Proc. Nati. Acad. Sci. U.SA
Vol. 74, No. 5, pp. 2054-20581, May 1977
Cell Biology
Triiodothyronine stimulates specifically growth hormone mRNA in rat pituitary tumor cells* (thyroid hormone/somatotropin/translation in cell-free system/mechanism of hormone action)
HISAO SEOt, GILBERT VASSARTt, HUGETTE BROCASt, AND SAMUEL REFETOFFt University of Chicago, School of Medicine, Chicago, Illinois 60637; and * Institute de Recherche Interdisciplinaire, School of Medicine, Free University of Brussels, 1000 Brussels, Belgium
t Thyroid Study Unit,
Communicated by Elwood V. Jensen, March 10, 1977
In a cell-free protein-synthesizing system from a rabbit reticulocyte lysate, total RNA extracted from cultured rat pituitary tumor (GH3) cells directed, in a dose-related manner, the synthesis of proteins that were precipitated by antisera specific to rat growth hormone (somatotropin) and rat prolactin. A marked decrease in growth hormone secretion and growth hormone mRNA activity was observed when cells were grown in a medium deficient in thyroid hormone. Addition of triiodothyronine in physiologic amounts both prevented and completely reversed this effect within 48 hr. Thyroid hormone had no effect on prolactin secretion or prolactin mRNA activity. These data suggest that thyroid hormone may stimulate synthesis of growth hormone throug induction of transcriptional activity. The possibility of an additional effect at the posttranscriptional level has not been excluded. Although thyroid hormone is believed to have a general effect on a variety of metabolic processes, some effects, at the molecular level, may be quite selective, as indicated by the observed changes in growth hormone but not prolactin mRNA activity. The GH3 cell model is useful in the study of triiodothyronine action because of independence from secondary hormonal effects caused by hypothyroidism and because simultaneous measurement of prolactin mRNA activity serves as a unique internal control.
ABSTRACT
The demonstration of triiodothyronine (T3) binding to nuclear proteins raises the possibility that thyroid hormone may regulate gene expression. Earlier work from Tata's laboratory showed that thyroid hormone-induced protein synthesis was preceded by formation of new RNA (1). Later, DeGroot et al. and Dilman et al. showed increase in the poly(A)-rich fraction of RNA (2, 3). Demonstrations of stimulation of a specific mRNA by thyroid hormones were recently provided by Kurtz et al. (4) and by Roy et al. (5) for a2u-globulin in the rat. However, because a number of hormones are known to stimulate the synthesis of this protein (6) and because thyroid hormone profoundly alters the level of such hormones (7), experiments done in the whole animal do not provide sufficient evidence for the direct induction of -d2u-globulin mRNA by thyroid hormone. Samuels et al. have reported a quantitative correlation between nuclear T3 receptor occupancy and stimulation of growth hormone (GH, somatotropin) synthesis in cultured rat pituitary cells (8). We have adopted a similar experimental model to study more directly the action of thyroid hormone on the induction of a specific mRNA. In a recent report, we have shown that GH and prolactin mRNA activities in the total RNA extracted from a rat pituitary cell line (GH3) that actively synthesizes both hormones can be quantitated using a rabbit reticulocyte lysate cell-free system (9). In this report, we show that Abbreviations: T3, triiodothyronine; GH, growth hormone (somatotropin); Tx, thyroidectomized; N, normal; NRS, normal rabbit serum; NMS, normal monkey serum; MEM, modified Eagle's medium. * Presented in part at the Fifty-Second Meeting of the American Thyroid Association, Inc., Toronto, Canada, September 15-18, 1976.
the T3-stimulated GH but not prolactin synthesis is accompanied by a specific increase in GH mRNA but not prolactin mRNA activity. Thus, the known effect of T3 on GH production may predominantly involve a selective stimulation of transcriptional activity. METHODS
Cell Culture. Rat pituitary tumor cells originating from the GH3 line were a gift from A. H. Tashjian. This line, derived from a single MtTW5 rat pituitary tumor cell, synthesizes and secretes both GH and PRL (10). Except for the medium used, the cells were propagated in culture as previously described (11). Cells were maintained in modified Eagle's medium (MEM) containing 10% fetal calf serum. For the experiments, cells were grown in MEM supplemented with 10% serum from either normal (N) or thyroidectomized (Tx) rats, with or without added T3. Each rat received 100,uCi of Na131I on the second day following surgical thyroidectomy and serum was harvested 4 weeks later. Concentrations of endogenous thyroxine and T3 in the normal rat serum were 5.4 ,ug/100 ml and 60 ng/100 ml, respectively, and in the Tx rat serum were 0.6 ,g/100 ml and less than 20 ng/100 ml, respectively. RNA Extraction. Each RNA preparation was obtained from 5 to 15 petri dishes or 30 to 70 X 106 cells treated in the same manner. The medium was removed and saved for hormone determination. The cells adhering to the bottom of the petri dishes were washed once with 3 ml of MEM, then harvested by gentle resuspension in the same medium, using a pasteur pipette. A 1 ml aliquot of the pooled 10 ml cell suspension was saved for total RNA (12) and DNA (13) determination and the remainder was centrifuged for 10 min at 600 X g. The cell pellet was resuspended in 1 ml of buffer containing 25 mM MgCl2/50 mM KCI/200 mM sucrose/200 mM Tris-HCl, pH 8.5. In a rapid succession 7 ml of buffer containing 1% sodium dodecyl sulfate/100 mM NaCl/5 mM EDTA/0.02% heparin/10 mM Tris, pH 8.5, was added, followed by 8 ml of phenol/chloroform (1:1 vol/vol) presaturated with a buffer containing 100 mM NaCl/1 mM EDTA/10 mM Tris. Extraction was carried out by manual agitation for 5 min. Following centrifugation for 10 min at 9000 X g the aqueous phase was removed, and the organic phase containing the denatured protein interphase was re-extracted after addition of 8 ml of the suspension buffer. The two aqueous phases were pooled and extracted four consecutive times with an equal volume of phenol/chloroform. During the final extraction no denatured protein interphase was present. Residual phenol in the aqueous phase was removed by a single extraction with chloroform, following which LiCl was added to a final concentration of 2 M. The RNA was allowed to precipitate at 4° overnight. After centrifugation for 10 min at 9000 X g the RNA pellet was 2054
Cell Biology: Seo, et al.
Proc. Nati. Acad. Sci. USA 74 (1977)
2055
Table 1. Radioactivity in O-RNA and in immunologic blanks
Radioactivity RNA,
Aug/ml 0 o 200 0 0 200
Specific antiserum (1st antibody)*
Normal carrier serumt
Anti-IgG serum (2nd
Monkey anti-rat GH
antibody)t
in precipitate, mean cpm ± range
Goat anti-monkey IgG Goat anti-monkey IgG Goat anti-monkey IgG Goat anti-rabbit IgG Goat anti-rabbit IgG Goat anti-rabbit IgG
NMS NMS
Rabbit anti-rat prolactin NRS NRS
1813 ± 1953 ± 1933 ± 913 ± 1058 ± 960 ±
187 17 33 13 167 16
For details in preparation, see Methods. Volume and dilution of sera added to 200 or 250 ,. of lysate are given in the footnotes. Acid-precipitable radioactivity was 16.76 X 106 cpm. * Monkey anti-rat GH: 5 Al (1:50); rabbit anti-rat prolactin: 5 pl (1:25). t NMS = normal monkey serum: 5 Ml (1:50); NRS = normal rabbit serum: 5 Ml (1:5). 1 Goat anti-monkey IgG: 25 Al (undiluted); goat anti-rabbit IgG: 5 Al (undiluted).
washed two times with cold 2 M LiCl followed by a single wash with 66% (vol/vol) ethanol in 0.3 M NaCl. The RNA was dissolved in water and absorbance was read at 260 and 280 nm. The ratio in all preparations ranged from 1.9 to 2.1. The amount of RNA was calculated assuming that 1 A260 = 40,ug. After an ethanol precipitation, the final RNA preparation was dissolved in water to the desired concentretion to be used for translation.
