Expression of the Type II Iodothyronine Deiodinase in Cultured Rat Astrocytes Is Selenium-dependent*

(Received for publication, November 19, 1996, and in revised form, April 9, 1997)

Sophie Pallud Dagger , Ana-Maria Lennon Dagger , Martine Ramauge Dagger , Jean-Michel Gavaret Dagger , Walburga Croteau §, Michel Pierre Dagger , Françoise Courtin Dagger and Donald L. St. Germain §

From Dagger  U96 INSERM-Unité de Recherche sur la Glande Thyroïde et la Regulation Hormonale, 80, rue du Général Leclerc, 94276 Le Kremlin-Bicêtre Cedex, France and the § Departments of Medicine and Physiology, Dartmouth Medical School, Lebanon, New Hampshire 03756

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

The iodothyronine deiodinases are a family of selenoproteins that metabolize thyroxine and other thyroid hormones to active and inactive metabolites in a number of tissues including brain. Using primary cultures of rat astroglial cells as a model system, we demonstrate that the mRNA for the type II iodothyronine deiodinase (DII) selenoenzyme is rapidly and markedly induced by forskolin and 8-bromo-cAMP. The induction of DII activity, however, was significantly impaired by culturing cells in selenium-deficient medium for 7 days. Under such conditions, the addition of selenium resulted in a rapid increase in cAMP-induced DII activity that was dose-dependent, with maximal effects noted within 2 h. Cycloheximide blocked this effect of selenium on restoring cAMP-induced DII activity, whereas actinomycin D did not. These data demonstrate that the DII selenoenzyme is expressed in cultured astrocytes and that the induction of DII activity by cAMP analogues appears to be mediated, at least in part, by pretranslational mechanisms. Furthermore, selenium deprivation impairs the expression of DII activity at the level of translation.


INTRODUCTION

Recent molecular cloning studies have identified cDNAs that code for a family of structurally related iodothyronine deiodinases (1-4). These oxidoreductases catalyze the removal of iodide from the phenolic (5'-position) or the tyrosyl (5-position) ring of thyroxine (T4)1 to form the active compound 3,5,3'-triiodothyronine (T3) or the inactive metabolite 3,3',5'-triiodothyronine, respectively (5). Based on both their unique functional properties and the structural information derived from cDNA sequences, three isoforms designated type I, II, and III iodothyronine deiodinases (DI, DII, and DIII, respectively) have been identified.

The cDNAs for all three deiodinase isoforms contain within their coding regions an in-frame TGA triplet that has been demonstrated to code for the uncommon amino acid selenocysteine (1, 2, 4, 6, 7). This residue plays an essential role in the function of these enzymes; mutagenesis of the TGA triplet to a cysteine codon markedly reduces catalytic activity (1-3). It thus appears that the more potent nucleophilic capability of selenium, as compared with sulfur, is required for efficient deiodination. This thesis has been underscored by the identification of cDNAs coding for homologues of these enzymes from several fish, amphibian, and mammalian species (3, 8-10). All such cDNAs isolated to date have demonstrated strict conservation of the active-site selenocysteine codon.

Complementary DNAs that code for DII from Rana catesbeiana (3), rat (designated rBAT1-1), and human have been identified most recently (4). The proteins coded by these cDNAs are highly conserved; rBAT1-1 DII demonstrates 73 and 87% amino acid identity to the R. catesbeiana and human homologues, respectively. In addition, expression studies have demonstrated that these enzymes manifest all of the unique characteristics of DII (3, 4, 7).

Evidence that the rBAT1-1 DII cDNA and its human and amphibian homologues code for selenoproteins includes the following. (a) The in-frame triplet in these cDNAs, which could serve either as a selenocysteine codon or as a termination codon, does not function in the latter capacity; site-directed mutagenesis of this triplet to an unambiguous TAA stop codon renders the encoded protein inactive (3). (b) A selenocysteine insertion sequence element in the 3'-untranslated region of the DII mRNAs is required for the synthesis and expression of a full-length functional protein (3, 4, 7). (c) The amino acid sequence RPLVVNFGSATSePPF, which appears to form the catalytic core in the DII proteins, is 80% identical to the active-site sequence in the DI and DIII isoenzymes, which are known to be selenoproteins (1, 6, 11). (d) 75Se is specifically incorporated into a protein of the expected size (31 kDa) when the human DII cDNA is expressed to high levels in a transient transfection system (7).

Cultures of neonatal rat astrocytes have proven to be an important model for studying the regulation of deiodination in the brain (12-15). Astrocytes contain only low basal levels of DII activity, but expression can be rapidly induced by treatment with hydrocortisone and catecholamines or cAMP analogues (14-16). Based on an inability to label a candidate DII protein with 75Se and the finding that relatively short-term selenium depletion failed to affect basal levels of DII activity, previous investigators suggested that DII expressed in astroglial cells was not a selenoprotein (17).

In this study, we have utilized the rBAT1-1 DII cDNA and conditions of long-term selenium depletion combined with stimulation by cAMP analogues to examine further the characteristics of DII expressed in cultured astroglial cells. We demonstrate that the rBAT1-1 mRNA, in concert with DII activity, is rapidly and markedly induced by forskolin in cultured astroglial cells and that the expression of this activity is highly dependent on the selenium status of the cells. These results strongly suggest that DII activity in cultured astroglial cells results from the expression of the selenodeiodinase coded by the rBAT1-1 DII mRNA and that pretranslational mechanisms play a major role in the induction of DII activity by cAMP analogues.


