Prolactinomas are the most common type of functioning pituitary adenoma in humans, but the control of lactotroph proliferation remains unclear. Here, using microarray analysis, we show that estrogen treatment increased expression of Wnt4 mRNA in adult Fischer rat pituitary tissue. Dual immunofluorescence analysis revealed that Wnt4 expression was not confined to lactotrophs, but that it was expressed in all anterior pituitary cell types. Estradiol induced proliferation in the somatolactotroph GH3 cell line, in parallel with Wnt4 mRNA and protein induction. A reporter gene assay for TCF- and LEF-dependent transcription revealed that there was no activation of the canonical Wnt pathway in GH3 cells upon stimulation with Wnt-conditioned culture medium or coexpression of constitutively active mutant β-catenin. Expression of β-catenin in both GH3 cells and normal rat anterior pituitary cells was restricted to the cell membrane and was unaltered by treatment with estradiol, with no nuclear β-catenin being detected under any of the conditions tested. We show for the first time that Wnt4 affects non-canonical signaling in the pituitary by inhibiting Ca2+ oscillations in GH3 cells, although the downstream effects are as yet unknown. In summary, Wnt4 is expressed in the adult pituitary gland, and its expression is increased by estrogen exposure, suggesting that its involvement in adult tissue plasticity is likely to involve β-catenin-independent signaling pathways.
Prolactinomas represent approximately 60% of all pituitary tumors in humans (Gurlek et al., 2007). They result from abnormal lactotroph cell proliferation and usually display only very slow growth, but the pathogenesis of prolactinoma formation and progression has remained elusive. Classical oncogenic mechanisms seem unlikely to be involved in most cases, and none of the common genetic mutations causing cancer has so far been found to operate in prolactinomas (Levy, 2008). A number of proteins have been implicated in pituitary adenoma development; pituitary tumor transforming gene (PTTG) (Kim et al., 2007), basic fibroblast growth factor (bFGF) (Zhang et al., 1999), vascular endothelial growth factor (VEGF) (McCabe et al., 2002), bone morphogenetic protein 4 (BMP4) (Labeur et al., 2010), pituitary tumor apoptosis gene (PTAG) (Bahar et al., 2004) and histone deacetylase 2 (HDAC2) (Bilodeau et al., 2006), among several others, have been demonstrated to play a role in pituitary tumorigenesis but the origins of many tumors are still unknown.
Estrogen has long been known to exert a proliferative effect on lactotroph cells. High circulating estrogen levels during pregnancy result in lactotroph hyperplasia, pituitary enlargement and increased circulating prolactin (PRL) levels (Asa et al., 1982; Goluboff and Ezrin, 1969; Lloyd et al., 1988). In vitro, estradiol (E2) induces proliferation of the somatolactotroph GH3 cell line (Horvath and Kovacs, 1988; Kansra et al., 2005; Lieberman et al., 1982; Song et al., 1989) and the effects of E2 on lactotroph proliferation can be studied in vivo using the estrogen-sensitive Fischer 344 rat. In this model, lactotroph hyperplasia, and eventual prolactinoma formation, can be induced by treatment with estradiol or the synthetic estrogen diethylstilbestrol (DES) (Heaney et al., 1999; Mucha et al., 2007; Phelps and Hymer, 1983; Wiklund et al., 1981).
Wnt molecules, a family of 19 secreted signaling proteins in humans, are expressed in overlapping temporal and spatial patterns during development (Yavropoulou and Yovos, 2007). They regulate diverse cellular processes, such as proliferation, differentiation, apoptosis and cell survival (Willert and Jones, 2006), and are crucially involved in embryonic development. In this capacity, they are thought to interact with other traditional signaling pathways regulating development, including the BMP, sonic hedgehog (Shh), sox and notch pathways, although the interactions are complex. Abnormalities in Wnt signaling pathways have been associated with numerous cancers (Giles et al., 2003) and, in most cases, activation of canonical Wnt signaling is involved in cancer progression (Reya and Clevers, 2005; Willert and Jones, 2006).