Cell-Free Protein Synthesis. Translation was carried out in rabbit reticulocyte lysate cell-free system as described by Palmiter et al. (14). After incubation for 1 hr at room temperature with RNA and [3H]leucine (Amersham, 53 Ci/mmol) at 170 MCi/ml, the reaction was stopped by the addition of 0.1 ml/ml of reaction mixture of a solution containing the following additives and bringing their concentrations to 0.02 M sodium phosphate buffer and 0.15 M NaCl, 1 % (vol/vol) Triton X-100, 1% sodium deoxycholate, and 0.10% unlabeled L-leucine, pH 7.5. The mixture was then centrifuged for 60 min at 100,000 X g in a Beckman rotor SW 60. The supernatant was used for trichloroacetic acid and immunoprecipitations described below. Blank reaction mixtures contained all additives except for RNA, which was substituted by an equal volume of water. Trichloroacetic Acid Precipitation. Five microliter aliquots of the centrifuged lysate were spotted onto Whatman GF/A filter disks and processed as described by Schimke et al. (15). Immunoprecipitation. GH and prolactin synthesis in the lysate was measured by double antibody precipitation. Antibodies against rat GH and rat prolactin were a gift from A. Parlow and second antibodies, anti-rabbit and anti-monkey IgG, were from V. S. Fang. Duplicate aliquots containing 200-250 Al of the lysate were incubated with 5,l of diluted first antibody for 1 hr at room temperature, and for 16 hr at 40, followed by incubation for 30 min at room temperature and for 8 hr at 40 with the appropriate second antibody (see legend to Table 1). The precipitate was centrifuged, washed with 1 ml of phosphate-buffered saline, resuspended, and cleaned by centrifugation in Microfuge tubes through 1 M sucrose as described by Rhoads et al. (16) and dissolved in 1 ml NCS (Amersham) prior to counting. The radioactivity in lysates incubated without added RNA (0-RNA blanks) and in immunoprecipitates in the presence of normal rabbit or monkey sera, rather than the specific antisera (immunologic blanks) was similar and constituted from 0.004 to 0.020% of the acid-precipitable radioactivity (Table 1). Specific immunoprecipitable radioactivity was calculated by subtraction of the appropriate blank. Precipitation with antiprolactin and anti-GH serum was carried
a
out on separate aliquots of lysate or on the same lysate in sequence. Either method achieved excellent recovery and reproducibility (Table 2). The specificity of the antibodies was evaluated by their ability to precipitate purified authentic rat GH and prolactin labeled with 125I. The antiprolactin serum was specific, and cross reaction of the anti-GH serum with prolactin was eliminated by addition of 1 ,ug of prolactin without affecting its reactivity with GH (Fig. 1). Furthermore, as previously shown by comigration of the immunoprecipitates of authentic hormones and cell-free translated material on sodium dodecyl sulfate/polyacrylamide gel electrophoresis (9), 95% of the radioactivity appeared as sharp single peaks. In accordance with findings by Sussman et al. (17), Evans et al. (18), and Maurer et al. (19), the hormones translated in the cell-free system were of slightly larger molecular weights than the authentic hormones. Quantitation of GH and Prolactin Concentration in Culture Media. GH and prolactin concentrations in the cell culture media were determined by radioimmunoassays as previously described (11). All media were changed 48 hr prior to termination of the experiments. Thus, concentration of hormones represent the cumulative secretion by the cells into the medium over 48 hr. The cells do not store significant amounts of hormone (11) and the rate of hormonal degradation has not been determined. Table 2. Recovery with single and sequential
immunoprecipitations 1st immunoprecipitation
Specific antiserum
2nd immunoprecipitation
Specific cpm* in precipitate
Specific antiserum
16,815 + 30
Anti-GH Antiprolactin
Specific cpm* in precipitate
Anti-
prolactin Antiprolactin Anti-GH
16,815 ± 30 7,514 ± 332
8,479
±
90
393 ± 30
Lysate was programmed with 200 Mug of RNA per ml of reaction mixture. Acid-precipitable radioactivity was 24.93 X 106 and 26.58 X 106 cpm in lysates with or without added RNA, respectively. Immunologic blanks on the 1st and 2nd immunoprecipitation had, respectively, 954 I 17 and 978 ± 45 cpm with NRS and 2223 ± 201 and 1915 I 232 cpm with NMS. 0-RNA blanks precipitated with antiprolactin and anti-GH serum had 1461 + 67 and 2754 1 46 cpm, respectively. All data are expressed as mean ± range. * cpm in immunoprecipitates minus blanks.
Proc. Nati. Acad. Sci. USA 74 (1977)
Cell Biology: Seo, et al.
2056
125l-Prolactin
_
x 103
96 14-
100r
I
7
V
C,
801
r-12-
-
C
D
.SCu
60
0 Q
86
O 40 E E
a
4-
E E
2-
0
201
.O
L.
NRS anti-Prolactin
L-
Unlabeled 0 prolactin, ng/tube
500
500
L610
1 000
E
C3
0 NMS
anti-GH
FIG. 1. Specificity of antisera to rat prolactin and GH. Antisera were incubated with 125I-labeled purified rat prolactin (1251-prolactin) or GH (1251-GH) under the standard reaction conditions described in the Methods section. Antiprolactin reacted only with 125I-prolactin. The amount of precipitable 125I-GH was similar to the NRS blank. In the absence of unlabeled prolactin, anti-GH reacted equally with 125I-GH and 1251-prolactin. Addition of 1000 ng of unlabeled prolactin reduced the immunoprecipitable 1251-prolactin to the level of the NMS blank without affecting the precipitability of 125I-GH. For abbreviations see legend to Table 1.