EXPERIMENTAL PROCEDURES

Materials

T4, T3, dithiothreitol, 8-bromo-cAMP, aurothioglucose, cycloheximide, forskolin, actinomycin D, glutathione, glutathione reductase, t-butyl hydroperoxide, and antibiotics were obtained from Sigma. Sodium selenite and silica gel plates were from Merck (Darmstadt, Germany). NADPH was obtained from Boehringer (Mannheim, Germany). [125I]T4 (1500 µCi/µg) and [35S]methionine (1000 Ci/mmol) were purchased from Amersham International (Buckinghamshire, United Kingdom). N-Bromoacetyl-T3 was a generous gift of Henning (Berlin, Germany). Na125I was from CIS BioInternational (Gif-sur-Yvette, France). Harlan Sprague Dawley rats were from Iffa-Credo (L'Albresle, France). Fetal calf serum (FCS) and culture media were from Life Technologies, Inc., and culture dishes were from Nunclon (Roskilde, Denmark).

Methods

Cell Culture Conditions

Cerebral hemispheres were removed from 2-day-old Harlan Sprague Dawley rats, and primary cultures of astroglial cells were prepared as described previously (18). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 6 g/liter glucose, 2.4 g/liter sodium bicarbonate, antibiotics (100 units/ml penicillin, 100 µg/liter streptomycin, and 0.25 µg/liter amphotericin B), and 10% FCS (DMEM/FCS). The medium was changed every 2-3 days until cells reached confluency at ~10 days. At this stage, cells were rendered selenium-deficient using culture methods adapted from Aizenman and de Vellis (19). The DMEM/FCS was removed, and the cells were washed with a chemically defined medium and cultured for 7 additional days in this medium with daily medium changes. The chemically defined medium was selenium-free and consisted of a 1:1 mixture of DMEM and Ham's F-12 medium supplemented with 5.2 g/liter glucose, 1.8 g/liter sodium bicarbonate, and the antibiotics listed above. Using these culture conditions, 95% of the cells contained immunoreactive glial fibrillary acidic protein as we have previously described (18). Astrocytes were treated with the test agents for the times and at the concentrations indicated in each experiment.

RNA Preparation and Northern Analysis

RNA was prepared from cultured glial cells by the methods of Chomczynski and Sacchi (20). Poly(A)+ RNA was isolated by one cycle of chromatography over oligo(dT)-cellulose (Collaborative Biomedical Products, Bedford, MA). RNA samples from rat tissues used as controls were prepared as described previously (21). Northern blots were prepared, hybridized with cDNA probes, and washed following published methods (21), with the final wash performed at 60 °C for 60 min. The probe was prepared by polymerase chain reaction using the rBAT1-1 DII cDNA as template and gene-specific sense and antisense primers that amplified a 590-base pair fragment of the coding region. After hybridization and analysis with the rBAT1-1 cDNA probe, blots were stripped and reprobed with a rat beta -actin probe. In one experiment, a rat DI cDNA probe was also used. Hybridization signals were quantified on a PhosphorImager 445SI with the ImageQuant program (Molecular Dynamics, Inc., Sunnyvale, CA) on a Macintosh computer.

Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) Assay

A coupled RT-PCR technique was also used to determine the presence of rBAT1-1-associated transcripts in samples of astrocyte RNA. Reactions utilized the Access RT-PCR system (Promega, Madison, WI) with 20 ng of total RNA as template. Reaction conditions were as specified by the manufacturer, except that 30 cycles were used in the PCR with an annealing temperature of 58 °C. The oligonucleotide primers were derived from the coding region sequence of the rBAT1-1 cDNA and have previously been described (4). An amplification product of 590 base pairs was expected. Reaction mixtures lacking reverse transcriptase or RNA template were used as controls. Products were separated on a 1.0% agarose gel, transferred to a Magna Charge nylon membrane (Micron Separations, Westborough, MA), and hybridized with a nested radiolabeled oligonucleotide rat DII probe as described previously (4). After washing, the blot was exposed to x-ray film for 12 h. The signals were quantified using PhosphorImager analysis as described above.

Nucleotide Sequencing

The 590-base pair product derived from the RT-PCR was purified by gel electrophoresis on a 1% agarose gel using the QIAquick gel extraction kit (QIAGEN Inc.) as described above. This DNA fragment was then sequenced on both strands using the same primers used for the PCR amplification and an automated sequencing system with fluorescent dye terminators (Applied Biosystems, Foster City, CA).

DII Assay

At the time of harvesting, the medium was aspirated, the cells were rinsed twice with 3 ml of ice-cold phosphate-buffered saline, and the culture dishes were frozen at -80 °C. Cells were processed later by scraping the cells from each dish into 0.2 ml of sample buffer (20 mM HEPES, 2 mM dithiothreitol, and 0.25 M sucrose, pH 7.4), followed by sonication for 5 s. DII activity was measured by incubating aliquots of the cell sonicate (containing 4-40 µg of protein) in an 80-µl final volume of 20 mM HEPES, pH 7.4, containing 20 mM dithiothreitol, 50 nM T3, and 1 nM [125I]T4 for 20-60 min at 37 °C. Reactions were stopped by adding 10 µl of 10 M NH4OH containing 10 µM T3 and 10 µM T4. The [125I]T3 produced was separated from [125I]T4 by descending paper chromatography (12). Deiodination was linear with both protein and time, and the quantity of protein assayed was adjusted to ensure that <30% of the substrate was consumed. Kinetic analysis was performed in sonicates of forskolin-treated cells using 0.05-2 nM [125I]T4 as substrate and 20 mM dithiothreitol as cofactor. Kinetic constants were estimated using double-reciprocal plots.