The canonical Wnt signaling pathway centers on activation of β-catenin. In unstimulated cells, β-catenin is either bound to E-cadherin, at the cell membrane, or resides unbound in the cytoplasm (Benjamin and Nelson, 2008). Cytoplasmic β-catenin is bound rapidly by adenomatous polyposis coli (APC) and axin, allowing glycogen synthase kinase (GSK)-3β to phosphorylate β-catenin, which promotes its degradation (Price, 2006). Wnt binding to its frizzled (Fzd) receptor prevents axin and APC binding to β-catenin, thus inhibiting its destruction (Rao and Kuhl, 2010). This allows β-catenin to accumulate in the cytoplasm and then translocate into the nucleus, where it interacts with the transcription factors T-cell-specific transcription factor (TCF) and lymphoid-enhancer-binding factor (LEF) to induce transcription of Wnt target genes (Widelitz, 2005).
Wnt molecules also signal through two non-canonical pathways; in the Wnt–Ca2+ pathway, binding of Wnt to Fzd induces Ca2+ influx through calcium-release-activated calcium (CRAC) channels to regulate downstream effectors, such as calcineurin and nuclear factor of activated T-cells (NFAT) (Medyouf and Ghysdael, 2008). The Wnt–planar cell polarity (PCP) pathway regulates the polar orientation of a cell using small GTPases, such as Cdc42 and Rho-A, to alter cell–cell adhesion through cadherin molecules (Widelitz, 2005). To date, these two pathways have been poorly defined, and their downstream effects appear to be tissue- and cell-type-specific. Wnt4 has been shown to activate both canonical and non-canonical pathways (Wang et al., 2007; Chang et al., 2007); however, there is presently no information in the literature regarding regulation of non-canonical Wnt pathways in the pituitary.
Wnt signaling might be involved in pituitary pathophysiology: Wnt4 affects expansion of specific cell types in the normal developing mouse pituitary. It is expressed from embryonic day 9.5 (E9.5) to E14.5, and Wnt4−/− mice have diminished cell numbers in the anterior pituitary (Treier et al., 1998; Potok et al., 2008). Wnt4 is rapidly upregulated by estrogen during uterine growth in mice and this is associated with activation of the canonical signaling pathway (Hou et al., 2004). Molecules associated with Wnt signaling, such as the frizzled receptor, APC, β-catenin and TCF, are expressed in the developing mouse pituitary (Douglas et al., 2001), and β-catenin has been shown to interact with Prop-1 to control key stages in cell fate determination in the developing pituitary (Olson et al., 2006). The evidence regarding the downstream effects of Wnt molecules in the adult pituitary is contentious. Semba and colleagues (Semba et al., 2001) found frequent nuclear accumulation of β-catenin in 57% of the human pituitary adenomas that they studied. However, in a similar study using 54 human pituitary adenomas, Miyakoshi and colleagues (Miyakoshi et al., 2008) found that, although Wnt4 expression was increased in adenomas producing growth hormone (GH), thyroid-stimulating hormone (TSH) and PRL, β-catenin was restricted to the cell membrane and was never found in the nucleus, suggesting a non-canonical action of Wnt4 (Miyakoshi et al., 2008). The same group also reported that Wnt4 was specifically expressed in the majority of somatotrophs and in a few thyrotrophs in the untreated rat pituitary, and that estrogen increased Wnt4 expression in these cell types (Miyakoshi et al., 2009). Finally, downregulation of Wnt inhibitory factor 1 (WIF1) has been reported in a series of human pituitary tumors, and this was associated with increased nuclear β-catenin accumulation, and transfection of GH3 cells with WIF1 decreased cell proliferation (Elston et al., 2008).
In the present study we conducted a microarray analysis on pituitary tissue obtained from estrogen-treated Fischer 344 rats in order to identify novel genes and pathways involved in lactotroph hyperplasia. Among numerous genes upregulated by estrogen, we noted induction of Wnt4 and Wnt10a. The induction of Wnt4 mRNA and protein was confirmed both in vivo, in the rat pituitary gland after estrogen treatment, and in vitro, in the somatolactotroph GH3 cell line. We sought to clarify whether Wnt4 acts via the canonical pathway in the pituitary gland, and found no evidence for activation of canonical Wnt signaling in either GH3 cells or primary Fischer 344 rat pituitary cells. However, Wnt4 did inhibit Ca2+ oscillations in GH3 cells, suggesting that non-canonical signaling pathways might be involved in the pituitary remodeling response to estrogen.