RESULTS Application of the Reticulocyte Lysate System to Quantitate GH and Prolactin mRNA Activity. We have previously shown a linear dose response between the radioactivity in GH immunoprecipitates and the input of RNA extracted from rat pituitary tumors transplanted in vvo. The linear dose response extended over the range of 10-300 Asg of RNA per ml of lysate (9). Results from a similar study conducted with total RNA extracted from the same cells grown in culture are shown in Fig. 2. There was a dose-dependent increase in both prolactin and GH immunoprecipitable radioactivity. Depending upon the lysate preparation used, 50-60% of the total radioactivity was acid-precipitable in the absence of exogenous RNA. With the addition of increasing amounts of RNA, and depending upon the lysate preparation, the acid-precipitable radioactivity usually but not always declined to a minimum of 25% when the largest amount of 200 ,g of RNA per ml of reticulocyte lysate was added. Depending upon the source and quantity of RNA added, 0.023-0.210% of the acid-precipitable radioactivity was immunoprecipitable with anti-GH or antiprolactin sera. Immunologic and 0-RNA blanks had 0.004-0.020% of the acidprecipitable radioactivity. Effect of Thyroid Hormone Deprivation on the Secretion and mRNA Activity of GH. In this study, cells plated in MEM containing fetal calf serum were divided into three groups of 15 plates. Medium in the first group was replaced with MEM supplemented with 10% normal rat serum; in the second group, with MEM containing 10% Tx rat serum; and in the third group the 10% Tx rat serum added to MEM was enriched with 80 ng of T3 per 100 ml to bring thyroid hormone concentration to a physiologic level. Following 6 days incubation in these conditions, media were collected, cells were harvested, RNA was extracted, and mRNA activity was measured. Results are shown in Fig. 3. Culture in conditions of thyroid hormone deficiency profoundly depressed GH secretion. This was prevented by the addition of T3 to the thyroid hormone-deficient serum. There
0Total RNA, pg/ml
FIG. 2. Relation of RNA input to cell-free synthesis of prolactin and GH. Total RNA extracted from cultured rat pituitary tumor cells was added to rabbit reticulocyte lysate cell-free protein-synthesizing system to yield the final concentration indicated in the abscissa.
were no changes in prolactin secretion. A corresponding decrease in GH mRNA activity was observed in cells grown in media supplemented with thyroid hormone-poor rat serum. Addition of T3 to the Tx rat serum prevented the decrease in GH mRNA activity. The total amount of RNA per cell was not affected by the conditions of culture. RNA to DNA ratios expressed as mean ± range in two experiments were as follows: 2.4 ± 0.3 for control cells grown in normal rat serum, 2.5 + 0.3 for cells grown in Tx serum, and 2.7 + 0.2 for cells grown in Tx serum supplemented with T3. To eliminate the possibility that decrease in GH mRNA activity observed in cells grown in Tx rat serum was due to inhibition of mRNA translation in the cell-free system, the relation of RNA input to GH synthesis in the lysate was studied. As shown in Fig. 4, the proportionality between RNA input and GH synthesis was preserved in the three RNA preparations. GH mRNA activity
GH and prolactin secretion into medium
= -
m
200
a z
160'
m
120
a
jL
In
GH Prolactin
r-4,
AI
10 x
0i
cm C C
T
E
4-
a CD n
3-
O
2-
80
a 0
~0c,
C
co 0 0
E E
b.
v-
N
I-
O0
Tx
N
Tx
TS
FIG. 3. Effect of thyroid hormone on GH and prolactin synthesis and on the GH mRNA activity. Rat pituitary tumor cells were cultured in media supplemented with 10% of thyroidectomized rat serum (Tx), normal rat serum (N), or thyroidectomized rat serum with added T3 (Tx + T3). Endogenous GH and prolactin synthesis, estimated from the amount secreted into the medium, and the GH mRNA activity were measured. Data are expressed as mean i range of duplicate determinations. For methodologic details see text.
Cell Biology: Seo et al.
Proc. Natl. Acad. Sci. USA 74 (1977) +T3 (80ng/dO)
2057
Table 3. Effect of T3 on GH and prolactin secretion and on mRNA activity
E 0
Ratio, mean ± range
N
3-
a)
IV
Hormone Measurement
la 0
GH
C
Tx N
Tx + T3 N
0.24 ± 0.16
0.82 ± 0.37
Secretion,
ng/gg of
E
DNA per 48 hr mRNA ac0
100 2 '00 Total RNA, jig /ml
FIG. 4. Immunoreactive GH in lysates programmed with increasing amounts of RNA derived from cells cultured under various conditions. For abbreviations see legend to Fig. 3. Methodological details appear in the text. Culture, treatment, and translation assays using the same reticulocyte lysate were carried out simultaneously.
Also, there was no evidence for inhibition of endogenous hemoglobin synthesis by the lysate, because acid-precipitable radioactivity was identical in all reactions with same RNA input, irrespective of its origin. Thus, quantitation of the mRNA activity most likely reflects mRNA content. Reversal of the Effects of Thyroid Hormone Deprivation by Addition of T3. In this experiment, cells were allowed to grow for 8 days in MEM containing 10% Tx rat serum, following which 80 ng of T3 per 100 ml of Tx rat serum was added and culture was continued for 2 more days. Control plates were grown in the presence of either 10% normal rat serum or Tx rat serum for the periods of 8 and 10 days. Results of GH and prolactin secretion into the media and the respective mRNA activities are shown in Table 3, and are for clarity expressed as the ratio of Tx/normal or Tx + T3/normal. As compared to normal, a 4-fold depression in GH secretion and 16-fold decrease in GH mRNA activity was found in cells grown in Tx rat serum for the entire period of incubation. Culture for 2 days after addition of T3 was sufficient to normalize both GH secretion and GH mRNA activity. Irrespective of hormonal treatment, there were no significant changes in prolactin secretion or prolactin mRNA activity. DISCUSSION Thyroid hormone is known to affect a variety of metabolic processes. While these effects are well recognized, the mechanism by which thyroid hormone influences these processes has remained largely a matter of speculation (20). The recent demonstration of a specific nuclear T3 receptor in various tissues (21-23), coupled with work indicating that treatment with thyroid hormone increases polymerase (24) or chromatin template activity (25) and the poly(A)-rich fraction of RNA (2, 3), suggests that thyroid hormone action may involve changes at the transcriptional level. The demonstration that the hormone induces specific mRNA species, coding for proteins known to be under hormonal regulation, would show how specificity is conferred and provide a more direct proof for the mechanism of thyroid horomone action. Such evidence appears to have been provided by Kurtz et al. (4) and by Roy et al. (5), who showed that treatment of male hypothyroid rats for 7-8 days with pharmacologic doses of thyroxine restored to normal both urinary excretion of and mRNA activity for the hepatic protein a2u-globulin. However, the interpretation of these experiments remains complicated by the fact that the synthesis of a2u-
tivity, cpm in immunopre-
cipitate Prolactin Secretion, ng/gg of DNA per 48 hr mRNA activ-
0.062 ± 0.021 0.84 ± 0.132
1.2 ± 0.05
1.4 ± 0.34
1.4 ± 0.08
1.3 ± 0.13
ity, cpm in immunopre-
cipitate
For abbreviations and experimental design, see text.
globulin is subject to influences by a variety of hormones of
ponthyroidal origin, such as growth hormone (6), the level of
which is known to be profoundly affected by thyroid hormone (7). In the present work we have shown that thyroid hormone deprivation decreases both the synthesis and mRNA activity of GH in rat pituitary cells grown in culture. The effect was not due to other metabolic alterations caused by hypothyroidism, because addition of thyroid hormone to the medium containing hypothyroid rat serum prevented the occurrence of the changes seen when hypothyroid rat serum was used alone (Fig. 3). The effect could be completely reversed within 48 hr following the addition to the medium of physiologic concentrations of T3. In preliminary experiments, induction of synthesis and mRNA activity of GH was observed as early as 6 hr after addition of T3 to cells cultured in Tx rat serum. Furthermore, it appears that the effect of thyroid hormone is selective on GH, because in the same cell preparations no changes occurred in prolactin synthesis or mRNA activity (Table 3). Data presented in abstract form by Martial et al. (26) are in agreeme..t with our observation of T3-induced GH mRNA activity, in another cell line of rat pituitary tumors. Because this cell line possesses T3 nuclear receptors and their ability to synthesize GH is quantitatively related to nuclear T3-receptor occupancy (8), the demonstration of selective induction of GH mRNA suggests that expression of thyroid hormone action requires the following sequence of events: hormone penetration within the cell, binding to nuclear receptor proteins, induction of transcriptional activity, synthesis of specific protein, and finally the metabolic effects induced by the latter. Although, in all experiments, the thyroid hormone-induced changes in the amount of GH were always associated with corresponding changes in GH mRNA activity, direct proportionality between medium GH and cell GH mRNA activity could not always be obtained. For example, while data presented in Fig. 3 show an average of 7.5-fold decrease in both GH synthesis and mRNA activity with thyroid hormone deprivation, in Table 3, under similar circumstances, an average
2058
Cell Biology: Seo et al.
