Thyrotropin-releasing Hormone and Epidermal Growth Factor Regulate Iron-regulatory Protein Binding in Pituitary Cells via Protein Kinase C-dependent and -independent Signaling Pathways*

Intracellular iron homeostasis is regulated, in part, by interactions between iron-regulatory proteins (IRP1 and IRP2) and iron-responsive elements (IREs) in ferritin and transferrin receptor mRNAs. In addition to iron, cellular oxidative stress induced by H2O2, nitric oxide, and hypoxia, and hormonal activation by thyroid hormone and erythropoeitin have each been shown to regulate IRP binding to IREs. Hormonal signals, in particular mediated through protein kinase C (PKC), play a central role in the modulation of IRP/IRE interactions since phorbol esters were shown to activate IRP binding (Eisenstein, R. S., Tuazon, P. T., Schalinske, K. L., Anderson, S. A., and Traugh, J. A. (1993) J. Biol. Chem. 268, 27363–27370). In pituitary thyrotrophs (TtT97), we found that thyrotropin releasing hormone (TRH) and epidermal growth factor (EGF) increased IRP binding to a ferritin IRE, dependent on PKC and mitogen-activated protein kinase (MAPK) activity. In contrast, TRH and EGF decreased IRP binding in pituitary lactotrophs (GH3), despite activation of PKC and MAPK. IRP1 and IRP2 levels remained constant and IRP2 binding was predominant throughout. TRH and EGF markedly decreased IRP binding in MAPK kinase inhibitor-treated GH3 cells, whereas, they increased IRP binding in phosphatase inhibitor-treated GH3 cells. IRE-dependent CAT reporter translational expression closely reflected IRP binding to the ferritin IRE in both GH3 and TtT97 cells. Interestingly, ferritin protein levels were regulated similarly by TRH in both cell lines. These data link two different cell receptor systems to common signaling pathways that regulate IRP binding and ferritin expression. Remarkably, for TRH and EGF, these effects may be PKC-dependent or -independent determined by the cell type.

Several factors have been shown to regulate ferritin and TfR gene expression at both transcriptional and post-transcriptional levels (5,6). A number of physiologically relevant hormones and cytokines modulate the ferritin transcription rate. For example, thyrotropin (thyroid-stimulating hormone (TSH)) increases H-ferritin mRNA and protein levels in rat thyroid cells due to increased transcription without a change in translation rate (13)(14)(15)(16). In human myoblasts and fibroblasts, tumor necrosis factor ␣ and interleukin-1␣ increase transcription of H-ferritin mRNA, but not L-ferritin mRNA, which increases ferritin protein levels also without a change in translation. In this system, however, once the supply of iron becomes limiting, translation rates decrease, leading to a reduction in ferritin synthesis (17,18). Phorbol ester (PMA) treatment of human promyelocytic HL60 cells also regulates H-and L-ferritin mRNA levels, in part, at the transcriptional level (19 -21).
In contrast, in human hepatoma cells (HepG2), interleukin-1␤ increases both H-and L-ferritin protein levels by translational enhancement of these mRNAs through the 5Ј-UTR acute box elements downstream of the IRE, without changing mRNA levels (22)(23)(24). Thyroid hormone induces a similar response in HepG2 cells, but this is associated with a decrease in IRP1 and IRP2 binding to the ferritin IRE (25). In erythroid cells, erythropoeitin regulates the stability of TfR mRNA via interactions between the IRP and IREs (26). In addition, protein kinase C (PKC) activation via PMA increases phosphorylation of IRP1 and binding of IRP1 to the L-ferritin IRE (27). Subsequently, it was shown that activation of PKC by PMA led to the direct phosphorylation of IRP1, which increased its bind-ing to a ferritin IRE, and also resulted in a functional increase of in vivo TfR mRNA levels (28,29). This was also proven to be the case for IRP2 (28). These studies established an unequivocal role for IRP phosphorylation in the non-iron-induced modulation of IRP binding activity and the regulation of expression of IRE-containing genes, including ferritin.
