Relationship between Thyroid and Glucocorticoid Hormone Receptor Occupancy, Growth Hormone Gene Transcription, and mrna Accumulation*

Transcription

1 THE JOURNAL OF BOLOGCAL CHEMSFY by The Amencan Society of B~olog~cal Chemists, nc. Val. 259, No. 20, mue of October 25, pp ,1984 Printed in U.S.A. Relationship between Thyroid and Glucocorticoid Hormone Receptor Occupancy, Growth Hormone Gene Transcription, and mrna Accumulation* (Received for publication, October 3, 1983) Jennifer K. Nyborg, Anh P. Nguyen, and Stephen R. Spindler From the Department of Biochemistry, University of California, Riverside, California The temporal relationship between the occupancy of When transcriptional activation did not occur, GH the thyroid hormone (3,5,3 -triiodo-~-thyronine (Ts)) mrna usually still increased 24 h after hormone treatand glucocorticoid (dexamethasone) receptors and the ment was begun, but not early in the induction. n rate of growth hormone (GH) gene transcription and contrast, a rapid rise in GH mrna was observed very mrna accumulation were investigated. The GC line early in the induction when increased transcription of of cultured rat pituitary tumor cells was used in these the gene occurred. These results suggest that glucocorstudies. The rate of GH gene transcription was mea- ticoids enhance GH mrna accumulation by post-transured by elongation of in vivo initiated RNA chains in scriptional as well as transcriptional means. isolated nuclei using radioactive precursors and hybridiation of labeled transcripts to immobilied GH gene sequences. T receptor occupancy was directly proportional to the rate of GH gene transcription during the initial phase of hormone induction. These results suggest that a single occupied receptor is suffi- cient to fully activate the gene and that the unoccupied receptor resides close to the gene regulatory site. A decrease in GH gene transcription rapidly followed the initial increase in activity. This decrease occurred in the absence of new protein synthesis and was not accompanied by a proportional decrease in the level of occupied TS receptor. An analysis of dexamethasone activation of GH gene transcription found maximum stable binding of occupied receptor to nuclei within 5 min of hormone addition, while transcription of the gene did not become fully activated until 30 to 60 min later. Transcription of the gene then declined to a rate 2 to 3 times that observed before addition of dexamethasone. These re- sults suggest that at least one relatively slow molecular step precedes glucocorticoid hormone-receptor enhancement of gene transcription and that a second molecular event later attenuates the response. The simultaneous addition of both hormones produced two peaks of enhanced transcription, each apparently due to the separate effects of the hormones. GH mrna levels were closely linked to the rate of transcription of the gene under all induction conditions. Comparison of the rates of decay of GH mrna in cells cultured in the presence or the absence of hormones suggests that GH mrna is much more stable when Ts is present. The glucocorticoid responses of the gene were consistently obtained only in the presence of Ts. These results and the transcriptional synergism of the two hormones suggest a direct interaction of the occupied receptors at their regulatory sites. The transcriptional effects of glucocorticoids were observed only in about half the experiments performed in the absence of Ts. * This research was supported by Research Grant AM from the National nstitutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked aduertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact, The GH lines of rat pituitary tumor cells express GH and prolactin in culture and provide useful model systems for the study of GH gene expression (1, 2). The rate of production of growth hormone in these cells is influenced by a number of factors, including glucocorticoid and thyroid hormones (2-6). Both hormones modulate GH production separately, and together their effects are synergistic (4,6). Measurement of GH mrna levels by cell-free translation of isolated mrna and by molecular hybridiation techniques has demonstrated that the hormones enhance the production of both mature and precursor forms of GH RNA (4-10). Recently, we have shown that both thyroid and glucocorticoid hormones enhance the rate of transcription of the GH gene by RNA polymerase 1 in the GC line of GH cells (10). Together, the hormones were found to be synergistic in their transcriptional effects. This work has been confirmed by others (11). Our studies also demonstrated that neither the presence of rapidly turning over protein nor the de mu0 hormone-induced synthesis of additional protein is necessary for the response of the gene to the hormones. These results suggest that the primary effects of occupied receptors on gene transcription may be effectively studied using this system. n order to obtain insight into the mechanisms of action of these hormones, the temporal relationship between the occupancy of the nuclear-bound thyroid and glucocorticoid hormone receptors, the rate of GH gene transcription, and the level of GH mrna have been determined. MATERALS AND METHODS Cell Culture-GC rat pituitary tumor cells (1, 2) were cultured in monolayer using medium containing 10% fetal calf serum. Subconfluent cultures were hormonally deinduced for 7 days in medium containing 10% calf serum which had been rendered hypothyroid and hypoglucocorticoid using AG -X10 anion exchange resin and activated charcoal (12, 13). Hormone inductions were performed by adjusting the culture medium to 10 nm T3 and/or 1 p c dexamethasone ~ for the times indicated. Measurement of Occupied Thyroid and Glucocorticoid Hormone Nuelear-bound Receptors-Deinduced GC cells were incubated with The abbreviations used are: GH, growth hormone; T,, 3,5,3 - triiodo-l-thyronine.

2 12378 Hormonal Regulation of Growth Hormone Gene Expression either 10 nm L-['~]T~ (34 Ci/mmol) or 1 PM [1,2,4,6,7-3H]dexamethasone (77 Ci/mmol) with and without excess unlabeled hormone. Occupied nuclear-bound glucocorticoid receptors were measured as described by Raaka and Samuels (12). Thyroid hormone receptor occupancy was determined by a method communicated to us by Dr. Norman Eberhard (University of California, San Francisco). Cell cultures were incubated with a 250-fold excess of unlabeled competitor T3 and/or radioactive Ta for various lengths of time, washed 3 times with ice-colddulbecco's phosphate-buffered saline (Gibco), drained well, and lysed by the addition of ice-cold 10 mm Tris-HC1, ph 7.4, 10 mm NaCl, 3 mm MgC12, 15 mm B-mercaptoethanol, and 0.5% Nonidet P-40. Nuclei were sedimented by centrifugation at 1000 X g for 5 min at 2 'C and washed 3 times with the above buffer. The radioactivity present in the nuclear pellets was determined using a y counter. The number of receptors present per cell was calculated using the specific activity of the radioactive T3 and the amount of DNA present in the pellet, as determined by the method of Burton (14). The GC cells used were found to contain 12.5 & 0.3 pg of DNA per cell. Determination of Growth Hormone Gene Activity-GH gene activity was determined as previously described (10). Briefly, nuclei were isolated from deinduced or hormonally induced GC cells and the in vivo initiated RNA polymerase allowed to continue RNA syntheses in the presence of radioactive precursors. The radioactively labeled nuclear RNA was purified and hybridied to GH cdna sequences immobilied on nitrocellulose. After extensive washing and digestion with T and pancreatic RNase to remove unhybridied RNA, the fraction of the total radioactive transcripts derived from exonic GH gene sequences was measured by liquid scintillation counting. The values obtained were corrected for the overall efficiency of the hybridiation reaction (20 to 25%), as judged by hybridiation of a ['H GH-cRNA internal standard. Background hybridiation to pbr322- containing filters was 0.2 to 0.5 ppm in the experiments shown. Hybridiations contained input radioactivity of 15 to 60 X 10' cpm. Specifically hybridied radioactivity bound linearly with respect to input radioactivity over this range (data not shown). Determination of GH mrna Leuels-Growth hormone mrna levels were determined using purified cytoplasmic RNA either by kinetic hybridiation analysis (Ra analysis) as previously described (10) or by dot-blot hybridiation analysis, as described by White and Bancroft (15). Autoradiograms were densitometrically scanned, the slopes of the dot intensities of each dilution series calculated for the linear portion of the curves by linear regression, and the relative intensity of each sample calculated at a common fold dilution. An internal standard of known GH mrna concentration, determined by &-hybridiation analysis (lo), was included in each group of samples, and the GH mrna level per cell estimated by comparison to this standard. RESULTS Deinduction of Growth Hormone Gene Expression-The rat pituitary tumor cells used in these studies are cultured in medium containing 10% fetal calf serum. This serum contains significant amounts of both thyroid and glucocorticoid hormones. n order to study the effects of these hormones on GH gene expression, the cells were maintained first for 7 days in medium containing serum from which these hormones had been extracted (hypomedium). Under these conditions, Ts was lost rapidly from its receptor (Fig. 1). nterestingly, the loss of hormone from the receptor in the hypomedium was much slower than the rate of exchange of bound with unbound T3. Exchange of half of the bound T3 required 30 min, while in hypomedium loss of half of the receptor-bound TS required about 4 h. Thus, a significant fraction of the TB which dissociated from receptor was "recaptured" before leaving the cell. Glucocorticoid hormone exits from the nucleus completely within 3 h of hormone removal (12). The number of GH mrna molecules per cell declined rapidly in the absence of the hormones (Fig. 1). After just 24 h in hypomedium, GH mrna was reduced to 8% of the initial level, suggesting a rapid reduction in the transcription rate of the gene. Analysis of this data suggests that GH mrna has a maximum half-life of 6 to 8 h in the absence of hormones, DAYS FG. 1. The effect of thyroid and glucocorticoid hormone deprivation on GH mrna levels. GC cells were shifted from medium containing 10% fetal calf serum to medium containing 10% hypothyroid and hypoglucocorticoid fetal calf serum and GH mrna levels (0) and T8 receptor occupancy (0) measured. n uiuo exchange of radioactively labeled receptor-bound Ta with a 250 X excess of unlabeled T3 is also shown (A) FG. 2. -Thyroid hormone effects on growth hormone gene transcription, mrna level, and receptor occupancy. GC cells were treated with T3 for the indicated times and GH gene transcription rate (O), GH mrna level (O), and T3 receptor occupancy (A) determined. Hybridiations were performed in duplicate, and duplicates varied by less than 20%. assuming that 3 to 4 half-lives were required to reduce mrna levels 92%. The cells cease to divide after a few hours in hypomedium,' thus the cell doubling rate does not affect the apparent half-life of the mrna. After 7 days in hypomedium, GH mrna declined to less than 2% of its initial level. For this reason, the experiments reported here were performed using cells preincubated in hypomedium for 7 days. Relationship between Thyroid Hormone Receptor Occupancy and GH Gene Actiuity-Because growth hormone gene transcription responds to thyroid and glucocorticoid hormones in GC cells (O), it was possible to correlate hormone-induced changes in transcription rate with occupancy of the receptors for each hormone. Fig. 2 presents the occupancy of thyroid ' J. K. Nyborg, A. P. Nguyen, and S. R. Spindler, unpublished observations.