Glutathione Peroxidase (GPx) Assay

GPx activity was measured using a modification of the methods of Flohé and Günzler (22) in the 10,000 × g supernatant of astroglial cell sonicates prepared as described above for the determination of DII activity. Assay tubes contained 20 mM potassium phosphate buffer, pH 7.4, 0.2 mM NADPH, 2 mM GSH, and 1 unit/ml glutathione reductase in a final volume of 500 µl. The samples (40 µl containing 15-40 µg of protein) were added to 450 µl of the reaction mixture and preincubated for 5 min at 25 °C before initiation of the reaction by addition of 10 µl of 35 mM t-butyl hydroperoxide. Absorbance at 340 nm was recorded for 10 min. The blank reaction, in which distilled water was substituted for sample, was subtracted from each assay.

Synthesis of BrAc[125I]T4

BrAc[125I]T4 was prepared by iodination of N-bromoacetyl-T3 as described previously (23). Its identity was confirmed by UV spectra, thin-layer chromatography, and high pressure liquid chromatography (Nucleosil RP18, Machery Nagel, Duren, Germany). The affinity label (specific activity > 3 mCi/µg) was >90% pure.

Affinity Labeling of Glial Cell Proteins

Aliquots of cell sonicates (50 µg of protein) prepared for DII activities were incubated for 20 min with 1 nM BrAc[125I]T4 at 37 °C. They were analyzed on 12.5% SDS-polyacrylamide gels according to Laemmli (24). Proteins were stained with Coomassie Brilliant Blue R-250, and the radiolabeled proteins were visualized by autoradiography.

Protein Determination

The protein content of cell sonicates was determined by the method of Bradford (25) using bovine serum albumin as standard.

Statistical Analysis

Statistical differences between groups were determined using Student's t test and were considered significant at p < 0.05.


RESULTS

Expression of the mRNA coding rBAT1-1 DII has been demonstrated in the rat anterior pituitary gland, BAT, the cerebral cortex, and the cerebellum (4). Northern analysis was performed to determine if this selenodeiodinase is also expressed in cultured astroglial cells. In these studies, cells were maintained in a selenium-fed state and treated prior to harvesting with agents known to induce DII or DIII activity (14-16, 18, 26, 27). The results from two experiments are shown in Fig. 1A. Control (unstimulated) cells contained no detectable rBAT1-1-associated transcripts, consistent with the very low levels of DII activity measured in these same cultures (Fig. 1B). Forskolin treatment, however, induced a marked increase in both DII activity and a 7.5-kilobase transcript that hybridizes with the rBAT1-1 cDNA. This transcript is the same size as that noted in a control sample of BAT RNA and is consistent with our previous demonstration of a 7.5-kilobase transcript in the rat cerebral cortex (4). Treatment of cells with either retinoic acid or T3 induced little or no increase in either DII activity or rBAT1-1-associated transcripts.


Fig. 1. Effects of agents known to induce DII or DIII activity on the expression of DII mRNA (A) and DII activity (B) in cultured astrocytes. The results of two separate experiments are shown. Prior to harvesting, selenium-fed cells were treated with forskolin (Forsk; 10 µM for 4 h in Experiment 1 and 10 µM for 6 h in Experiment 2), all-trans-retinoic acid (Ret. Acid, RA; 5 µM for 48 h), or T3 (10 nM for 24 h). Northern analysis was performed using ~10 µg of poly(A)+ RNA/lane and a radiolabeled probe derived from the coding region of the rBAT1-1 cDNA. RNA prepared from BAT of a cold-exposed rat was included as a positive control (Cont). After autoradiography and PhosphorImager analysis, the blot was stripped and hybridized with a rat beta -actin probe. DII activity was determined in sonicates of the same cell cultures. kb, kilobase.
[View Larger Version of this Image (32K GIF file)]

The rapidity with which forskolin stimulates the level of DII mRNA is shown in Fig. 2A. When corrected using actin for RNA loading, a rise in DII transcripts was clearly evident by Northern analysis within 1 h of stimulation, with maximal levels achieved at 4 h (Fig. 2B). Furthermore, when the more sensitive technique of RT-PCR was applied to these samples, a marked increase in the DII amplicon was noted in the RNA sample obtained at 30 min after forskolin addition (Fig. 3). Notably, DII activity was not increased in these same cultures at 30 min relative to control levels (2.1 and 1.6 fmol of T3/min/mg of protein, respectively). After 8 h of exposure, however, activity was increased 30-fold (49.3 fmol of T3/min/mg of protein), consistent with both the marked increase in DII mRNA levels and prior reports of the time course of forskolin stimulation of DII activity (14, 15).


Fig. 2. Time course of stimulation of DII mRNA by forskolin treatment of cultured astrocytes. Astrocytes were treated with 10 µM forskolin and then harvested at the times indicated for the preparation of RNA. A, Northern analysis using ~1 µg of astrocyte poly(A)+ RNA/lane and radiolabeled probes derived from the rBAT1-1 cDNA, a rat DI cDNA, and a beta -actin cDNA; B, ratio of DII to beta -actin mRNA based on PhosphorImager analysis. Units are arbitrary. kb, kilobase pairs.
[View Larger Version of this Image (23K GIF file)]


Fig. 3. RT-PCR analysis of the time course of the stimulation of DII transcripts by forskolin treatment of cultured astrocytes. Astrocytes were treated with 10 µM forskolin and then harvested at the times indicated for the preparation of RNA. The PCR products from each amplification were separated on a 1% agarose gel and transferred by capillary blotting to a nylon membrane. The membrane was hybridized with a nested radiolabeled primer and exposed to film. Control lanes included mixtures containing no reverse transcriptase (RT), no RNA template (Temp), and a reaction mixture containing pituitary RNA as a positive control. Essentially identical results were obtained in a second amplification using samples of RNA from a separate experiment.
[View Larger Version of this Image (28K GIF file)]

Probing of the Northern blot depicted in Fig. 2A with a labeled cDNA for rat DI failed to detect any hybridizing species. As a control in this same hybridization reaction, a 2.1-kilobase band was demonstrated in a sample of RNA from rat liver (data not shown), consistent with the reported size of the DI mRNA (1, 28).