As an active secretory organ, the pituitary gland is critically dependent upon a regulated supply of iron to meet its metabolic requirements. However, to date, little is known of the factors and mechanisms associated with this process. The anterior pituitary gland secretes a diverse set of hormones, including TSH and prolactin (PRL), in response to specific hypothalamic peptides (e.g. thyrotropin-releasing hormone (TRH)) and growth factors (e.g. epidermal growth factor (EGF)). TRH regulates the synthesis and secretion of TSH and PRL by thyrotroph and lactotroph cells, respectively (30). In addition, TRH regulates proliferation of specific pituitary cells (31,32) and has been implicated in the formation of pituitary adenomas (33). TRH binds and activates two G protein-coupled receptors, TRHR1 and TRHR2 (34), which transduce signals predominantly via the PKC and mitogen-activated protein kinase (MAPK) pathways (35)(36)(37). Thus, TRH plays a central role in regulating the metabolism of specific anterior pituitary cells. This is particularly relevant given that excess iron accumulation in the anterior pituitary in the iron storage disease, hereditary hemochromatosis, results in hormone insufficiency and pituitary failure (38,39). These data emphasize the need to elucidate the pathways involved in regulating iron homeostasis in the pituitary.
Interestingly, similar to TRH, EGF increases PRL and TSH secretion in pituitary cells and also regulates pituitary cell proliferation (40,41). EGF is a potent mitogen that binds and activates the EGF receptor (EGFR), a type 1 tyrosine kinase receptor (42). The EGFR is overexpressed in lactotroph, thyrotroph, and gonadotroph tumor-derived cells (43), and may play a role in the development of pituitary tumorigenesis (43,44). Binding of EGF to the EGFR activates many signaling pathways, including MAPK and PKC (45,46). Other important links between the EGFR and TRHR pathways have recently been established. In thyrotrophs and lactotrophs, binding of TRH transactivates EGFRs, leading to a reduction in EGFR number and mRNA (47). In addition, stimulation of lactotrophs and thyrotrophs by EGF decreases TRHR number and mRNA levels (30,44). Based on the above, we reasoned that TRH and EGF may play an important role in the control of iron homeostasis in the pituitary.
To investigate the role of TRH and EGF in the regulation of iron homeostasis in the pituitary, we examined the effect of each ligand on the regulation of IRP binding to a ferritin IRE in two different pituitary cell lineages. Interestingly, we found that IRP2 was the predominant ferritin-IRE binding IRP. TRH and EGF up-regulated binding of IRP1 and IRP2 to a ferritin IRE in thyrotrophs (TtT97 cells), associated with reduced translation of a ferritin-IRE reporter and ferritin synthesis in whole cells. In marked contrast, TRH and EGF reduced IRP binding in lactotrophs (GH3 cells) despite an associated reduction in ferritin protein levels. Remarkably, we found that the effect was PKC-dependent in thyrotrophs, but not in lactotrophs. In addition, MAPK and phosphatases were found to play an integral role in the regulation of IRP binding activity in lactotrophs. These data provide novel links between two different cell receptor systems, common signaling pathways, regulated IRP binding and ferritin expression.

EXPERIMENTAL PROCEDURES
Cell Culture, Transient Transfections, and DNA Constructs-Rat lactotroph GH3 (ATCC CCL 81) cells and human hepatoma HepG2 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum. Murine thyrotrope TtT97 tumors were grown in thyroid-ablated LAF1 mice (48). After tumor removal from the mice, cells were plated at a density of 10 7 cells/ml in DMEM supplemented with 10% fetal calf serum (48). Transient transfections were performed on TtT97, GH3, and HepG2 cells with FuGENE (Roche Molecular Biochemicals). The TRH receptor (TRHR) construct, pCDM8wtTRHR, contains the full-length cDNA as described (49). HIRE-CAT contains the human ferritin IRE upstream of the CAT coding region as described (25) and RSVLUC is as described (50). pIRE contains a conserved ferritin IRE (25) cloned into the BamHI and HindIII sites of pBlueScript(KS) in the T7 sense orientation.
Western Analysis-For Western analysis, proteins were separated by 6% SDS-PAGE, transferred to polyvinylidene difluoride membrane (Roche), which was incubated with 20 g of IRP1 and/or 20 g of IRP2 immune serum (1:2000) (51) before detection of IRPs using ECL-Plus (Amersham Pharmacia Biotech) detection.