of 4-fold decrease in GH synthesis was associated with a 16-fold decrease in GH mRNA activity. While this discrepancy may be methodologic in origin, the possibility of an additional effect of thyroid hormone at the posttranscriptional level cannot be excluded (27-29). The following methodologic problems should be considered. The endogenous synthesis of GH and prolactin was estimated from secretion rate over 48 hr, in turn calculated from the hormonal concentration in the medium and cell number at the termination of the experiment. Rates of cell division and GH secretion may vary over 48 hr under different culture conditions. The rate of hormonal degradation has not been taken into account. On the other hand, although the intracellular hormone content was not measured, at 48 hr it represents less than 3% of the medium content. From the point of view of mRNA quantitation, the turnover rate of the messenger and the residence time of T3 at the receptor site should be considered. Acute induction of thyroid hormone deprivation causes a very gradual decline in GH mRNA activity over at least 6 and possibly 10 days. Furthermore, it is practically impossible to prepare a serum totally devoid of thyroid hormone. Attempts to extract the hormone modify the serum composition in a variety of other substances and serumless media are inadequate for the period of culture extending beyond 2 days that is necessary to induce a state of thyroid hormone deprivation. Thus, under the present experimental conditions, it is impossible to determine whether thyroid hormone may not indeed switch de novo the GH gene. We have eliminated the possibility of inhibition of mRNA translation in the cell-free system by showing a good dose response with various RNA preparations, and the preservation of prolactin mRNA activity in preparations with virtually no GH mRNA activity. In fact, the former serves as a unique internal control for the relative activity of the cell-free protein-synthesizing system, for possible errors from extraneous factors affecting cell-free translation, for mRNA recovery, and for the relative biologic quality of various RNA preparations tested. The authors are grateful to Dr. Armen H. Tashjian, Jr. for providing the GHS cell line, to Dr. Albert Parlow for making available rat GH and prolactin antisera, to Dr. Victor S. Fang for the anti-rabbit and anti-monkey IgG sera, to Dr. Gerald Burke for the T3 antiserum, to the National Pituitary Agency for GH and prolactin radioimmunoassay kits, and to Dr. Ruben Matalon for assistance and advice in cell culture
methodology. The technical assistance of Mrs. Ofelia Gomez and Mr. Swen R. Hagen, and the secretarial help of Mrs. Yolanda W. Richmond are also acknowledged. This work was supported in part by U.S. Public Health Service Grant AM-15070. G.V. is Charg6 de Recherche at the Belgian Fonds National de la Recherche Scientifique. His travel expenses were supported through a generous gift from Mr. and Mrs. Harry Katz. The costs of publication of this article were defrayed in part by the payment of page charges from funds made available to support the research which is the subject of the article. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
Proc. Natl. Acad. Sci. USA 74 (1977) 1. Tata, J. R. (1969) Gen. Comp. Endocrinol. 2,385-397. 2. DeGroot, L. J. & Rue, P. A. (1976) Fifth International Congress of Endocrinology, Hamburg, Germany, Abstr. no. 578. 3. Dillman, W. H., Mendecki, J., Koerner, D. W. & Oppenheimer, J. H. (1976) Fifth International Congress of Endocrinology, Hamburg, Germany, Abstr. no. 581. 4. Kurtz, D. T., Sippel, A. S. & Feigelson, P. (1976) Biochemistry 15, 1031-1036. 5. Roy, A. K., Schiop, M. J. & Dowbenko, D. J. (1976) FEBS Lett. 64,396-399. 6. Roy, A. K. (1973) J. Endocrinol. 56,295-301. 7. Hervas, F., Morreale de Escobar, G. & Escobar del Rey, F. (1975)
Endocrinology 97,91-101. 8. Samuels, H. H., Shapiro, L. E. & Tsai, J. S. (1976) 58th Annual meeting of the Endocrine Society, San Francisco, California, Abstr. no. 266. 9. Brocas, H., Seo, H., Refetoff, S. & Vassart, G. (1976) FEBS Lett. 70, 175-179. 10. Tashjian, A. H., Jr., Bancroft, F. C. & Levine, L. (1970) J. Cell Biol. 47,61-70. 11. Seo, H., Refetoff, S. & Fang, V. S. (1977) Endocrinology 100, 216-226. 12. Munro, M. N. & Fleck, A. D. (1966) in Methods of Biochemical Analysis, ed. Glick, D. (Interscience Publishers, New York), Vol. 14, pp. 113-176. 13. Burton, K. (1956) Biochem. J. 62, 315-323. 14. Palmiter, R. D., Oka, T. & Schimke, R. T. (1973) J. Biol. Chem. 248,2031-2039. 15. Schimke, R. T., Rhoads, R. E. & McKnight, G. S. (1974) in Methods in Enzymology, eds. Moldave, K. & Grossman, L. (Academic Press, New York), Vol. 30, pp. 694-701. 16. Rhoads, R. E., McKnight, G. S. & Schimke, R. T. (1973) J. Biol. Chem. 248, 2031-2039. 17. Sussman, P. M., Tushinski, R. J. & Bancroft, F. C. (1976) Proc. Natl. Acad. Sci. USA 73,29-33. 18. Evans, G. A. & Rosenfeld, M. G. (1976) J. Biol. Chem. 251, 2842-2847. 19. Maurer, R. A., Stone, R. & Gorski, J. (1976) J. Biol. Chem. 251, 2801-2807. 20. Bernal, J. & Refetoff, S. (1977) Clin. Endocrinol., in press. 21. Oppenheimer, J. H., Schwartz, H. L. & Surks, M. I. (1972) J. Clin.
Endocrinol. Metab. 35, 330-33.