3 Hormonal Regulation of Growth Hormone Gene Expression hormone receptor and the transcription of the GH gene after T3 administration. At the hormone concentration used, the T3 receptor was 89% occupied 60 min after addition of hormone. Consistent with this level of occupancy, transcription increased to 87% of its maximum rate by this time. Analysis of the relationship between receptor occupancy and gene transcription rate indicates that they are directly proportional between 0 and 1 h of induction, indicating that a single occupied receptor is sufficient to activate transcription of the GH gene. The close correspondence between these parameters suggests gene activation by the receptor is very rapid, and thus the unoccupied T3 receptor may reside on or near the DNA site which is involved in regulation of the gene. A significant feature of the T3 response is the transient nature of the peak of gene transcription. n the experiment shown, after 4 h of hormone stimulation transcription declined to about 40% of the maximum rate attained. A corresponding decrease in receptor level was not detected, suggesting that the decrease was not related to a decline in the level of occupied receptor. The rate of the decay which occurred with time in T3- stimulated gene transcription was variable. Often it decreased significantly in only a few hours. However, occasionally it did not decline significantly until 8 h or more after hormone administration (see Fig. 6A). The reason for this variability is unknown. t did not correlate with the density of the cells, the length of time since subculturing of the cells, or the use of different batches of serum. Relationship between Glucocorticoid Hormone Receptor Occupancy and GH Gene Activity-The kinetics of the transcriptional effects of glucocorticoids were also studied. At the hormone concentration used, stable nuclear binding of the dexamethasone-receptor complex was maximal within 5 min after hormone addition (Fig. 3). However, maximum levels of transcription of the gene were not attained until 25 to 55 min later (also see Fig. 4). n these studies, the rate of transcription of the genebegan to decrease rapidly after reaching this maximum. Two hours after hormone addition, transcription decreased to a new level of activity varying between 1.5- and %fold of the uninduced level. No concomitant decrease was detected in the nuclear receptor level. Glucocorticoid enhancement of gene transcription and mrna accumulation was not reproducibly obtained in the absence of T3 (10). n approximately half of the experiments conducted using dexamethasone alone, little or no increase in transcription rate was detected. When such results were found, little change occurred in GH mrna levels until after 6 h of induction. Thereafter, a gradual 2-fold increase in mrna level was often observed, suggesting that enhanced mrna stability also may be involved in the glucocorticoid response. Since T3 is apparently important in obtaining a reproducible glucocorticoid transcriptional response, the effects of chronic T3 stimulation on the kinetics of the glucocorticoid transcriptional response were examined (Fig. 4). When dexamethasone was added to cells which had been pretreated for 12 h with T3, there was a rapid increase in transcription rate, followed by a rapid decrease to a steady-state rate of transcription which was 2 to 3 times that found previously. No change in the kinetics of the induction was found (Fig. 3). However, the transcriptional response was always observed under these conditions. The Effects of Simultaneous Administration of T3 and Dexamethasone on GH Gene Expression-The kinetics of the response of the gene to the simultaneous administration of both hormones appear to be a composite of the effects of each hormone alone (Fig. 5). A peak of gene activity resembling the early gene response to dexamethasone was found after 1 h of induction, and a second peak of transcriptional activity resembling that induced by T3 alone occurred after 2 h. n this experiment, gene activity decreased significantly after only 3 h of induction, although in other similar experiments gene activity decreased more slowly (Fig. 6B). The reason for this variability is unknown, but it is not related to cell density, growth rate of the cells, or the serum batches used. n the absence of protein synthesis, the kinetics of the gene response to TS and dexamethasone were qualitatively very similar to the response which occurred in their presence (data not shown). However, there was at least a 2-fold enhancement of the peak of T3 or dexamethasone and T3-stimulated gene activity when cycloheximide was present 30 min before and during the hormone induction (data not shown). Such an increase was noted after 4 h of induction with either or both hormones (lo), and these data demonstrate that it is not an artifact of altered kinetics of gene induction as had been proposed. Accumulation of GH mrna in Response to Hormones- - c G t- E4 u V cn a u - W2 W FG. 3. Glucocorticoid hormone effects on growth hormone gene activity, mrna level, and receptor occupancy. GC cells were hormonally induced for the indicated times with dexamethasone, and GH gene transcription rate (O), GH mrna level (O), and glucocorticoid receptor occupancy (A) determined. - 0 O FG. 4. Glucocorticoid effects on the rate of CH gene transcription in cells chronically treated with Ts. GC cells were deinduced for 7 days in hypomedium and treated for 12 h with T,. Dexamethasone was added to the cultures, and the GH gene transcription rate (0) and mrna level (0) were determined at the indicated times.