To verify the identity of the transcripts hybridizing with the rBAT1-1 cDNA probe in astroglial cells, the 590-base pair PCR product shown in Fig. 3 was purified and sequenced on both strands. As shown in Fig. 4, the sequence of the PCR product was identical to that of the rBAT1-1 cDNA (4), including the presence of the in-frame TGA triplet that codes for selenocysteine.


Fig. 4. Comparison of the sequence of the glial cell RNA-derived PCR product shown in Fig. 3 with the sequence of the coding region of the rBAT1-1 DII cDNA. Underlined sequences represent the primers used for the PCR amplification and DNA sequencing reactions and the nested primer used as a probe for the RT-PCR experiment shown in Fig. 3.
[View Larger Version of this Image (56K GIF file)]

Given these results demonstrating that the mRNA coding the type II selenodeiodinase is highly induced in astroglial cells by forskolin, we undertook additional studies using this model system to re-examine the effects of selenium on the expression of DII activity. Cells were depleted of selenium by culturing them in selenium-free medium, which was changed daily, for 7 days. As a control, the activity of another selenoprotein, GPx, was also determined. Under these culture conditions, GPx activity was decreased by 70-90% as compared with the values observed in selenium-replete astrocytes. DII activity in selenium-replete and selenium-depleted cells was determined both before and after stimulation with forskolin or 8-bromo-cAMP. As shown in Fig. 5, selenium depletion for 7 days had little or no effect on the low level of DII activity present in control astrocytes. However, in selenium-depleted cells, the stimulatory effect of forskolin was significantly blunted; stimulated activity in depleted cells was only 50% of the value observed in selenium-fed astrocytes. In a second experiment, DII activity was stimulated using 8-bromo-cAMP in astrocytes cultured for varying periods of time in selenium-deficient medium. Stimulated levels of activity progressively declined as the period of selenium depletion increased from 1 to 4 days and then plateaued at a value approximately one-third of that noted in cells maintained in selenium-containing medium (Fig. 6).


Fig. 5. Effect of selenium depletion on DII activity in cultured astrocytes. After reaching confluence, cells were incubated for 7 days in serum-free, selenium-deficient medium ± added selenium (30 nM). Forskolin (10 µM) was added to some cultures 4 h prior to harvesting. DII activity was determined in cell sonicates. Closed bars represent activity in cells maintained in selenium-containing medium, whereas open bars are data from cells in selenium-deficient medium. Data are the means ± S.D. of results obtained from three dishes. **, p < 0.005 versus forskolin-treated cells maintained in selenium-containing medium. Similar results were observed in two additional experiments.
[View Larger Version of this Image (10K GIF file)]


Fig. 6. Time course of the effect of selenium depletion on 8-bromo-cAMP-induced DII activity in cultured astrocytes. After reaching confluence, cells were placed in serum-free medium supplemented with 30 nM selenium. The medium was changed daily, with some cells being switched to selenium-deficient medium at 1, 2, 3, 4, or 7 days prior to harvesting. On the 8th day, astrocytes were incubated with 1 mM 8-bromo-cAMP for 6 h, and then all cells were harvested, and DII activity was determined in cell sonicates. Data are the means ± S.D. of results obtained from three dishes. *, p < 0.01; **, p < 0.005 versus selenium-replete cells.
[View Larger Version of this Image (9K GIF file)]

This impaired expression of DII activity in selenium-depleted cells did not result from a change in the kinetic characteristics of the enzyme. Using T4 as substrate, Km values of 0.42 and 0.40 nM were determined for DII activity in selenium-replete and selenium-depleted cells, respectively. This low Km and the insensitivity of this DII activity to inhibition by propylthiouracil and aurothioglucose (data not shown) confirm prior studies (14, 15) that 5'-deiodination in cultured rat astrocytes, as in brain (29), is catalyzed by a DII.

The effects of selenium repletion were studied in cells previously depleted of this element for 7 days. In the experiment depicted in Fig. 7A, cells were treated with 8-bromo-cAMP for 6 h prior to harvesting. Selenium supplementation (30 nM) was begun at the indicated times prior to harvesting. As shown, supplementation for as short a time as 1 h prior to harvesting induced a significant increase in the cAMP-stimulated DII activity. The effect was maximal (5-fold increase in stimulated activity) after 2 h of selenium supplementation and was maintained with longer exposure times. The effects of selenium repletion on GPx activity were also studied. The restoration of its activity required at least 24-48 h (Fig. 7B). Fig. 8 demonstrates that the restoration of normal DII responsiveness to forskolin stimulation by selenium supplementation occurs in a dose-dependent fashion. Thus, the addition for 6 h of 0.3 nM selenium to the culture medium of selenium-depleted cells partially restored forskolin-induced stimulation, whereas 1 nM selenium induced a maximal effect.