22. Samuels, H. H. & Tsai, J. S. (1973) Proc. Natl. Acad. Sci. USA 70,
3488-3492. 23. DeGroot, L. J., Refetoff, S., Strausser, J. & Barsano, C. (1974) Proc. Natl. Acad. Sci. USA 71, 4042-4046. 24. Griswold, M. D. & Cohen, P. P. (1972) J. Biol. Chem. 247,
353-359. 25. Kim, K.-H. & Cohen, P. P. (1966) Biochemistry 55, 12511255. 26. Martial, J. A., Seeberg, P. H., Goodman, H. M. & Baxter, J. D. (1976) 52nd Meeting of the American Thyroid Association, Toronto, Canada, Abstr. no. T-10. 27. Sokoloff, L., Roberts, P. A., Januska, M. M. & Kline, J. E. (1968) Proc. Natl. Acad. Sci. USA 60,652-659. 28. Yang, S. S. & Sanadi, D. R. (1969) J. Biol. Chem. 244, 50815082. 29. Mathews, R. W., Oronsky, A. & Haschemeyer, A. E. V. (1973)
J. Biol. Chem. 248,1329-1333.
Vol. 74, No. 5, pp. 2054-20581, May 1977
Cell Biology
Triiodothyronine stimulates specifically growth hormone mRNA in rat pituitary tumor cells* (thyroid hormone/somatotropin/translation in cell-free system/mechanism of hormone action)
HISAO SEOt, GILBERT VASSARTt, HUGETTE BROCASt, AND SAMUEL REFETOFFt University of Chicago, School of Medicine, Chicago, Illinois 60637; and * Institute de Recherche Interdisciplinaire, School of Medicine, Free University of Brussels, 1000 Brussels, Belgium
t Thyroid Study Unit,
Communicated by Elwood V. Jensen, March 10, 1977
In a cell-free protein-synthesizing system from a rabbit reticulocyte lysate, total RNA extracted from cultured rat pituitary tumor (GH3) cells directed, in a dose-related manner, the synthesis of proteins that were precipitated by antisera specific to rat growth hormone (somatotropin) and rat prolactin. A marked decrease in growth hormone secretion and growth hormone mRNA activity was observed when cells were grown in a medium deficient in thyroid hormone. Addition of triiodothyronine in physiologic amounts both prevented and completely reversed this effect within 48 hr. Thyroid hormone had no effect on prolactin secretion or prolactin mRNA activity. These data suggest that thyroid hormone may stimulate synthesis of growth hormone throug induction of transcriptional activity. The possibility of an additional effect at the posttranscriptional level has not been excluded. Although thyroid hormone is believed to have a general effect on a variety of metabolic processes, some effects, at the molecular level, may be quite selective, as indicated by the observed changes in growth hormone but not prolactin mRNA activity. The GH3 cell model is useful in the study of triiodothyronine action because of independence from secondary hormonal effects caused by hypothyroidism and because simultaneous measurement of prolactin mRNA activity serves as a unique internal control.
ABSTRACT
The demonstration of triiodothyronine (T3) binding to nuclear proteins raises the possibility that thyroid hormone may regulate gene expression. Earlier work from Tata's laboratory showed that thyroid hormone-induced protein synthesis was preceded by formation of new RNA (1). Later, DeGroot et al. and Dilman et al. showed increase in the poly(A)-rich fraction of RNA (2, 3). Demonstrations of stimulation of a specific mRNA by thyroid hormones were recently provided by Kurtz et al. (4) and by Roy et al. (5) for a2u-globulin in the rat. However, because a number of hormones are known to stimulate the synthesis of this protein (6) and because thyroid hormone profoundly alters the level of such hormones (7), experiments done in the whole animal do not provide sufficient evidence for the direct induction of -d2u-globulin mRNA by thyroid hormone. Samuels et al. have reported a quantitative correlation between nuclear T3 receptor occupancy and stimulation of growth hormone (GH, somatotropin) synthesis in cultured rat pituitary cells (8). We have adopted a similar experimental model to study more directly the action of thyroid hormone on the induction of a specific mRNA. In a recent report, we have shown that GH and prolactin mRNA activities in the total RNA extracted from a rat pituitary cell line (GH3) that actively synthesizes both hormones can be quantitated using a rabbit reticulocyte lysate cell-free system (9). In this report, we show that Abbreviations: T3, triiodothyronine; GH, growth hormone (somatotropin); Tx, thyroidectomized; N, normal; NRS, normal rabbit serum; NMS, normal monkey serum; MEM, modified Eagle's medium. * Presented in part at the Fifty-Second Meeting of the American Thyroid Association, Inc., Toronto, Canada, September 15-18, 1976.
the T3-stimulated GH but not prolactin synthesis is accompanied by a specific increase in GH mRNA but not prolactin mRNA activity. Thus, the known effect of T3 on GH production may predominantly involve a selective stimulation of transcriptional activity. METHODS
Cell Culture. Rat pituitary tumor cells originating from the GH3 line were a gift from A. H. Tashjian. This line, derived from a single MtTW5 rat pituitary tumor cell, synthesizes and secretes both GH and PRL (10). Except for the medium used, the cells were propagated in culture as previously described (11). Cells were maintained in modified Eagle's medium (MEM) containing 10% fetal calf serum. For the experiments, cells were grown in MEM supplemented with 10% serum from either normal (N) or thyroidectomized (Tx) rats, with or without added T3. Each rat received 100,uCi of Na131I on the second day following surgical thyroidectomy and serum was harvested 4 weeks later. Concentrations of endogenous thyroxine and T3 in the normal rat serum were 5.4 ,ug/100 ml and 60 ng/100 ml, respectively, and in the Tx rat serum were 0.6 ,g/100 ml and less than 20 ng/100 ml, respectively. RNA Extraction. Each RNA preparation was obtained from 5 to 15 petri dishes or 30 to 70 X 106 cells treated in the same manner. The medium was removed and saved for hormone determination. The cells adhering to the bottom of the petri dishes were washed once with 3 ml of MEM, then harvested by gentle resuspension in the same medium, using a pasteur pipette. A 1 ml aliquot of the pooled 10 ml cell suspension was saved for total RNA (12) and DNA (13) determination and the remainder was centrifuged for 10 min at 600 X g. The cell pellet was resuspended in 1 ml of buffer containing 25 mM MgCl2/50 mM KCI/200 mM sucrose/200 mM Tris-HCl, pH 8.5. In a rapid succession 7 ml of buffer containing 1% sodium dodecyl sulfate/100 mM NaCl/5 mM EDTA/0.02% heparin/10 mM Tris, pH 8.5, was added, followed by 8 ml of phenol/chloroform (1:1 vol/vol) presaturated with a buffer containing 100 mM NaCl/1 mM EDTA/10 mM Tris. Extraction was carried out by manual agitation for 5 min. Following centrifugation for 10 min at 9000 X g the aqueous phase was removed, and the organic phase containing the denatured protein interphase was re-extracted after addition of 8 ml of the suspension buffer. The two aqueous phases were pooled and extracted four consecutive times with an equal volume of phenol/chloroform. During the final extraction no denatured protein interphase was present. Residual phenol in the aqueous phase was removed by a single extraction with chloroform, following which LiCl was added to a final concentration of 2 M. The RNA was allowed to precipitate at 4° overnight. After centrifugation for 10 min at 9000 X g the RNA pellet was 2054
Cell Biology: Seo, et al.