Wnt signaling components are upregulated in estrogen-induced pituitary hyperplasia
Treatment of female Fischer 344 rats with DES for three weeks resulted in a twofold increase in uterus weight, a threefold increase in pituitary weight and a twofold increase in PRL mRNA expression (Fig. 1). Microarray analysis was conducted on the estrogen-treated pituitary tissue using an Affymetrix rat genome 2.0 array. Analysis of the array data revealed significant increases in a series of mRNAs that we expected to identify, including those encoding galanin, Pttg1 and transforming growth factor alpha (TGFα) (see Table 1 for selected examples, a more complete data set is shown in supplementary material Table S1). A number of genes involved in cell proliferation were also upregulated, including those encoding calpain 8, calbindin 3, cyclin A2, cyclin B2 and Ki67 (Mki67). Wnt4 and Wnt10a, which are both known to play roles in the developing pituitary, were both markedly increased (75- and 6.6-fold respectively). Other genes related to Wnt signaling pathways were also upregulated, including those encoding carboxypeptidase Z (Cpz), protein kinase C (PKC)-β1 and Wif1. Jagged 2 and delta-like 3, two members of the notch signaling pathway, which is known to interact with the Wnt pathway, were also increased (fourfold and sevenfold, respectively). To validate the microarray analysis, quantitative RT–PCR analysis confirmed significant increases in Wnt10a, Wnt4 and Cpz expression in rat pituitary tissue (Fig. 2).
Wnt4 protein expression in rat anterior pituitary tissue
In order to identify the Wnt4-expressing endocrine cell types in the pituitary, adult female rat pituitary tissue was analyzed by dual immunofluorescence staining for Wnt4 with GH, PRL, adrenocorticotropic hormone (ACTH), luteinizing hormone (LH) and TSH-β (Fig. 3C–G). The specificity of the anti-Wnt4 antibody was assessed using adult Fischer 344 rat kidney tissue, and we confirmed expression of Wnt4 in cortical renal tubules and absence of expression in adjacent cells (Fig. 3A), in keeping with the previously observed expression patterns in this tissue (Terada et al., 2003). All of the endocrine cell types in the anterior pituitary were found to express Wnt4, although the prevalence was highest in somatotrophic cells. No major alterations in coexpression patterns were seen in animals treated with estrogen.
Estradiol induces Wnt4 expression in somatolactotroph GH3 cells
To investigate the effect of estrogen on Wnt signaling in the pituitary, we used the somatolactotroph GH3 cell line as a model system. Estradiol induced a slow proliferative response in the rat pituitary GH3 cell line; cell number increased by 50% after stimulation for 4 days, and was almost doubled after 7 days (Fig. 4A). Quantitative RT–PCR (Q-PCR) analysis confirmed induction of PRL mRNA in GH3 cells, by eightfold at 24 hours and 14-fold at 72 hours (Fig. 4B). The level of Wnt4 mRNA was unchanged at 24 hours, but was increased 2.5-fold at 72 hours (Fig. 4C), and the level of Wnt4 protein was increased at 72, 120 and 168 hours (Fig. 4D).
Estradiol does not induce canonical Wnt signaling in GH3 cells
Activation of the canonical Wnt signaling pathway was assessed using the TopFlash reporter gene, which displays a transcriptional response to activation of TCF and LEF by β-catenin, inducing luciferase expression. No induction of luciferase activity occurred in GH3 cells after treatment with either estradiol or Wnt4-conditioned medium. Lithium chloride and Wnt3a-conditioned medium, two well-characterized inducers of canonical Wnt signaling in many cell lines, were also unable to induce TCF-mediated gene expression in GH3 cells, but gave a robust induction in human embryonic kidney HEK-293 cells (27-fold and sevenfold, respectively; Fig. 5A). RT–PCR analysis of Fzd receptor expression confirmed that Fzd2, 4, 5 and 6 were all expressed in GH3 cells (data not shown), indicating that their lack of response to Wnt ligands was not related to the absence of receptor expression. Furthermore, GH3 cells cotransfected with plasmids encoding a constitutively active mutant β-catenin (mβ-Cat) and the TopFlash reporter showed no induction of luciferase, whereas HEK-293 cells treated under the same conditions exhibit a 14-fold induction of luciferase (Fig. 5B). Taken together, these results indicate that canonical Wnt signaling is not inducible in pituitary GH3 cells.