Preparation of RNA Transcripts-Linearized templates of pIRE and pBlueScript(KS) (HindIII digests) were used with T7 polymerase (Life Technologies, Inc.) in transcription reactions containing [ 32 P]UTP (3000 Ci/mmol; Amersham Pharmacia Biotech) as described (25) to produce transcripts with a specific activity of approximately 5 ϫ 10 9 cpm/g RNA. Full-length transcripts were purified as described (52).
RNA Electrophoretic Mobility Shift Assay (REMSA), Supershift Assay, and UV Cross-link Analysis-REMSAs were performed as described previously (52), RNA-protein complexes detected by phosphorimaging (PhosphorImager, Molecular Dynamics) and quantified by ImageQuant (Molecular Dynamics) analysis. Supershift assays were performed with 10 g of IRP1 immune mouse serum or IRP2 immune rabbit serum added after heparin for a 30-min incubation prior to electrophoresis. For UV cross-link analysis, the protocol used was as described (52), using 7.5% SDS-PAGE gels, prior to phosphorimaging.
Chloramphenicol Acetyltransferase (CAT) and Luciferase (LUC) Enzyme Assays-Cells were grown in 10-cm 2 dishes, transfected and treated in triplicate, washed in phosphate-buffered saline, and resuspended in 1.5 ml of phosphate-buffered saline, centrifuged at 2,000 ϫ g for 2 min at 4°C and lysed in 200 l of LUC buffer (25 mM glycyl glycine, 15 mM magnesium sulfate). LUC assays were performed in a Berthold AutoLumat LB953 that added 60 g/ml beetle luciferin (Promega) and 15 g/ml ATP in LUC buffer. CAT assays were performed as described previously (25).
RNase Protection Analysis-CAT and LUC mRNAs from transfected cells were characterized by RNase protection as described (25), using 5 ϫ 10 8 cpm/g SP6 transcribed 33 P-radiolabeled CAT and LUC transcripts (25). The full-length probes, which also contain plasmid leader sequence, were 261 and 460 nucleotides for CAT and LUC, respectively. The protected cRNAs were 250 and 350 nucleotides for CAT and LUC, respectively. The protected cRNAs were separated by electrophoresis on a 6% acrylamide, 8 M urea gel and visualized by phosphorimaging.

Pituitary Cells Contain IRP1 and IRP2-REMSA and UV
cross-linking analysis were used to identify two IRPs, IRP1 and IRP2, in GH3 (rat pituitary lactotrophs) ( Fig. 1) and TtT97 (mouse primary pituitary thyrotrophs) (data not shown) cells. IRP1 and IRP2 from these cells bound specifically to a ferritin IRE in REMSA (Fig. 1, lane 1) and in UV cross-link assay (Fig.  1, lane 4), where the identity of each RNA-protein complex was confirmed by supershift with IRP1-and IRP2-specific antibodies ( Fig. 1, lanes 2 and 3). In both cell types, the majority of binding to the IRE was by IRP2. However, as demonstrated for other cell types (25), the binding affinity of IRP1 could be increased and that of IRP2 decreased by the addition of ␤-mercaptoethanol to GH3 and TtT97 cell extracts (data not shown). This indicated that the binding affinity of IRPs in these pituitary cells could be altered without changes in IRP levels.