Proc. Nati. Acad. Sci. USA 74 (1977)
2055
Table 1. Radioactivity in O-RNA and in immunologic blanks
Radioactivity RNA,
Aug/ml 0 o 200 0 0 200
Specific antiserum (1st antibody)*
Normal carrier serumt
Anti-IgG serum (2nd
Monkey anti-rat GH
antibody)t
in precipitate, mean cpm ± range
Goat anti-monkey IgG Goat anti-monkey IgG Goat anti-monkey IgG Goat anti-rabbit IgG Goat anti-rabbit IgG Goat anti-rabbit IgG
NMS NMS
Rabbit anti-rat prolactin NRS NRS
1813 ± 1953 ± 1933 ± 913 ± 1058 ± 960 ±
187 17 33 13 167 16
For details in preparation, see Methods. Volume and dilution of sera added to 200 or 250 ,. of lysate are given in the footnotes. Acid-precipitable radioactivity was 16.76 X 106 cpm. * Monkey anti-rat GH: 5 Al (1:50); rabbit anti-rat prolactin: 5 pl (1:25). t NMS = normal monkey serum: 5 Ml (1:50); NRS = normal rabbit serum: 5 Ml (1:5). 1 Goat anti-monkey IgG: 25 Al (undiluted); goat anti-rabbit IgG: 5 Al (undiluted).
washed two times with cold 2 M LiCl followed by a single wash with 66% (vol/vol) ethanol in 0.3 M NaCl. The RNA was dissolved in water and absorbance was read at 260 and 280 nm. The ratio in all preparations ranged from 1.9 to 2.1. The amount of RNA was calculated assuming that 1 A260 = 40,ug. After an ethanol precipitation, the final RNA preparation was dissolved in water to the desired concentretion to be used for translation.
Cell-Free Protein Synthesis. Translation was carried out in rabbit reticulocyte lysate cell-free system as described by Palmiter et al. (14). After incubation for 1 hr at room temperature with RNA and [3H]leucine (Amersham, 53 Ci/mmol) at 170 MCi/ml, the reaction was stopped by the addition of 0.1 ml/ml of reaction mixture of a solution containing the following additives and bringing their concentrations to 0.02 M sodium phosphate buffer and 0.15 M NaCl, 1 % (vol/vol) Triton X-100, 1% sodium deoxycholate, and 0.10% unlabeled L-leucine, pH 7.5. The mixture was then centrifuged for 60 min at 100,000 X g in a Beckman rotor SW 60. The supernatant was used for trichloroacetic acid and immunoprecipitations described below. Blank reaction mixtures contained all additives except for RNA, which was substituted by an equal volume of water. Trichloroacetic Acid Precipitation. Five microliter aliquots of the centrifuged lysate were spotted onto Whatman GF/A filter disks and processed as described by Schimke et al. (15). Immunoprecipitation. GH and prolactin synthesis in the lysate was measured by double antibody precipitation. Antibodies against rat GH and rat prolactin were a gift from A. Parlow and second antibodies, anti-rabbit and anti-monkey IgG, were from V. S. Fang. Duplicate aliquots containing 200-250 Al of the lysate were incubated with 5,l of diluted first antibody for 1 hr at room temperature, and for 16 hr at 40, followed by incubation for 30 min at room temperature and for 8 hr at 40 with the appropriate second antibody (see legend to Table 1). The precipitate was centrifuged, washed with 1 ml of phosphate-buffered saline, resuspended, and cleaned by centrifugation in Microfuge tubes through 1 M sucrose as described by Rhoads et al. (16) and dissolved in 1 ml NCS (Amersham) prior to counting. The radioactivity in lysates incubated without added RNA (0-RNA blanks) and in immunoprecipitates in the presence of normal rabbit or monkey sera, rather than the specific antisera (immunologic blanks) was similar and constituted from 0.004 to 0.020% of the acid-precipitable radioactivity (Table 1). Specific immunoprecipitable radioactivity was calculated by subtraction of the appropriate blank. Precipitation with antiprolactin and anti-GH serum was carried
a
out on separate aliquots of lysate or on the same lysate in sequence. Either method achieved excellent recovery and reproducibility (Table 2). The specificity of the antibodies was evaluated by their ability to precipitate purified authentic rat GH and prolactin labeled with 125I. The antiprolactin serum was specific, and cross reaction of the anti-GH serum with prolactin was eliminated by addition of 1 ,ug of prolactin without affecting its reactivity with GH (Fig. 1). Furthermore, as previously shown by comigration of the immunoprecipitates of authentic hormones and cell-free translated material on sodium dodecyl sulfate/polyacrylamide gel electrophoresis (9), 95% of the radioactivity appeared as sharp single peaks. In accordance with findings by Sussman et al. (17), Evans et al. (18), and Maurer et al. (19), the hormones translated in the cell-free system were of slightly larger molecular weights than the authentic hormones. Quantitation of GH and Prolactin Concentration in Culture Media. GH and prolactin concentrations in the cell culture media were determined by radioimmunoassays as previously described (11). All media were changed 48 hr prior to termination of the experiments. Thus, concentration of hormones represent the cumulative secretion by the cells into the medium over 48 hr. The cells do not store significant amounts of hormone (11) and the rate of hormonal degradation has not been determined. Table 2. Recovery with single and sequential
immunoprecipitations 1st immunoprecipitation
Specific antiserum
2nd immunoprecipitation
Specific cpm* in precipitate
Specific antiserum
16,815 + 30
Anti-GH Antiprolactin
Specific cpm* in precipitate
Anti-
prolactin Antiprolactin Anti-GH
16,815 ± 30 7,514 ± 332
8,479
±
90
393 ± 30
Lysate was programmed with 200 Mug of RNA per ml of reaction mixture. Acid-precipitable radioactivity was 24.93 X 106 and 26.58 X 106 cpm in lysates with or without added RNA, respectively. Immunologic blanks on the 1st and 2nd immunoprecipitation had, respectively, 954 I 17 and 978 ± 45 cpm with NRS and 2223 ± 201 and 1915 I 232 cpm with NMS. 0-RNA blanks precipitated with antiprolactin and anti-GH serum had 1461 + 67 and 2754 1 46 cpm, respectively. All data are expressed as mean ± range. * cpm in immunoprecipitates minus blanks.
Proc. Nati. Acad. Sci. USA 74 (1977)
Cell Biology: Seo, et al.
2056
125l-Prolactin
_
x 103
96 14-
100r
I
7
V
C,
801
r-12-
-
C
D
.SCu
60
0 Q
86
O 40 E E
a
4-
E E
2-
0
201
.O
L.
NRS anti-Prolactin
L-
Unlabeled 0 prolactin, ng/tube
500
500
L610
1 000
E
C3
0 NMS
anti-GH
FIG. 1. Specificity of antisera to rat prolactin and GH. Antisera were incubated with 125I-labeled purified rat prolactin (1251-prolactin) or GH (1251-GH) under the standard reaction conditions described in the Methods section. Antiprolactin reacted only with 125I-prolactin. The amount of precipitable 125I-GH was similar to the NRS blank. In the absence of unlabeled prolactin, anti-GH reacted equally with 125I-GH and 1251-prolactin. Addition of 1000 ng of unlabeled prolactin reduced the immunoprecipitable 1251-prolactin to the level of the NMS blank without affecting the precipitability of 125I-GH. For abbreviations see legend to Table 1.