Translocation of β-catenin from the cytoplasm to the nucleus is a key feature of activation of the canonical Wnt pathway. Immunocytochemical analysis of GH3 cells and primary pituitary cells showed that β-catenin was predominantly at the cell membrane in unstimulated cells, with no apparent nuclear staining. Neither treatment with estradiol nor LiCl was able to induce nuclear translocation of β-catenin, and membrane staining remained unchanged, further indicating that canonical signaling was not activated in GH3 cells or in primary rat pituitary cell cultures (Fig. 6).
Wnt4 inhibits Ca2+ oscillations in GH3 cells
GH3 cells are known to have spontaneous oscillations in intracellular Ca2+ concentrations (Haymes and Hinkle, 1993). To assess whether Wnt4 signals through the Wnt–Ca2+ pathway, we carried out live-cell Ca2+ imaging on GH3 cells loaded with Fluo-4 and measured whether treatment with Wnt4 recombinant protein modulated the spontaneous Ca2+ oscillations (Fig. 7). The addition of Wnt4 partially or completely suppressed oscillations in intracellular Ca2+ in 63% of cells. By comparison, no change was observed in the oscillations following the addition of a control medium (Fig. 7E). Wnt4 induced the total inhibition of oscillations for long periods in some cells (Fig. 7A), whereas in other cells it induced a reduction in frequency and amplitude of oscillations (Fig. 7B), gave a temporary inhibition of oscillations (Fig. 7C) or had no effect (Fig. 7D). Immunoneutralization, by prior incubation with an anti-Wnt4 antibody, completely prevented the inhibition of the Ca2+ transient currents by Wnt4 (Fig. 7F). The proportion of the cells displaying complete suppression, partial inhibition or no effect is shown in Fig. 7G. Flow cytometry analysis indicated that 63% of GH3 cells in these conditions were in the G1 phase of the cell cycle, 36% were in S phase and 1% were in G2 or M phase (data not shown), suggesting a potential link between the Ca2+ response to Wnt4 and the cell cycle.
Here, we have provided evidence for the involvement of Wnt4 in the proliferative response of the pituitary gland to estrogen. We have shown that Wnt4 expression was increased upon estrogen treatment in pituitary tissue in vivo and that its expression is also induced by estrogen in the somatolactotroph GH3 cell line. Expression of Wnt4 in the pituitary was not restricted to lactotrophic cells, and Wnt4 action in pituitary cells did not involve the canonical β-catenin signaling pathway, implying that other pathways are likely to be involved.
Wnt4 is a growth factor involved in many developmental processes and is thought to have a role in fetal pituitary development (Treier et al., 1998; Potok et al., 2008; Brinkmeier et al., 2009). Wnt4 is expressed in the developing mouse pituitary between e9.5 and e14.5 and might regulate differentiation of ventral cell types (Potok et al., 2008; Treier et al., 1998), perhaps through interaction with Prop1 (Olson et al., 2006). Wnt4-deficient mice display pituitary hypoplasia, at least affecting the somatotroph and thyrotroph lineages (Potok et al., 2008); however, until now, very little information regarding Wnt4 in the adult rat pituitary has been available. In the present study, we found that the expression of Wnt4 in the adult rat was not confined to the lactotroph population but was readily detectable in all of the endocrine cell types in the pituitary. Estrogen treatment increased the proportion of lactotrophic cells in the pituitary (data not shown), as expected, but did not markedly alter the proportions of cell types that expressed Wnt4. Previous data from Miyakoshi and colleagues (Miyakoshi et al., 2009) suggested that Wnt4 expression was confined to the GH-producing cells and a minority of the TSH-producing cells in the rat pituitary. The reasons for the discrepancy between the Miyakoshi study and the present study are not clear, but we found that Wnt4 expression was seen in over 90% of somatotrophic cells. Other reasons for this difference could relate to the different rat strains used (Fischer 344 in the present study and Sprague Dawley in the Miyakoshi study), or the antibodies used for immunohistochemistry; the antibody used in the previous report was raised against a 14-amino-acid oligopeptide, whereas our antibody was produced against a 100-residue peptide. Antibody specificity was confirmed in the present study by checking for the specific staining in renal tubules, the lack of non-specific staining in the absence of primary antibody and by establishing that a protein of the appropriate size was detected by western blotting.