4 12380 Hormonal Regulation of Growth Hormone Gene Expression f \ FG. 5. GH gene transcription rate after the simultaneous addition of dexamethasone and Ts to hypomedium-deinduced cells. Hormones were administered to GC cells, and GH gene transcription rate (0) and GH mrna level (0) measured at the indicated times DAYS FG. 6. Kinetics of GH gene activation and mrna accumulation after hormone addition to deinduced GC cells. Shown are T3 addition (panel A) and simultaneous T3 and dexamethasone addition (panel B). GH gene transcription rate (0) and GH mrna level (0) are depicted. After the addition of T3 and/or dexamethasone to deinduced cells, GH mrna levels were closely correlated with the rate of transcription of the gene (Figs. 2-6). Despite the reduction in the rate of transcription of the gene 1 day after T3 or dexamethasone and T administration, GH mrna remained at relatively high levels until at least 3 days after hormone addition. Since the division time of the cells averages about 36 and 72 h in the presence of T3 and of T3 and dexamethasone, respectively, these data suggest that the mrna is very stable in the presence of T3. This stability is in marked contrast to the apparent instability of the mrna after shift of the cells into hormone-free medium (Fig. 1). DSCUSSON The data presented here indicate that the occupancy of the T3 receptor is closely linked temporally to the transcriptional activity of the GH gene. The T3 receptor is a chromosomal protein in both the presence and absence of the hormone, and the rapidity of the gene response suggests the receptor resides on or near the site at which it activates gene expression. The stable binding of the occupied glucocorticoid receptor to its nuclear acceptor sites is apparently much more rapid than the rate of GH gene activation, suggesting the existence of at least one rate-limiting step before activation of the gene. Unoccupied steroid hormone receptors apparently are nuclear proteins which translocate to acceptor sites after hormone binding (16, 17). The lag between stable nuclear binding of the occupied receptor and gene activation apparently occurs distal to this step. The lag is not related to the synthesis of new proteins, since dexamethasone activates the gene in the absence ofnew protein synthesis (10). n addition, the lag probably is not related to the transit time required for the polymerase to transcribe the gene. Assuming an in uiuo transcription rate of 30 nucleotides/s, the entire GH gene could be transcribed in about 1 min. The lag may involve modification of the receptor, association of the receptor with a low abundance chromosomal protein, and/or association of the receptor with specific gene regulatory site(s). The nature of the gene regions involved in the hormone responsiveness of the GH gene is at present undefined. However, the glucocorticoid-responsive elements found upstream from the start of transcription of the proviral form of mouse mammary tumor virus (18) appear to share a degenerate hexanucleotide consensus sequence, AGA4CA(G)$, at the sites of hormone-receptor binding (19, 20) which is homologous to sequences present 5' to the cap site of the rat GH gene. These sequences are centered at -155 (identical in 7 of 8 nucleotides), -177 (identical in 7 of 8 nucleotides), and -321 (identical in 6 of 7 nucleotides), suggesting that these sites may be involved in the glucocorticoid responsiveness of this gene.2 The transcriptional synergism of the two hormones suggests that their occupied receptors may interact directly when bound to their regulatory sites. t is possible that reproducible and stable association of the occupied glucocorticoid receptor to the putative regulatory sites near the GH gene may require the presence of occupied thyroid hormone receptor. This stabiliing interaction maybe the reason for the variable nature of the glucocorticoid response. The simplest model would predict that the T3 regulatory region should be adjacent to the glucocorticoid-receptor-binding site of the gene. The rapid attenuation of both the dexamethasone- and T3- induced peaks of GH gene transcription suggests that the hormone-receptor complexes undergo a time-dependent modification, association with accessory regulatory proteins, or movement to a second regulatory site which renders the gene less efficiently transcribed. The receptor may be displaced from the regulatory site by the transcription reaction, with an attendant time-dependent decrease in initiation rate. t appears unlikely that a general depletion of receptor is responsible for the reduced transcription rate, since changes in receptor number later in the induction do not correlate with changes in the transcription rate of the gene. GH mrna is apparently much more stable when thyroid or glucocorticoid hormones are present. This observation draws into question the conclusions of several studies of the glucocorticoid responsiveness of transvected GH genes. A 2- to %fold glucocorticoid enhancement of GH mrna levels was interpreted as evidence for transcriptional regulation of the genes (21,22). t now appears possible that post-transcriptional effects of the hormones on GH mrna stability could explain the results reported. However, we are now using the detailed description of the transcriptional response of the genes presented here to evaluate the gene regions involved in each of the complex steps of the hormone responses observed.