TRH Divergently Modulates IRP Binding in Thyrotrophs and Lactotrophs-TRH binds with high affinity to TRH-Rs on lactotrophs and thyrotrophs (34,36,53) to stimulate secretion of prolactin and TSH, respectively. Thus, these two cell types provide an ideal comparative model to investigate the effects of TRH on IRP binding activity and the posttranscriptional regulation of iron homeostasis in the pituitary. In GH3 cells, TRH (100 nM) decreased IRP-IRE complexes (predominantly IRP2) ϳ2-fold within 30 min ( Fig. 2A). This was sustained at 2 h. In marked contrast, in TtT97 cells IRP2-and IRP1-IRE complex formation increased ϳ2-fold within 30 min of treatment with TRH (100 nM) ( Fig. 2A). This was more obvious after 2 h. Changes in IRP1-IRE complex mirrored that of IRP2-IRE complex formation in both cell types, although the IRP1-IRE complex was always represented at lower levels. Although IRP1 and IRP2 were abundant in GH3 cells (Fig. 2B), IRP2 formed the predominant IRE-IRP complex ( Fig. 2A) (5% IRP1:95% IRP2). The ratio of IRP2-IRE to IRP1-IRE in GH3 cells was subject to minor change dependent on treatment conditions (see below), but the IRP2-IRE complex was always predominant. IRP2 binding was not as predominant in TtT97 cells, where the binding complex ratio was 40% IRP1:60% IRP2 ( Fig.  2A). As for GH3 cells, TtT97 cells contained abundant amounts of IRP1 and IRP2 protein (data not shown). To determine the effect of TRH on IRP1 and IRP2 protein levels, we performed a Western assay. As can be seen in Fig. 2B, TRH treatment did not alter IRP1 or IRP2 protein levels over a 2-h time course. No changes in IRP levels were observed with any of the treatments in all experiments performed (see below). Taken together, these data indicate that TRH is a novel modulator of IRP2 and IRP1 binding activity, and TRH-induced changes in complex formation were not due to changes in IRP1 and IRP2 protein levels.
TRH Modulates Ferritin Levels in GH3 and TtT97 Cells-Given this cell-specific divergent effect of TRH on IRP binding, we analyzed the effect of TRH on ferritin expression in each cell line. Remarkably, we found that TRH rapidly decreased heavy (H) and light (L) chain ferritin levels in both TtT97 and GH3 cells within 2 h of treatment (Fig. 3A). Indeed, there was almost complete reduction of the H-ferritin subunit in GH3 cells (Fig.  3A), whereas the effect in TtT97 cells was less marked (Fig.  3A). The down-regulation of ferritin in TtT97 cells was consistent with the increased IRP binding we had observed in REMSA. However, in GH3 cells, the down-regulation of ferritin levels, despite decreased IRP binding, suggested that transcriptional repression, and/or translational repression via the ferritin acute box (22)(23)(24), and/or increased ferritin dissociation (54) were the major determinants of ferritin levels. These data indicate that TRH may regulate pituitary cell iron homeostasis through different mechanisms in a cell-specific manner.
Regulation of IRE-dependent translation by TRH Is TRH Receptor-dependent-To determine whether the regulation of IRP binding by TRH was dependent upon TRH-Rs, we used HepG2 liver cells that are devoid of TRH-Rs for transfections. HIRE-CAT was used as a reporter because it reflects the degree of IRP binding to the 5Ј-UTR IRE in the translation construct (23). When HepG2 cells were transfected with HIRE-CAT without the TRH-R, no reporter activity was observed (Fig. 3B). However, when the TRH-R was cotransfected with the HIRE-CAT, TRH induced a time-dependent increase in CAT-activity (Fig. 3B). Similar experiments were performed in GH3 and TtT97 cells. TRH induced CAT activity in GH3 cells in a timedependent manner, but decreased it in TtT97 cells (Fig. 3B). This is consistent with the differential regulation of IRP binding that we had observed above. We next performed RNase protection assays with [ 33 P]CAT and [ 33 P]LUC probes in the cells transfected with HIRE-CAT and RSVLUC. TRH did not alter CAT or control LUC mRNA levels compared with control cells (Fig. 3C) 1. Pituitary cells contain IRP1 and IRP2. GH3 cell extract (10 g) was incubated