RESULTS Application of the Reticulocyte Lysate System to Quantitate GH and Prolactin mRNA Activity. We have previously shown a linear dose response between the radioactivity in GH immunoprecipitates and the input of RNA extracted from rat pituitary tumors transplanted in vvo. The linear dose response extended over the range of 10-300 Asg of RNA per ml of lysate (9). Results from a similar study conducted with total RNA extracted from the same cells grown in culture are shown in Fig. 2. There was a dose-dependent increase in both prolactin and GH immunoprecipitable radioactivity. Depending upon the lysate preparation used, 50-60% of the total radioactivity was acid-precipitable in the absence of exogenous RNA. With the addition of increasing amounts of RNA, and depending upon the lysate preparation, the acid-precipitable radioactivity usually but not always declined to a minimum of 25% when the largest amount of 200 ,g of RNA per ml of reticulocyte lysate was added. Depending upon the source and quantity of RNA added, 0.023-0.210% of the acid-precipitable radioactivity was immunoprecipitable with anti-GH or antiprolactin sera. Immunologic and 0-RNA blanks had 0.004-0.020% of the acidprecipitable radioactivity. Effect of Thyroid Hormone Deprivation on the Secretion and mRNA Activity of GH. In this study, cells plated in MEM containing fetal calf serum were divided into three groups of 15 plates. Medium in the first group was replaced with MEM supplemented with 10% normal rat serum; in the second group, with MEM containing 10% Tx rat serum; and in the third group the 10% Tx rat serum added to MEM was enriched with 80 ng of T3 per 100 ml to bring thyroid hormone concentration to a physiologic level. Following 6 days incubation in these conditions, media were collected, cells were harvested, RNA was extracted, and mRNA activity was measured. Results are shown in Fig. 3. Culture in conditions of thyroid hormone deficiency profoundly depressed GH secretion. This was prevented by the addition of T3 to the thyroid hormone-deficient serum. There
0Total RNA, pg/ml
FIG. 2. Relation of RNA input to cell-free synthesis of prolactin and GH. Total RNA extracted from cultured rat pituitary tumor cells was added to rabbit reticulocyte lysate cell-free protein-synthesizing system to yield the final concentration indicated in the abscissa.
were no changes in prolactin secretion. A corresponding decrease in GH mRNA activity was observed in cells grown in media supplemented with thyroid hormone-poor rat serum. Addition of T3 to the Tx rat serum prevented the decrease in GH mRNA activity. The total amount of RNA per cell was not affected by the conditions of culture. RNA to DNA ratios expressed as mean ± range in two experiments were as follows: 2.4 ± 0.3 for control cells grown in normal rat serum, 2.5 + 0.3 for cells grown in Tx serum, and 2.7 + 0.2 for cells grown in Tx serum supplemented with T3. To eliminate the possibility that decrease in GH mRNA activity observed in cells grown in Tx rat serum was due to inhibition of mRNA translation in the cell-free system, the relation of RNA input to GH synthesis in the lysate was studied. As shown in Fig. 4, the proportionality between RNA input and GH synthesis was preserved in the three RNA preparations. GH mRNA activity
GH and prolactin secretion into medium
= -
m
200
a z
160'
m
120
a
jL
In
GH Prolactin
r-4,
AI
10 x
0i
cm C C
T
E
4-
a CD n
3-
O
2-
80
a 0
~0c,
C
co 0 0
E E
b.
v-
N
I-
O0
Tx
N
Tx
TS
FIG. 3. Effect of thyroid hormone on GH and prolactin synthesis and on the GH mRNA activity. Rat pituitary tumor cells were cultured in media supplemented with 10% of thyroidectomized rat serum (Tx), normal rat serum (N), or thyroidectomized rat serum with added T3 (Tx + T3). Endogenous GH and prolactin synthesis, estimated from the amount secreted into the medium, and the GH mRNA activity were measured. Data are expressed as mean i range of duplicate determinations. For methodologic details see text.
Cell Biology: Seo et al.
Proc. Natl. Acad. Sci. USA 74 (1977) +T3 (80ng/dO)
2057
Table 3. Effect of T3 on GH and prolactin secretion and on mRNA activity
E 0
Ratio, mean ± range
N
3-
a)
IV
Hormone Measurement
la 0
GH
C
Tx N
Tx + T3 N
0.24 ± 0.16
0.82 ± 0.37
Secretion,
ng/gg of
E
DNA per 48 hr mRNA ac0
100 2 '00 Total RNA, jig /ml
FIG. 4. Immunoreactive GH in lysates programmed with increasing amounts of RNA derived from cells cultured under various conditions. For abbreviations see legend to Fig. 3. Methodological details appear in the text. Culture, treatment, and translation assays using the same reticulocyte lysate were carried out simultaneously.
Also, there was no evidence for inhibition of endogenous hemoglobin synthesis by the lysate, because acid-precipitable radioactivity was identical in all reactions with same RNA input, irrespective of its origin. Thus, quantitation of the mRNA activity most likely reflects mRNA content. Reversal of the Effects of Thyroid Hormone Deprivation by Addition of T3. In this experiment, cells were allowed to grow for 8 days in MEM containing 10% Tx rat serum, following which 80 ng of T3 per 100 ml of Tx rat serum was added and culture was continued for 2 more days. Control plates were grown in the presence of either 10% normal rat serum or Tx rat serum for the periods of 8 and 10 days. Results of GH and prolactin secretion into the media and the respective mRNA activities are shown in Table 3, and are for clarity expressed as the ratio of Tx/normal or Tx + T3/normal. As compared to normal, a 4-fold depression in GH secretion and 16-fold decrease in GH mRNA activity was found in cells grown in Tx rat serum for the entire period of incubation. Culture for 2 days after addition of T3 was sufficient to normalize both GH secretion and GH mRNA activity. Irrespective of hormonal treatment, there were no significant changes in prolactin secretion or prolactin mRNA activity. DISCUSSION Thyroid hormone is known to affect a variety of metabolic processes. While these effects are well recognized, the mechanism by which thyroid hormone influences these processes has remained largely a matter of speculation (20). The recent demonstration of a specific nuclear T3 receptor in various tissues (21-23), coupled with work indicating that treatment with thyroid hormone increases polymerase (24) or chromatin template activity (25) and the poly(A)-rich fraction of RNA (2, 3), suggests that thyroid hormone action may involve changes at the transcriptional level. The demonstration that the hormone induces specific mRNA species, coding for proteins known to be under hormonal regulation, would show how specificity is conferred and provide a more direct proof for the mechanism of thyroid horomone action. Such evidence appears to have been provided by Kurtz et al. (4) and by Roy et al. (5), who showed that treatment of male hypothyroid rats for 7-8 days with pharmacologic doses of thyroxine restored to normal both urinary excretion of and mRNA activity for the hepatic protein a2u-globulin. However, the interpretation of these experiments remains complicated by the fact that the synthesis of a2u-
tivity, cpm in immunopre-
cipitate Prolactin Secretion, ng/gg of DNA per 48 hr mRNA activ-
0.062 ± 0.021 0.84 ± 0.132
1.2 ± 0.05
1.4 ± 0.34
1.4 ± 0.08
1.3 ± 0.13
ity, cpm in immunopre-
cipitate
For abbreviations and experimental design, see text.