Fig. 7. Effect of selenium repletion on DII (A) and GPx (B) activities in cultured astrocytes. After reaching confluence, cells were incubated for 7 days in serum-free, selenium-deficient medium. Selenium (30 nM) was added to the culture medium at the times indicated prior to harvesting. For determining the effects of selenium repletion on DII activity, cells were stimulated with 8-bromo-cAMP (1 mM) at 6 h prior to harvesting. DII activity was determined in cell sonicates. Data are the means ± S.D. of results obtained from three dishes. *, p < 0.02; **, p < 0.005 versus the respective controls in selenium-depleted cells. Similar results were observed in two additional experiments.
[View Larger Version of this Image (17K GIF file)]


Fig. 8. Effect of selenium concentration on DII activity in cultured astrocytes. After reaching confluence, cells were incubated for 7 days in serum-free, selenium-deficient medium. Selenium, at the concentrations indicated, was then added to the culture medium at 6 h prior to harvesting. Cells were stimulated with forskolin (10 µM) for the last 4 h of the incubation period. DII activity was determined in cell sonicates. Data are the means ± S.D. of results obtained from three dishes. *, p < 0.01; **, p < 0.005 versus the control in selenium-depleted cells.
[View Larger Version of this Image (13K GIF file)]

We also studied the effect of selenium deficiency on the abundance of the putative 30-kDa (p30) substrate-binding subunit of DII (23, 30). In selenium-depleted cells treated with 8-bromo-cAMP, the affinity labeling of p30 with BrAc[125I]T4 was reduced in comparison with that observed in cells cultured in the presence of 30 nM selenium (Fig. 9). No effect of selenium was observed on the labeling of p30 under basal conditions, consistent with the lack of an effect of selenium deprivation on basal DII activity.


Fig. 9. Effect of selenium depletion on BrAc[125I]T4 incorporation into the p30 protein. After reaching confluence, cells were incubated for 7 days in serum-free, selenium-deficient medium. For the last 2 days of the culture period, half of the cells were repleted with selenium by the addition of 30 nM selenium to the culture medium. Cells were stimulated with 8-bromo-cAMP (1 mM) for 6 h prior to harvesting. Aliquots of cell sonicates were affinity-labeled and analyzed by SDS-polyacrylamide gel electrophoresis as described under "Experimental Procedures."
[View Larger Version of this Image (13K GIF file)]

To determine if the impairment in cAMP-stimulated DII activity observed in selenium-depleted cells resulted from an increased rate of loss or inactivation of the DII protein, DII activity in selenium-fed and selenium-depleted cells was quantified after blocking protein synthesis. Cells were first treated with 1 mM 8-bromo-cAMP to induce DII activity and were then exposed 6 h later to cycloheximide (18 µM). Estimates for the half-life of DII activity were the same in selenium-depleted and selenium-replete cells (1.3 ± 0.4 and 1.4 ± 0.4 h, respectively; n = four experiments). This indicates that selenium depletion does not increase the rate of loss of DII activity, but rather decreases the rate of DII production.

Additional studies in selenium-depleted cells examined whether the increased DII activity observed after selenium repletion was dependent on protein synthesis. Selenium-depleted cells were first incubated with 10 µM forskolin. After 3.5 h, cycloheximide (18 µM) or medium (as a control) was added, followed 30 min later by the addition of selenium (30 nM) to one-half of the cultures. Incubations were continued for an additional 1 h. The results are shown in Fig. 10A. In forskolin-pretreated cells not exposed to cycloheximide, the addition of selenium for 1 h induced a marked increase in DII activity, as demonstrated previously (Fig. 5). In contrast, cycloheximide treatment completely abolished this selenium-induced increase in DII activity. (In control experiments, we demonstrated that pretreatment with cycloheximide at 18 µM for 30 min blocked protein synthesis at 1 h by 95% as indicated by the incorporation of [35S]methionine.)


Fig. 10. Effects of cycloheximide (A) and actinomycin D (B) on selenium induction of DII activity in cultured astrocytes. After reaching confluence, cells were incubated for 7 days in serum-free, selenium-deficient medium. Cells were then treated with 10 µM forskolin to induce DII activity. At 3.5 h after the addition of forskolin, cycloheximide (Cyclohex.; 18 µM), actinomycin D (Actino. D; 6 µM), or medium (as a control) was added, followed 30 min later by the addition of selenium (30 nM) to one-half of the cultures. Incubations were continued for an additional 1 h (A) or 2 h (B) before harvesting and the determination of DII activity in cell sonicates. Data are the means ± S.D. of results obtained from three dishes. *, p < 0.02; **, p < 0.005 versus the respective values in selenium-depleted cells. Similar results were observed in two additional experiments.
[View Larger Version of this Image (18K GIF file)]

To determine whether the stimulatory effect of selenium was dependent on transcriptional activation, experiments similar to those just described were repeated using actinomycin D. In preliminary experiments and consistent with the report of others (15), we observed that a 30-min pretreatment of astrocytes with actinomycin D (6 µM) completely blocked the induction of DII activity by a 2-h treatment with 10 µM forskolin (data not shown). In the experiment presented in Fig. 10B, selenium-depleted cells were first treated with forskolin for 3.5 h, and then actinomycin D or medium (as a control) was added. Selenium (30 nM) was added 30 min later to one-half of the cultures, and incubations were continued for an additional 2 h. In cells not treated with actinomycin D, selenium again enhanced the response to forskolin ~2-fold. In contrast to the effects of cycloheximide, actinomycin D did not block this stimulatory effect of selenium on DII activity. This suggests that the effect of selenium is largely independent of transcriptional activity and is mediated by an enhanced rate of translation.


DISCUSSION

The tissue-specific patterns of deiodinase expression serve a critical role in the regulation of thyroid hormone action (5, 31). This appears to be particularly true in the central nervous system, which is the only tissue in rodents and humans that expresses both an activating (DII) and an inactivating (DIII) deiodinase during the developmental period and in adulthood (32-35). These enzymes are likely responsible in large part for the relative stability of brain T3 levels during states of altered thyroid function (36-38). Thus, DII activity, and hence the relative efficiency of T3 production, is markedly increased in hypothyroidism (38), whereas T3 degradation is markedly reduced (39).