with 5 ϫ 10 4 counts (10 pg of RNA) 32 P-labeled ferritin IRE for 30 min prior to RNase T1 and heparin (52). IRP-specific isoform antibodies (Ab) (10 g of IRP1-Ab and IRP2-Ab) were added to lanes 2 and 3, respectively, for 30 min after heparin. IRP-IRE complexes were separated by REMSA (n ϭ 3 independent assays) and visualized by phosphorimaging. For UV cross-linking (lane 4), 20 g of GH3 cell extract was incubated with 20 pg of the IRE probe, samples subjected to UV irradiation for 10 min, RNase digestion, and SDS-PAGE (n ϭ 3 independent assays) (see "Experimental Procedures"), and complexes visualized by phosphorimaging. FIG. 2. TRH modulates IRP1 and IRP2 binding to a ferritin IRE in GH3 and TtT97 cells. A, REMSA was performed as described for Fig. 1 with GH3 and TtT97 cell extracts Ϯ TRH (100 nM) (0, 30, 120 min) with the ferritin IRE probe. Results are represented as -fold changes in IRP2-IRE relative to control, and are the means of triplicates Ϯ standard error (S.E.) (n ϭ 5 independent assays). Variation ranged from 5% to 12% in all REMSAs performed. B, 6% SDS-PAGE was performed with GH3 cell extract Ϯ TRH (100 nM). Proteins were transferred to polyvinylidene difluoride membrane, incubated with IRP1-Ab and IRP2-Ab, and visualized by ECL (performed in duplicate; n ϭ 3 independent assays). Variation in replicates was 0 -10%. an important role in the pituitary by regulating PRL and TSH secretion, cell proliferation, as well as TRHR number and mRNA levels (44,47). Indeed, the EGFR and the TRHR activate common pathways. In this context, we investigated the effect of EGF treatment on IRP1 and IRP2 binding profiles in GH3 and TtT97 cells. Interestingly, the effect of EGF on IRP-IRE interactions mimicked the effect seen with TRH. EGF down-regulated IRP binding 2-fold within 30 min in GH3 cells (Fig. 4), which was more pronounced at 2 h, but up-regulated IRP binding in TtT97 cells (Fig. 4). These data suggest a convergence of signaling pathways utilized by each ligand to regulate IRP1 and IRP2 binding in pituitary cells.
PKC Modulates IRP Binding Activity in Pituitary Cells-One of the major signaling cascades that is activated by TRH and EGF is the PKC pathway (35,53). In GH3 cells, TRH has been shown to activate specific isoforms of PKC (␣, ␤, ␥, ␦, ⑀) (35,53). To further define pathways involved for each ligand, GH3 and TtT97 cells were treated with either TRH (100 nM), EGF (100 nM), or phorbol esters (PMA, 100 nM). In preliminary studies, we found that TRH, EGF, and PMA mobilized and activated all the PKC isoforms tested (␣, ␤, ␥, ⑀, ␦) (peak at 5 min for GH3 cells; peak at 10 min in TtT97 cells; data not shown). To determine the effect of PKC activation on IRP binding, we treated several different cell lines with PMA (100 nM) and analyzed them in REMSA. PMA increased IRP-IRE binding by 2-fold in HL60 myeloid leukemic cells (Fig. 5A), consistent with the results of Schalinske and Eisenstein (28), which demonstrated this to be a result of increased IRP phosphorylation by PKC. Interestingly, PMA increased IRP-IRE binding by 2-fold in both GH3 (Fig. 5A) and TtT97 cells (Fig. 5A). This contrasted with the TRH-induced reduction in IRP binding in GH3 cells ( Fig. 2A).
To investigate further the role of PKC, cells were treated with either a PKC inhibitor peptide (peptide 19 -31, 100 nM) or PMA (100 nM) for 24 h before activation by TRH. For GH3 cells, TRH still down-regulated IRP1 and IRP2 binding ϳ2-fold in the presence of the 19 -31 peptide. Interestingly, TRH downregulated IRP2 binding, but did not greatly affect IRP1 binding, in GH3 cells pretreated with PMA for 24 h (Fig. 5B). In TtT97 cells, pretreatment with the 19 -31 peptide or PMA abolished the TRH-induced up-regulation of IRP binding (Fig.  5B), but did not alter the relative IRP1:IRP2 binding ratio. Addition of PMA to 19 -31-treated GH3 and TtT97 cells did not alter IRP1 and IRP2 binding, indicating that PKC had been inactivated in these cells (data not shown). These findings suggest that TRH-induced modulation of IRP-IRE binding is PKC-dependent in TtT97 cells, and PKC-independent in GH3 cells.