Wnt4 was recently found to be highly expressed in several human pituitary adenoma types, including lactotroph, somatotroph and thyrotroph adenomas (Miyakoshi et al., 2008). The authors of that report speculated that Wnt4 might be involved in proliferation of those cell lineages, and the present findings support the general hypothesis that Wnt4 is involved in plasticity of function and structure in the adult pituitary gland. In addition, a recent study found that expression of WIF1, an extracellular inhibitor of Wnt signaling, was significantly reduced in pituitary adenomas, and that its overexpression reduced GH3 cell proliferation, further implicating Wnt signaling in pituitary growth (Elston et al., 2008).
The signaling pathway involved in Wnt action in the pituitary was previously unclear, but here we have shown that canonical signaling was not activated in the rat pituitary GH3 cell line, either by estrogen or by Wnt4 or Wnt3a. Furthermore, a constitutively active mutant β-catenin protein, which markedly activated TCF- and LEF-dependent transcriptional signaling in HEK-293 cells, had no effect in pituitary GH3 cells. We have found no evidence for nuclear localization of β-catenin in GH3 cells, in primary cultures of rat pituitary cells or in intact rat pituitary tissue (data not shown), and similarly no evidence has been found for nuclear expression of β-catenin in those pituitary adenomas that displayed Wnt4 overexpression (Miyakoshi et al., 2008).
Non-canonical Wnt signaling pathways remain less well defined than the canonical cascade, but in different systems have been found to target Ca2+ signaling or kinase pathways that might affect PCP (Rao and Kuhl, 2010). One target of the Wnt–PCP pathway is E-cadherin, through which Wnt ligands might alter cell–cell adhesion in order to control the orientation and development of a number of organs (Fanto and McNeill, 2004). The cellular patterns of cadherin expression in the pituitary have recently been identified (Chauvet et al., 2009), and the interaction of β-catenin and E-cadherin at the cell membrane might play an important role in the development and plasticity of cell networks. We show here, for the first time, that Wnt4 has an impact upon non-canonical signaling in pituitary cells by inhibiting spontaneous Ca2+ oscillations. The mechanism by which this inhibition occurs has not been studied but might well involve inactivation of CRAC channels (Gwack et al., 2007), although further study will need to be performed to elucidate the details of this signaling pathway in the pituitary. The Wnt–Ca2+ pathway has recently been implicated in the progression of a number of cancers, where it has both suppressive and inductive properties (McDonald and Silver, 2009), and, therefore, study into the effects of the non-canonical actions of Wnt4 in the pituitary is likely to throw new light on the nature of adult pituitary remodeling and adenoma formation.
Materials and Methods
Female Fischer 344 rats were implanted with slow-release subcutaneous pellets (containing 10 mg of DES, n=5, or placebo, n=5; Innovation Research, Novi, MI) for 3 weeks. Animal experiments were performed according to UK Home Office guidelines and rats were killed by a schedule 1 method. Pituitary glands were harvested, then washed with ice-cold PBS and snap-frozen on dry ice. RNA was extracted using the Qiagen RNeasy mini kit according to the manufacturer's instructions. Expression profiling was performed using an Affymetrix gene chip rat genome 2.0 array (no. 230; three chips per group with individual animals for each chip). Background correction, quantile normalization and gene expression analysis were performed using the robust multichip average (RMA) function in Bioconductor (Bolstad et al., 2003).
GH3 cells and HEK-293 cells (ATCC, Rockville, MD) were grown at 37°C under a 5% CO2 atmosphere, in Phenol-Red-free Dulbecco's modified Eagle's medium (DMEM) supplemented with 1 g of glucose/l, 10% fetal bovine serum (FBS) and 1% Glutamax (all Gibco). Two stably transfected cell lines were used to obtain conditioned media: LM (TK−) cells transfected with a Wnt3a expression vector (ATCC) and NIH-3T3 cells transfected with a Wnt4 expression vector (kindly donated by Andreas Kispert, Institute of Molecular Biology, Hannover Medical School, Hannover, Germany). The conditioned media were generated as previously described (Willert et al., 2003).