5 Hormonal Regulation of Growth Hormone Gene Expression n summary, the results presented suggest that: 1) T3 receptor resides on or near the DNA site from which it acts to modulate GH gene expression; 2) only one occupied TS receptor is required to activate transcription of the gene; 3) the occupied glucocorticoid receptor requires a relatively slow molecular step before it mediates activation of GH gene transcription; 4) the receptors apparently act independently but interact physically at their regulatory sites; 5) the rate of hormone-induced gene transcription is attenuated by protein synthesis-independent processes during induction of the gene; and 6) the hormones have post-transcriptional effects which apparently stabilie GH mrna. REFERENCES 1. Tashjian, A. H., Jr., Yasumura, Y., Levine, L., Sato, G. H., and Parker, M. L. (1968) Endocrinology 82, Tashjian, A. H., Jr., Bancroft, F. C., and Levine, L. (1970) J. Cell BWl Bancroft, F. C., Levine, L., and Tashjian, A. H., Jr. (1969) J. Cell BWl. 43, Shapiro, L. E., Samuels, H. H., and Yaffee, B. M. (1978) Proc. Natl. Acad. Sci. U. S. A. 75, Martial, J. A., Baxter, J. D., Goodman, H. M., and Seeburg, P. H. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, Martial, J. A,, Seeburg, P. H., Gueni, D., Goodman, H. M., and Baxter, J. D. (1979) Proc. Natl. Acad. Sci. U. S. A. 74, Seo, H., Vassart, G., Brocas, H., and Refetoff, S. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, Wegne, M., Schachter, B. S., Baxter, J. D., and Martial, J. A. (1982) DNA (N. Y.) 1, Dobner, P. R., Kawasaki, E. S., Yu, L-Y., and Bancroft, F. C. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, Spindler, S. R., Mellon, S. H., and Baxter, J. D. (1982) J. Bwl. Chem. 267, Evans, R. M., Birnberg, W. C., and Rosenfeld, M. G. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, Raaka, B. M., and Samuels, H. H. (1983) J. BWl. Chem. 268, Samuels, H. H., Stanley, F., and Casanova, J. (1979) Endocn nol- Ogy 106, Burton, K. (1956) Biochem. J. 62, White, B.A., and Bancroft, F. C. (1982) J. BWl. Chem. 257, Grady, W. W., Schrader, W. T., and O Malley,B. W. (1982) Endocr. Rev. 4, Welshons, W. V., Leiberman, M. E., and Gorski, J. (1984) Nature (Lond.) 307, Chandler, V. L., Maler, B. A., and Yamamoto, K. R. (1983) Cell 34, Payvar, F., DeFranco, D., Firestone, G. L., Edger, B., Wange, o., Okret, S., Gustafsson, J.-A., and Yamamoto, K. R. (1983) Cell 35, Scheidereit, C., Geisse, S., Westphal, H. M., and Beato, M. (1983) Nature (Lord.) 304, Doechmer, J., Barinaga, M., Vale, W., Rosenfeld, M. G., Verma,. M., and Evans, R. M. (1982) Proc. Natl. Acad. Sei. U. S. A. 79, Robins, D. M., Paek,., Seeburg, P. H., and Axel, R. (1982) Cell 29,