Substantial Involvement of MAPK in the Modulation of IRP-IRE Binding in Pituitary
Cells-Recent studies have demonstrated that TRH mediates some of its effects in GH3 cells through PKC-dependent and -independent activation of mitogen activated protein kinases (MAPKs, p42, p44) (35,53). This led us to analyze GH3 and TtT97 cells treated with TRH, EGF, and PMA for phosphorylation and activation of MAPKs. In preliminary experiments, we found that TRH (100 nM) and EGF (4 nM) induced MAPK activation (p42, p44) which peaked at 5 min in GH3 cells and at 15 min in TtT97 cells (data not shown). In the following experiments, Western blot demonstrated that there were no changes in the level of IRP1 and IRP2 by any of the treatments in GH3 cells over the 2-h time course (Fig. 6B). GH3 cells were treated with the MAPK kinase (MAPKK) inhibitor (PD098059, 40 M), which blocks the phosphorylation and activation of MAPK. This reduced IRP-IRE FIG. 4. EGF regulates IRP binding activity in GH3 and TtT97 cells. REMSAs were performed as described for Fig. 1 using the ferritin riboprobe with GH3 and TtT97 cell extracts treated with EGF (4 nM) (0, 5, 30, and 120 min). Experiments were performed in triplicate (n ϭ 5 independent assays). Results are expressed as CAT activity after normalizing with luciferase activity, and are the means of triplicates Ϯ S.E.; n ϭ 3 independent assays (variation of 5-15%). C, 20 g of total RNA from GH3-and TtT97-transfected cells and 100 ng of RSVLUC and RSVCAT mRNA was hybridized with 33 P-labeled RSVLUC (410 nucleotides) and RSVCAT (261 nucleotides) probes. S1 nucleaseprotected double-stranded RNAs were separated on 6% acrylamide, 8 M urea gels and detected by phosphorimaging. Control RSVLUC and CAT complexes are larger due to leader sequence that is unprotected in RNA from GH3 and TtT97 transfectants. Variation in replicates was 3-7%. Experiments were performed in triplicate (n ϭ 2 independent assays).
TRH, EGF, and PMA Increase IRP-IRE Binding in Protein Phosphatase-inhibited (Pinh) Cells-Changes in IRP phosphorylation status can modulate IRP binding activity to an IRE (28,29). The data above support this concept and are consistent with ligand-induced activation of the PKC and MAPK pathways in the pituitary. In contrast, treatment of cells with phosphatase inhibitors (e.g. okadaic acid) has been shown previously to modify the effects of NO and H 2 O 2 on IRP binding activity (55,56). To investigate the potential role of phosphatases in the regulation of IRP binding activity by TRH and EGF, GH3 cells were treated with two phosphatase inhibitors, okadaic acid (5 M) (that inhibits protein phosphatase 1 (PP1) and PP2A (Ref. 57)), and sodium pervanadate (5 mM) (that inhibits protein tyrosine phosphatase (PTB) (Ref. 57)). In contrast to the down-regulation observed with TRH or EGF alone (Fig. 7,  lanes 1-3 and 13-15, respectively), TRH and EGF increased IRP-IRE binding in Pinh-treated cells (Fig. 7, lanes 4 -6 and  16 -18, respectively). However, PMA induced a 5-fold increase in Pinh cells (Fig. 7, lanes 10 -12), compared with a 2-fold increase in non-Pinh-treated cells (Fig. 7, lanes 7-9). These data suggest that phosphatases play an integral role in the regulation of IRP-IRE binding by TRH and EGF in GH3 cells. DISCUSSION This report demonstrates that the neuropeptide, TRH, and growth factor, EGF, activate signaling pathways that directly modulate IRP binding activity and regulate IRE-dependent gene expression. This observation adds to the known changes in IRP function induced by iron influx and redox change (58), nitric oxide (51,55), intra-and extracellular oxidative stress (H 2 O 2 ) (55, 59), hypoxia (60,61), phorbol esters (PMA) (27)(28)(29), erythropoeitin (26), and thyroid hormone (25). We developed thyrotroph and lactotroph pituitary cell models to examine the signal transduction pathways that may phosphorylate IRPs and the resultant downstream effects on ferritin gene expression.