For the verification of microarray results, RNA from all five animals from each group was pooled for quantitative real-time PCR (Q-PCR) analysis. For cell culture timecourse experiments, GH3 cells were grown in 25 cm2 flasks (~5×106 cells), serum-starved for 24 hours and then stimulated with 10 nM E2 (Sigma) for the indicated times. Cells were harvested by trypsinization and washed twice in ice-cold PBS. Total RNA was isolated using the Qiagen RNeasy kit according to the manufacturer's instructions.
cDNA was generated with the Omniscript RT system (Qiagen). Q-PCR was performed using the Stratagene Mx3000 P thermocycler (Stratagene) and the SYBRgreen Jump Start Taq ready mix (Sigma). Cyclophilin was used as the housekeeping gene for normalization. The following primers were used: cyclophilin, 5′-TTTTCGCCGCTTGCTGCAGAC-3′ and 5′-CACCCTGGCACATGAATCCTGGA-3′; PRL, 5′-AGCCAAGTGTCAGCCCGGAAAG-3′ and 5′-TGGCCTTGGCAATAAACTCACGA-3′; Wnt4, 5′-acgaggggccgcacttgcaaca-3′ and 5′-ggcacaggcgggcacggtcac-3′; Cpz, 5′-ccccagggcgcgtaggcagc-3′ and 5′-cccgggccgtggaggtggacat-3′; and Wnt10a, 5′-ccagcttcagtgcattgcccaaca-3′ and 5′-agctcaggcaggtgggggtggtag-3′.
GH3 cell proliferation assay
GH3 cells were plated at a density of 1×104 cells per well in 96-well plates. They were starved for 24 hours in 0.25% BSA, before stimulation at defined timepoints. The cell number was measured using the CellTiter 96 AQueous one solution cell proliferation assay (MTS; Promega) according to the manufacturer's instructions.
Primary antibodies used were against: Wnt4 (rabbit Ig, 1:1000 for western blot and 1:50 for immunohistochemistry; SDI, Newark, Germany), α-tubulin (mouse Ig, 1:25000; Abcam, Cambridge UK), GH (goat Ig, 1:50; R&D Systems, Abingdon, UK), PRL (mouse Ig, 1:4000, Pierce, Rockford, IL), TSH (guinea-pig Ig, 1:100; NIDDK, Bethesda, MD), ACTH (mouse Ig, 1:200; Novocastra, Milton Keynes, UK) or LH (mouse Ig, 1:1000; kindly provided by Janet Roser, University of California-Davis, Davis, CA), R51 PRL (rabbit Ig, 1:500, kindly donated by Alan McNeilly, MRC Human Reproductive Sciences Unit, Edinburgh, UK) and β-catenin (mouse Ig, 1:400; BD Transduction Laboratories, Oxford, UK).
Secondary antibodies were: anti-(rabbit Ig)–HRP-conjugated (donkey Ig, 1:2000) and anti-(mouse Ig)–HRP (donkey Ig, 1:25000) (both from Santa Cruz Biotechnology), anti-(rabbit Ig)–Alexa-Fluor-546 (donkey Ig, 1:500, Invitrogen), anti-(mouse Ig)–Alexa-Fluor-488 (donkey Ig, 1:1000, Invitrogen), anti-(guinea-pig Ig)–FITC (goat Ig, 1:64, Sigma), anti-(goat Ig)–FITC (donkey Ig, 1:500, Santa Cruz Biotechnology) and anti-(rabbit Ig)–Texas-Red (donkey 1:500, Santa Cruz Biotechnology).
Western blotting analysis
GH3 cells were plated at ~7.5×105 cells per well in six-well plates and serum-starved in 10% dextran-charcoal-treated FBS (DCT-FBS) (Perbio Scientific, Cramlington, UK) for 24 hours before stimulation with either control (DCT-FBS) or E2 (10 nM, Sigma). Cells were washed twice with ice-cold PBS and lysed using RIPA buffer with Complete mini EDTA-free protease inhibitors (Roche). Cells were scraped, agitated for 30 minutes on a rocker and centrifuged for 10 minutes at 12,000 g at 4°C, and the supernatant was collected and stored at −80°C. Samples were subjected to SDS-PAGE (10% gels), before transfer onto nitrocellulose membrane. Primary antibodies were applied overnight at 4°C, then secondary antibodies were applied for 1 hour at room temperature. Staining was detected with EZ-ECL (Pierce) using Kodak Biomax XAR film.