Previous studies have provided convincing evidence that mammalian DII coded by the rBAT1-1 cDNA is a selenoprotein (3, 4, 7). In the rat, Northern analysis has demonstrated that this deiodinase is expressed in BAT, the anterior pituitary gland, the cerebral cortex, and the cerebellum (4). In humans, expression appears to be more widespread, with mRNA and DII activity being detected also in placenta, skeletal muscle, heart, and the thyroid gland (4, 40). This study extends these previous observations and demonstrates that cultured rat astroglial cells also express the rBAT1-1 selenodeiodinase and that, not unexpectedly, the degree of expression is highly dependent on the presence of inducers of DII activity and on the selenium status of the cells. Based on the experimental results utilizing cycloheximide and actinomycin D, the impairment of induction of DII activity by selenium depletion appears to occur at the level of translation. Such a finding is consistent with the known mechanisms whereby selenocysteine is incorporated into proteins (41). Indeed, the effect of selenium deficiency on impairing selenoprotein translation has previously been demonstrated for DI (21, 42).

Our demonstration that the selenium status of astroglial cells markedly alters the expression of DII activity contrasts with the prior study of Safran et al. (17), who noted no effect of medium selenium concentrations on basal DII activity in this same model system. However, there are two significant methodological differences between their study and ours. First, the degree of selenium depletion in this study, in which selenium-free culture medium was changed daily for up to 7 days, is likely greater than that induced by Safran et al. (17). These investigators, after initially washing cells in selenium-free medium, cultured them for shorter time periods with no further medium changes. Second, DII activity in the study of Safran et al. (17) was determined only in the basal unstimulated state, where enzyme activity is low and thus likely to be less sensitive to selenium deprivation. In our study, basal DII activity also was unaffected by selenium depletion, whereas stimulated activity was markedly diminished.

The ability to radiolabel selenoproteins with 75Se represents an important means of detecting and studying their expression (43). It is therefore somewhat surprising, given the present results, that Safran et al. (17) were unable to label a candidate DII selenoprotein using this isotope in astroglial cells stimulated with cAMP. Our experience has been similar.2 This may reflect compartmentalization of selenium stores within the cell, the relatively low abundance of the DII protein, and/or the insensitivity of this labeling technique. Regarding the latter two possibilities, Salvatore et al. (7) have labeled human DII with 75Se in a transient transfection system. Of note, however, the levels of DII activity expressed in their system were 10-20-fold higher than those observed in forskolin-stimulated astroglial cells.

The effects of selenium deficiency on deiodinase activities in vivo are dependent on both the degree of selenium deprivation and which tissues are examined (44-46). In the liver and kidney, DI activity is markedly impaired by nutritional selenium deprivation in the rat, whereas thyroid DI activity is preserved, reflecting the ability of this organ to conserve its selenium stores (43, 44, 47). A similar resistance to selenium depletion has been noted in the brain (43), and DII and DIII levels in this tissue are only minimally altered by nutritional selenium deficiency (45, 48). Notably, this ability of some tissues to retain selenium appears to be maintained when cells from these organs are placed into primary culture (46). Such a phenomenon may explain why relatively prolonged selenium deprivation is necessary to impair DII expression in cultured astroglial cells.

Four days of selenium deprivation impaired cAMP-stimulated DII activity in our cultured astroglial cells by ~65%. However, maintenance of cells in selenium-free medium for an additional 3 days did not result in further impairment of stimulated DII activity (Fig. 6), suggesting that the culture conditions employed do not allow for complete cellular depletion of selenium stores. In this regard, it is not known what amount of selenium must be retained by the cell to maintain partial DII responsiveness to cAMP stimulation, although conceivably such amounts might be quite small if selenium is sequestered within intracellular compartments involved in protein synthesis and/or is used preferentially for DII translation as compared with other selenoproteins. Alternatively, a minor component of DII activity in astroglial cells may derive from a selenium-independent catalytic mechanism, although to date there is no compelling evidence that non-selenoprotein deiodinases exist.

In this study, selenium deprivation impaired the expression of GPx activity to a significantly greater extent than DII activity. Furthermore, repletion with selenium for 2 h resulted in maximal responsiveness of DII activity to cAMP stimulation, whereas at least 2 days of selenium refeeding were required to restore GPx activity to the level observed in selenium-fed control cells. Although differences in the rates of turnover of these proteins may be responsible in part for these observations, the results support the emerging concept that there is a hierarchy of expression of selenoproteins within a given cell or tissue (49, 50). We have previously reported similar preferential expression of DIII after selenium repletion of glial cells (51). The mechanisms responsible for this differential utilization of selenium within the cell remain poorly defined, but could be secondary to selective alterations in selenoprotein mRNA levels induced by selenium deficiency (52-54) or differences in the translational efficiency of deiodinase versus GPx mRNAs. This latter effect might result from variation in the effectiveness with which the DII and GPx selenocysteine insertion sequence elements direct selenocysteine incorporation during protein synthesis (55).

Finally, our results provide insight into the mechanism by which cAMP increases DII activity in cultured glial cells. It has recently been proposed that this effect is due to a post-translational mechanism that involves the translocation within the cell of preformed enzyme (56). The present results strongly suggest that pretranslational mechanisms are also involved, in that DII mRNA is rapidly and markedly induced by forskolin and 8-bromo-cAMP. Furthermore, our finding that selenium depletion markedly impairs cAMP stimulation of DII activity is further evidence that synthesis of new enzyme is required for this effect.

In conclusion, the present results provide persuasive evidence that the selenoenzyme encoded by the rBAT1-1 cDNA is induced in cultured rat astroglial cells by cAMP and forskolin and that DII activity in these cells is dependent on the availability of selenium. Furthermore, DII expression in this model system appears to have a high priority relative to that of another selenoprotein, GPx. Such a privileged status may reflect the importance of this enzyme in the homeostatic control of thyroid hormone action in the central nervous system.


