO1 alone was insufficient to induce this change.

The CA-FOXO1 mouse model demonstrated that early activation of FOXO1 is not sufficient to prevent the emergence of any of the five trophic cell types, on the basis of protein data at e18.5. Early signaling and transcription factors specific to the different lineages were reduced at e15.5, including Nr5a1 and Tbx19. Tshb was severely decreased at e15.5 but recovered to some degree by e18.5, potentially indicating a delay in the differentiation of thyrotropes. By e18.5, the Tshb transcript level was significantly lower in CA-FOXO1 mice than in controls, although it was adequate to drive protein production. Prolactin, luteinizing hormone, GH, TSH, and ACTH were all detected by IHC in mutant animals. However, the relative population of the cells was not intact. Thyrotropes and somatotropes were reduced in relative numbers compared with those of WT controls. These cell types, as well as lactotropes, arise from a shared POU1F1-positive precursor. It is unknown whether lactotropes would be similarly reduced because of the perinatal lethality of the model. Therefore, it is possible that FOXO1 exerts effects that influence early factors preceding Pou1f1 expression, indirectly influencing the lineage, or that its effects persist enough through progenitor commitment to terminal differentiation to alter the number of somatotropes and thyrotropes. It is unclear why corticotropes and gonadotropes are not similarly affected, as their transcript expression was also significantly reduced. IHC is not a quantitative technique, and it is possible the corticotropes and gonadotropes detected were subfunctional despite hormone protein detection. In addition, other groups have highlighted the complexity of directly correlating mRNA to protein in biological systems, aptly demonstrating that transcription, mRNA stability, and response to adverse effects are important to consider (46). This is why both mRNA and protein were evaluated when possible to form conclusions.

CA-FOXO1 mice were alive through e18.5, as evidenced by their response to touch when removed from the uterus; however, within minutes of birth, the CA-FOXO1 pups had died (data not shown). The pituitary gland of the mutant animals had significantly fewer corticotropes than did WT littermates owing to the fact that the gland was hypoplastic [also see cell counts, Fig. 7(b)], and the pituitary-adrenal axis contributes to the development of the lung, gastrointestinal tract, liver, and heart (47). Insufficiencies in the function of any of these tissues could then be responsible for the lethality of the mutation. Specifically, mice may be dying because of respiratory distress syndrome, which can occur when there is insufficient corticosterone or thyroid hormone during development. Both of these hormones are necessary for proper lung development, and respiratory distress syndrome has been observed in pituitary-disease mouse models, including Prop1 knockout mice (48). It is also possible that the lethality was due to cell death in the telencephalon or foregut. TUNEL assays at early developmental time points showed increased apoptosis in these regions, which also undergo cre-mediated recombination (data not shown). Overall, although of unknown origin, the perinatal lethality allowed for only the investigation of ectopic FOXO1 expression during prenatal time points.

In sum, premature expression of FOXO1 during development led to adenohypophyseal hypoplasia. Although the different anterior lobe hormone-producing cells were able to functionally differentiate, the relative numbers of somatotropes and thyrotropes were reduced compared with those of controls. The major cause of hypoplasia most likely lies in failure to establish pituitary identity.

Glossary

Abbreviations:

BrdU

5-bromo-2′-deoxyuridine

DAPI

4′,6-diamidino-2-phenylindole

e

embryonic day

FGF

fibroblast growth factor

FGFR

fibroblast growth factor receptor

H&E

hematoxylin and eosin

IHC

immunohistochemistry

LHB

LH β

PBS-T

PBS-Tween

pHH3

phospho-histone H3

Polr2b

RNA polymerase II subunit b

PRL

prolactin

RT-qPCR

quantitative RT-PCR

SA-HRP

streptavidin–horseradish peroxidase

Tfrc

transferrin receptor

TRITC

tetramethylrhodamine isothiocyanate

TSA

tyramide signal amplification

TSB

tyramide signal amplification kit blocking solution

TSHB

TSH β

WT

wild-type