Female Fischer 344 rats were killed by a schedule 1 method and pituitary glands were removed for either wax-embedding or dissociation. Pituitary glands were fixed for 2 hours in 4% PFA before wax-embedding and sectioning at a thickness of 5 μm. Deparaffinization of wax-embedded slices was conducted in xylene, and sections were subsequently re-hydrated in decreasing concentrations of ethanol. Antigen retrieval was performed by boiling in 10 mM sodium citrate for 20 minutes. Pituitary cells were dissociated as previously described (Sartor et al., 2004), and cultures of primary pituitary cells or GH3 cells were plated on poly-(L-lysine)-coated glass coverslips and stimulated as indicated, before fixation in 4% PFA.
Cells or tissue were blocked in 5% donkey serum in PBS for 1 hour, and incubated with the first primary antibody overnight at 4°C followed by the secondary antibody for 2 hours at room temperature. Samples were blocked again in 5% donkey serum, before an overnight incubation with the second primary antibody at 4°C, followed by the secondary antibody for 2 hours at room temperature. Samples were treated with DAPI (0.1 μg/ml, Sigma) for 20 minutes at room temperature, and then mounted for analysis in Permafluor (Thermo Scientific). Images were collected using a Nikon C1 confocal microscope (Bioimaging Facility, Faculty of Life Sciences, University of Manchester, Manchester, UK).
Reporter gene assays
GH3 and HEK-293 cells were plated in sterile white 96-well plates at ~1×104 cells per well and ~5×103 cells per well, respectively. Cells were transfected using Fugene 6 transfection reagent (Roche) according to the manufacturer's instructions. Cells were transfected with either Super 8× TopFlash, Super 8× FopFlash negative control (both at 0.1 μg per well; Addgene, Cambridge, MA) or a vector encoding mβ-Cat [0.1 μg per well; a kind gift from Hans Clevers (Hubrecht Institute, Utrecht, The Netherlands) (Morin et al., 1997)] with pRL-TK Renilla (Promega, 0.1 μg per well for GH3 cells and 0.01 μg per well for HEK-293 cells). Fugene:DNA ratios were 3:1 for GH3 cells and 6:1 for HEK-293 cells. Cells were left for 24 hours, then stimulated as specified. Luminescence was measured using the Dual-Glo luciferase assay system (Promega) according to the manufacturer's instructions.
Live-cell Ca2+ imaging
GH3 cells were seeded in glass-bottomed dishes (Iwaki) and left to settle for 24 hours. Cells were loaded with Fluo-4 (Invitrogen) for 30 minutes and then the dish was transferred onto the stage of a Zeiss Axiovert 200 microscope with an attached XL incubator (at 37°C and under a humid 5% CO2 atmosphere). Cells were either stimulated with vehicle (DCT-FBS) or Wnt4 recombinant protein (Novus Biologicals, Littleton, CO). For the immunoneutralization experiment, Wnt4 protein was incubated with Wnt4 antibody for 1 hour at room temperature, before addition to cells in a 10:1 antibody:protein molar ratio. Images were taken every 3 seconds, areas of interest were drawn around cells and mean intensity throughout the experiment was calculated using Kinetic Imaging AQM6 software (Andor, Belfast, UK).
We thank the Wellcome Trust, UK for funding this work (project grant 082794, to J.R.E.D. and G.B.). A.G. was supported by a PhD studentship from the Biotechnology and Biological Sciences Research Council (BBSRC), UK. C.V.H. was supported by the BBSRC (BBF0059381). The Centre for Cell Imaging has been supported through BBSRC REI (grant BBE0129651), and Hamamatsu Photonics and Carl Zeiss Limited provided technical support. We are grateful to Dave Spiller for assistance in the Centre for Cell Imaging, University of Liverpool, and to Leo Zeef, University of Manchester Genomic Technology Facility, for valuable help in the analysis of the array data. The Developmental Biomedicine Research Group is supported by the Manchester Academic Health Sciences Centre (MAHSC) and the NIHR Manchester Biomedical Research Centre. Deposited in PMC for release after 6 months.