FOOTNOTES

*   This work was supported in part by the Association pour la Recherche contre le Cancer (to M. P.) and by National Institutes of Health Grant DK-42271 (to D. L. S.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U53505.


   To whom correspondence should be addressed. Fax: 33-1-49-59-85-40; E-mail: courtin{at}kb.inserm.fr.
1   The abbreviations used are: T4, thyroxine; BrAcT4, N-bromoacetylthyroxine; T3, 3,5,3'-triiodothyronine; BAT, brown adipose tissue; DI, type I iodothyronine deiodinase; DII, type II iodothyronine deiodinase; DIII, type III iodothyronine deiodinase; FCS, fetal calf serum; DMEM, Dulbecco's modified Eagle's medium; RT-PCR, reverse transcriptase-polymerase chain reaction; GPx, glutathione peroxidase.
2   A. M. Lennon, unpublished data.

REFERENCES

  1. Berry, M. J., Banu, L., and Larsen, P. R. (1991) Nature 349, 438-440 [CrossRef][Medline] [Order article via Infotrieve]
  2. St. Germain, D. L., Schwartzman, R., Croteau, W., Kanamori, A., Wang, Z., Brown, D. D., and Galton, V. A. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 7767-7771 [Abstract] ; Correction (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 11282
  3. Davey, J. C., Becker, K. B., Schneider, M. J., St. Germain, D. L., and Galton, V. A. (1995) J. Biol. Chem. 270, 26786-26789 [Abstract/Free Full Text]
  4. Croteau, W., Davey, J. C., Galton, V. A., and St. Germain, D. L. (1996) J. Clin. Invest. 98, 405-417 [Abstract/Free Full Text]
  5. St. Germain, D. L. (1994) Trends Endocrinol. Metab. 5, 36-42
  6. Salvatore, D., Low, S. C., Berry, M., Maia, A. L., Harney, J. W., Croteau, W., St. Germain, D. L., and Larsen, P. R. (1995) J. Clin. Invest. 96, 2421-2430 [Medline] [Order article via Infotrieve]
  7. Salvatore, D., Tibor, B., Harney, J. W., and Larsen, P. R. (1996) Endocrinology 137, 3308-3315 [Abstract]
  8. Mandel, S. J., Berry, M. J., Kieffer, J. D., Harney, J. W., Warne, R. L., and Larsen, P. R. (1992) J. Clin. Endocrinol. & Metab. 75, 1133-1139 [Abstract]
  9. Becker, K. B., Schneider, M. J., Davey, J. C., and Galton, V. A. (1995) Endocrinology 136, 4424-4431 [Abstract]
  10. Valverde-R, C., Croteau, W., LaFleur, G. L., Jr., Orozco, A., and St. Germain, D. L. (1997) Endocrinology 138, 642-648 [Abstract/Free Full Text]
  11. Croteau, W., Whittemore, S. L., Schneider, M. J., and St. Germain, D. L. (1995) J. Biol. Chem. 270, 16569-16575 [Abstract/Free Full Text]
  12. Courtin, F., Chantoux, F., and Francon, J. (1986) Mol. Cell. Endocrinol. 48, 167-178 [CrossRef][Medline] [Order article via Infotrieve]
  13. Cavalieri, R. R., Gavin, L. A., Cole, R., and de Vellis, J. (1986) Brain Res. 364, 382-385 [CrossRef][Medline] [Order article via Infotrieve]
  14. Courtin, F., Chantoux, F., Pierre, M., and Francon, J. (1988) Endocrinology 123, 1577-1581 [Abstract]
  15. Leonard, J. L. (1988) Biochem. Biophys. Res. Commun. 151, 1164-1172 [Medline] [Order article via Infotrieve]
  16. Courtin, F., Chantoux, F., Gavaret, J.-M., Toru-Delbauffe, D., Jacquemin, C., and Pierre, M. (1989) Endocrinology 125, 1277-1281 [Abstract]
  17. Safran, M., Farwell, A. P., and Leonard, J. L. (1991) J. Biol. Chem. 266, 13477-13480 [Abstract/Free Full Text]
  18. Esfandiari, A., Gagelin, C., Gavaret, J.-M., Pavelka, S., Lennon, A.-M., Pierre, M., and Courtin, F. (1994) Glia 11, 255-261 [Medline] [Order article via Infotrieve]
  19. Aizenman, Y., and de Vellis, J. (1987) Brain Res. 414, 301-308 [CrossRef][Medline] [Order article via Infotrieve]
  20. Chomczynski, P., and Sacchi, N. (1987) Anal. Biochem. 162, 156-159 [CrossRef][Medline] [Order article via Infotrieve]
  21. DePalo, D., Kinlaw, W. B., Zhao, C., Engelberg-Kulka, H., and St. Germain, D. L. (1994) J. Biol. Chem. 269, 16223-16228 [Abstract/Free Full Text]
  22. Flohé, L., and Günzler, W. A. (1984) Methods Enzymol. 105, 114-121 [Medline] [Order article via Infotrieve]
  23. Lennon, A.-M., Esfandiari, A., Gavaret, J.-M., Courtin, F., and Pierre, M. (1994) J. Neurochem. 62, 2116-2123 [Medline] [Order article via Infotrieve]
  24. Laemmli, U. K. (1970) Nature 227, 680-685 [Medline] [Order article via Infotrieve]
  25. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 [CrossRef][Medline] [Order article via Infotrieve]