TRH down-regulated H-and L-ferritin protein expression similarly in both cell lines. However, this was not a reflection of similar effects of TRH on IRP binding to the ferritin IRE. TRH signaling caused down-regulation of both IRP1 and IRP2 binding in GH3 cells, but up-regulation in TtT97 cells. The TRHinduced changes in IRP2-and IRP1-IRE complex formation were not associated with changes in IRP levels. Our transfection data showed that the TRH-induced modulation of IRP-IRE binding closely reflected the changes in IRE-driven CAT gene expression in both TtT97 and GH3 cells. The studies in TtT97 cells, in which TRH increased IRP binding and decreased ferritin-CAT translation, are consistent with the IRE-dependent down-regulation of ferritin protein levels in response to TRH. However, other regulatory events must override the effect of IRP/IRE interactions to control steady-state ferritin levels in GH3 cells. This effect could result from changes in transcription rate, or be due to altered translation by other 5Ј-UTR elements in ferritin mRNAs (23), or even be the result of increased ferritin dissociation (54). The ferritin response in GH3 cells is likely due to one or a combination of these alternate pathways and is currently under investigation. The cellspecific and divergent responses of TRH on IRP binding imply that the regulation of iron homeostasis involves a greater level of complexity than that afforded by the IRP/IRE interaction alone in pituitary cells.
The similarity of effect on IRP binding between TRH and EGF suggested that both ligands were converging on a common signaling pathway. Schalinske and Eisenstein (28) have shown previously that activation of PKC by PMA leads to phosphorylation of IRP1 and IRP2 in HL-60 cells, which increases their binding affinity to an IRE, as reflected in increased TfR mRNA stability. Indeed, differences in IRP2:IRP1 binding affinity and response to stimuli may reflect differences in basal IRP2 and IRP1 phosphorylation status. In HL60 cells, IRP2 is maintained in a hyperphosphorylated state compared with IRP1 (28,29), and phosphopeptide mapping of IRP1 and IRP2 shows that, after PMA treatment and activation of PKC, IRP1 and IRP2 are further phosphorylated at different sites. We found a similar PMA-induced increase in IRP binding in GH3 and TtT97 cell lines in addition to HL-60 cells. Furthermore, TRH, EGF, and PMA activated PKC in pituitary cells, and PKC activation was obligatory for TRH-induced responses in TtT97 cells, but not in GH3 cells. Interestingly, our data indicate that the cell-specific changes in IRP binding activity induced by TRH and EGF may, in part, be due to activation of other intracellular signaling pathways. Notably, a MAPK consensus motif was reported to be present in IRP2 (28). In this regard, GH3 cells exhibited a marked augmentation of TRH and EGF action with MAPKK inhibition. Our results suggest a novel interplay between PKC and MAPK is critical for regulating IRP binding activity in pituitary cells, and that MAPK may play an important role in maintaining constitutive or basal IRP binding activity. Phosphatase inhibition studies provide further insight into the mechanisms of TRH and EGF action in the pituitary. Because the use of okadaic acid alone may induce differential effects on IRP binding (55), we used a combination of inhibitors to block phosphatases PP1, PP2A, and PTB. With this combination, we abolished the TRH-and EGF-induced down-regulation of IRP binding activity in GH3 cells. This implied a role for dephosphorylation and activation of phosphatases in the TRH-and EGF-induced modulation of IRP binding activity in pituitary cells. This would provide a novel mechanism of IRP modulation, which to date has centered on the activation of pathways leading to IRP phosphorylation. The particular phosphatases are yet to be identified, but may include PP1, PP2A and PTB. In conclusion, we propose that both ligands activate PKC and MAPK in GH3 cells, but that concurrent activation of specific phosphatases predominates leading to relative IRP dephosphorylation and reduced IRE binding affinity.