  26. Courtin, F., Liva, P., Gavaret, J.-M., Toru-Delbauffe, D., and Pierre, M. (1991) J. Neurochem. 56, 1107-1113 [Medline] [Order article via Infotrieve]
  27. Esfandiari, A., Courtin, F., Lennon, A.-M., Gavaret, J.-M., and Pierre, M. (1992) Endocrinology 131, 1682-1688 [Abstract]
  28. St. Germain, D. L., Dittrich, W., Morganelli, C. M., and Cryns, V. (1990) J. Biol. Chem. 265, 20087-20090 [Abstract/Free Full Text]
  29. Visser, T. J., Leonard, J. L., Kaplan, M. M., and Larsen, P. R. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, 5080-5084 [Abstract]
  30. Farwell, A. P., and Leonard, J. L. (1989) J. Biol. Chem. 264, 20561-20567 [Abstract/Free Full Text]
  31. Köhrle, J. (1994) Exp. Clin. Endocrinol. 102, 63-89 [Medline] [Order article via Infotrieve]
  32. Kaplan, M. M., and Yaskoski, K. A. (1980) J. Clin. Invest. 66, 551-562 [Medline] [Order article via Infotrieve]
  33. Kaplan, M. M., and Yaskoski, K. A. (1981) J. Clin. Invest. 67, 1208-1214 [Medline] [Order article via Infotrieve]
  34. Campos-Barros, A., Hoell, T., Musa, A., Sampaola, S., Stoltenburg, G., Pinna, G., Eravci, M., Meinhold, H., and Baumgartner, A. (1996) J. Clin. Endocrinol. & Metab. 81, 2179-2185 [Abstract]
  35. Karmarkar, M. G., Prabarkaran, D., and Godbole, M. M. (1993) Am. J. Clin. Nutr. 57, (suppl.) 291S-294S
  36. van Doorn, J., Roelfsema, F., and van der Heide, D. (1986) Acta Endocrinol. 113, 59-64 [Medline] [Order article via Infotrieve]
  37. van Doorn, J., van der Heide, D., and Roelfsema, F. (1984) Endocrinology 115, 174-182 [Abstract]
  38. Calvo, R., Obregón, M. J., Ruiz de Ona, C., Escobar del Rey, F., and de Escobar, G. M. (1990) J. Clin. Invest. 86, 889-899 [Medline] [Order article via Infotrieve]
  39. Silva, J. E., and Matthews, P. S. (1984) J. Clin. Invest. 74, 1035-1049 [Medline] [Order article via Infotrieve]
  40. Salvatore, D., Tu, H., Harney, J. W., and Larsen, P. R. (1996) J. Clin. Invest. 98, 962-968 [Abstract/Free Full Text]
  41. Stadtman, T. C. (1991) J. Biol. Chem. 266, 16257-16260 [Free Full Text]
  42. Berry, M. J., Harney, J. W., Ohama, T., and Hatfield, D. L. (1994) Nucleic Acids Res. 22, 3753-3759 [Abstract]
  43. Behne, D., Hilmert, H., Scheid, S., Gessner, H., and Elger, W. (1988) Biochim. Biophys. Acta 966, 12-21 [Medline] [Order article via Infotrieve]
  44. Vadhanavikit, S., and Ganther, H. E. (1993) J. Nutr. 123, 1124-1128 [Medline] [Order article via Infotrieve]
  45. Meinhold, H., Campos-Barros, A., Walzog, B., Köhler, R., Müller, F., and Behne, D. (1993) Exp. Clin. Endocrinol. 101, 87-93 [Medline] [Order article via Infotrieve]
  46. Beech, S. G., Walker, S. W., Beckett, G. J., Arthur, J. R., Nicol, F., and Lee, D. (1995) Analyst 120, 827-831 [Medline] [Order article via Infotrieve]
  47. Chanoine, J., Braverman, L. E., Farwell, A. P., Safran, M., Alex, S., Dubord, S., and Leonard, J. L. (1993) J. Clin. Invest. 91, 2709-2713 [Medline] [Order article via Infotrieve]
  48. Chanoine, J. P., Safran, M., Farwell, A. P., Tranter, P., Ekenbarger, D. M., Dubord, S., Alex, S., Arthur, J. R., Becker, G. J., Braverman, L. E., and Leonard, J. L. (1992) Endocrinology 130, 479-484
  49. Gross, M., Oertel, M., and Köhrle, J. (1995) Biochem. J. 306, 851-856 [Medline] [Order article via Infotrieve]
  50. Bermano, G., Nicol, F., Dyer, J. A., Beckett, G. J., Arthur, J. R., and Hesketh, J. E. (1995) Biochem. J. 311, 425-430 [Medline] [Order article via Infotrieve]
  51. Ramauge, M., Pallud, S., Esfandiari, A., Gavaret, J.-M., Lennon, A.-M., Pierre, M., and Courtin, F. (1996) Endocrinology 137, 3021-3025 [Abstract]
  52. Christensen, M. J., and Burgener, K. W. (1992) J. Nutr. 122, 1620-1626 [Medline] [Order article via Infotrieve]
  53. Toyoda, H., Himeno, S., and Imura, N. (1989) Biochim. Biophys. Acta 1008, 301-308 [Medline] [Order article via Infotrieve]
  54. Lei, X. G., Evenson, J. K., Thompson, K. M., and Sunde, R. A. (1995) J. Nutr. 125, 1438-1446 [Medline] [Order article via Infotrieve]
  55. Berry, M. J., Banu, L., Harney, J. W., and Larsen, P. R. (1993) EMBO J. 12, 3315-3322 [Abstract]
  56. Safran, M., Farwell, A. P., and Leonard, J. L. (1996) J. Biol. Chem. 271, 16363-16368 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.