Although our results show that pituitary cells have abundant amounts of the two major IRPs (IRP1 and IRP2), IRP2 predominates in binding to the ferritin IRE, particularly in GH3 lactotroph cells. Differential regulation of IRP1 and IRP2 by TRH and EGF was not observed, unlike the effects in rat hepatoma cells where NO modulated IRP1 binding to a ferritin IRE, but not IRP2 (51). The binding affinity of IRP2 for the ferritin IRE is greater than for other IREs (62) found in the TfR, mitochondrial aconitase, erythroid 5Ј-aminolaevulinate synthase (62,63), DMT1 (64), and IREG1/ferroportin 1 (65) mRNAs. This suggests that in the pituitary, the major role for IRP2 will be in the regulation of ferritin expression, rather than other IRE-containing genes. Interestingly, IRP1:IRP2 ratios do vary among different cell types (63). For example, in liver, kidney, and intestines, IRP1 Ͼ IRP2 (66), and in rat reticulocytes, IRP1 ϭ IRP2 (62). In addition, mouse pro-B lymphocytes, Ba/F3 cells, contain no detectable IRP1 mRNA or protein (67). In these cells, IRP2 alone, at levels comparable to other cell types that also contain IRP1, regulated ferritin, and TfR protein similarly in response to iron (67). Whether a similar response in TfR and ferritin regulation would occur with signals that alter the phosphorylation profile of IRP2, as with TRH and EGF, remains to be tested. It may be that, under these circumstances, variations in ferritin regulation may be more pronounced compared with other IRE-containing genes.
The divergent IRP/ferritin IRE responses to TRH in two pituitary cell types raise the possibility that similar responses will occur with other IRE-containing genes. Generally, binding of IRPs to IREs in the 3Ј-UTR of the TfR mRNA stabilizes the mRNA resulting in increased intracellular uptake of iron (58). A similar pathway may be controlled by the 3Ј-UTR of the divalent metal iron transporter, DMT1 (64). In contrast, increased IRP binding to the 5Ј-UTR IREs in mammalian mitochondrial aconitase (69) and erythroid 5Ј-aminolaevulinate synthase (70) mRNAs would normally decrease levels of these proteins due to translational inhibition, resulting in lower Krebs cycle activity and less heme production, respectively. The recent cloning of a novel iron exporter (known as IREG1 or ferroportin 1) (65, 71) that is expressed in the basolateral membrane of epithelial cells of the duodenum, in Kupffer cells, and in macrophages, has added new insight into the regulation of iron homeostasis. IREG1/ferroportin 1 is considered to play a major role in iron export and is up-regulated in the iron overload disease, hereditary hemochromatosis (65). Interestingly, IREG1/ferroportin1 contains an IRE in the 5Ј-UTR. Unexpectedly, IREG1/ferroportin1 expression and mRNA levels increase with increased IRP binding, prompting speculation that the 5Ј-UTR IRE in IREG1/ferroportin 1 is either an mRNA stability determinant, or a translational enhancer control element (71). It will be of interest to determine the effect of TRH and EGF on each of these interactions in target cells.
The pituitary failure that accompanies iron overload in hereditary hemochromatosis (39) suggests a particular sensitivity of this organ to iron. This may reflect abnormalities such as the continued uptake of iron, due to aberrant expression of a second TfR similar to that recently described in an animal model of hemochromatosis (68). Abnormality of iron export due to dysregulation of IREG1/ferroportin 1 expression could also contribute to the accumulation of iron in the anterior pituitary. It remains unknown how ferritin expression is regulated when these cells are subjected to iron overload. At least one patient with hemochromatosis and pituitary insufficiency has recovered pituitary function after venesection and reduction in iron levels (38), suggesting that the defects may be reversible. Thus, investigation into the regulation of expression of the genes encoding iron storage and transport proteins in the pituitary in animal models of hemochromatosis may provide new insight into the tissue-specific control of iron homeostasis in this disorder.
In summary, TRH and EGF represent new regulators of IRP binding and iron homeostasis in the pituitary, where IRP2 binding to a ferritin IRE is predominant. Remarkably, TRH and EGF regulate IRP binding in a cell-specific manner that is PKC-dependent in one pituitary cell type, but not in the other. MAPK and protein phosphatases also play an important part in regulating IRP binding activity in pituitary cells. Together with the work of others (3,(27)(28)(29), our data emphasize the importance of IRP phosphorylation status in the regulation of IRP binding and the subsequent control of cellular iron homeostasis. Finally, these studies illustrate novel complexity in the regulation of IRP binding activity within different cell types of a single organ, and provide new insight into mechanisms by which TRH and EGF may modulate iron homeostasis and, consequently, cell growth and proliferation.