globulin is subject to influences by a variety of hormones of
ponthyroidal origin, such as growth hormone (6), the level of
which is known to be profoundly affected by thyroid hormone (7). In the present work we have shown that thyroid hormone deprivation decreases both the synthesis and mRNA activity of GH in rat pituitary cells grown in culture. The effect was not due to other metabolic alterations caused by hypothyroidism, because addition of thyroid hormone to the medium containing hypothyroid rat serum prevented the occurrence of the changes seen when hypothyroid rat serum was used alone (Fig. 3). The effect could be completely reversed within 48 hr following the addition to the medium of physiologic concentrations of T3. In preliminary experiments, induction of synthesis and mRNA activity of GH was observed as early as 6 hr after addition of T3 to cells cultured in Tx rat serum. Furthermore, it appears that the effect of thyroid hormone is selective on GH, because in the same cell preparations no changes occurred in prolactin synthesis or mRNA activity (Table 3). Data presented in abstract form by Martial et al. (26) are in agreeme..t with our observation of T3-induced GH mRNA activity, in another cell line of rat pituitary tumors. Because this cell line possesses T3 nuclear receptors and their ability to synthesize GH is quantitatively related to nuclear T3-receptor occupancy (8), the demonstration of selective induction of GH mRNA suggests that expression of thyroid hormone action requires the following sequence of events: hormone penetration within the cell, binding to nuclear receptor proteins, induction of transcriptional activity, synthesis of specific protein, and finally the metabolic effects induced by the latter. Although, in all experiments, the thyroid hormone-induced changes in the amount of GH were always associated with corresponding changes in GH mRNA activity, direct proportionality between medium GH and cell GH mRNA activity could not always be obtained. For example, while data presented in Fig. 3 show an average of 7.5-fold decrease in both GH synthesis and mRNA activity with thyroid hormone deprivation, in Table 3, under similar circumstances, an average
2058
Cell Biology: Seo et al.
of 4-fold decrease in GH synthesis was associated with a 16-fold decrease in GH mRNA activity. While this discrepancy may be methodologic in origin, the possibility of an additional effect of thyroid hormone at the posttranscriptional level cannot be excluded (27-29). The following methodologic problems should be considered. The endogenous synthesis of GH and prolactin was estimated from secretion rate over 48 hr, in turn calculated from the hormonal concentration in the medium and cell number at the termination of the experiment. Rates of cell division and GH secretion may vary over 48 hr under different culture conditions. The rate of hormonal degradation has not been taken into account. On the other hand, although the intracellular hormone content was not measured, at 48 hr it represents less than 3% of the medium content. From the point of view of mRNA quantitation, the turnover rate of the messenger and the residence time of T3 at the receptor site should be considered. Acute induction of thyroid hormone deprivation causes a very gradual decline in GH mRNA activity over at least 6 and possibly 10 days. Furthermore, it is practically impossible to prepare a serum totally devoid of thyroid hormone. Attempts to extract the hormone modify the serum composition in a variety of other substances and serumless media are inadequate for the period of culture extending beyond 2 days that is necessary to induce a state of thyroid hormone deprivation. Thus, under the present experimental conditions, it is impossible to determine whether thyroid hormone may not indeed switch de novo the GH gene. We have eliminated the possibility of inhibition of mRNA translation in the cell-free system by showing a good dose response with various RNA preparations, and the preservation of prolactin mRNA activity in preparations with virtually no GH mRNA activity. In fact, the former serves as a unique internal control for the relative activity of the cell-free protein-synthesizing system, for possible errors from extraneous factors affecting cell-free translation, for mRNA recovery, and for the relative biologic quality of various RNA preparations tested. The authors are grateful to Dr. Armen H. Tashjian, Jr. for providing the GHS cell line, to Dr. Albert Parlow for making available rat GH and prolactin antisera, to Dr. Victor S. Fang for the anti-rabbit and anti-monkey IgG sera, to Dr. Gerald Burke for the T3 antiserum, to the National Pituitary Agency for GH and prolactin radioimmunoassay kits, and to Dr. Ruben Matalon for assistance and advice in cell culture
methodology. The technical assistance of Mrs. Ofelia Gomez and Mr. Swen R. Hagen, and the secretarial help of Mrs. Yolanda W. Richmond are also acknowledged. This work was supported in part by U.S. Public Health Service Grant AM-15070. G.V. is Charg6 de Recherche at the Belgian Fonds National de la Recherche Scientifique. His travel expenses were supported through a generous gift from Mr. and Mrs. Harry Katz. The costs of publication of this article were defrayed in part by the payment of page charges from funds made available to support the research which is the subject of the article. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
Proc. Natl. Acad. Sci. USA 74 (1977) 1. Tata, J. R. (1969) Gen. Comp. Endocrinol. 2,385-397. 2. DeGroot, L. J. & Rue, P. A. (1976) Fifth International Congress of Endocrinology, Hamburg, Germany, Abstr. no. 578. 3. Dillman, W. H., Mendecki, J., Koerner, D. W. & Oppenheimer, J. H. (1976) Fifth International Congress of Endocrinology, Hamburg, Germany, Abstr. no. 581. 4. Kurtz, D. T., Sippel, A. S. & Feigelson, P. (1976) Biochemistry 15, 1031-1036. 5. Roy, A. K., Schiop, M. J. & Dowbenko, D. J. (1976) FEBS Lett. 64,396-399. 6. Roy, A. K. (1973) J. Endocrinol. 56,295-301. 7. Hervas, F., Morreale de Escobar, G. & Escobar del Rey, F. (1975)
Endocrinology 97,91-101. 8. Samuels, H. H., Shapiro, L. E. & Tsai, J. S. (1976) 58th Annual meeting of the Endocrine Society, San Francisco, California, Abstr. no. 266. 9. Brocas, H., Seo, H., Refetoff, S. & Vassart, G. (1976) FEBS Lett. 70, 175-179. 10. Tashjian, A. H., Jr., Bancroft, F. C. & Levine, L. (1970) J. Cell Biol. 47,61-70. 11. Seo, H., Refetoff, S. & Fang, V. S. (1977) Endocrinology 100, 216-226. 12. Munro, M. N. & Fleck, A. D. (1966) in Methods of Biochemical Analysis, ed. Glick, D. (Interscience Publishers, New York), Vol. 14, pp. 113-176. 13. Burton, K. (1956) Biochem. J. 62, 315-323. 14. Palmiter, R. D., Oka, T. & Schimke, R. T. (1973) J. Biol. Chem. 248,2031-2039. 15. Schimke, R. T., Rhoads, R. E. & McKnight, G. S. (1974) in Methods in Enzymology, eds. Moldave, K. & Grossman, L. (Academic Press, New York), Vol. 30, pp. 694-701. 16. Rhoads, R. E., McKnight, G. S. & Schimke, R. T. (1973) J. Biol. Chem. 248, 2031-2039. 17. Sussman, P. M., Tushinski, R. J. & Bancroft, F. C. (1976) Proc. Natl. Acad. Sci. USA 73,29-33. 18. Evans, G. A. & Rosenfeld, M. G. (1976) J. Biol. Chem. 251, 2842-2847. 19. Maurer, R. A., Stone, R. & Gorski, J. (1976) J. Biol. Chem. 251, 2801-2807. 20. Bernal, J. & Refetoff, S. (1977) Clin. Endocrinol., in press. 21. Oppenheimer, J. H., Schwartz, H. L. & Surks, M. I. (1972) J. Clin.
Endocrinol. Metab. 35, 330-33.
22. Samuels, H. H. & Tsai, J. S. (1973) Proc. Natl. Acad. Sci. USA 70,
3488-3492. 23. DeGroot, L. J., Refetoff, S., Strausser, J. & Barsano, C. (1974) Proc. Natl. Acad. Sci. USA 71, 4042-4046. 24. Griswold, M. D. & Cohen, P. P. (1972) J. Biol. Chem. 247,
353-359. 25. Kim, K.-H. & Cohen, P. P. (1966) Biochemistry 55, 12511255. 26. Martial, J. A., Seeberg, P. H., Goodman, H. M. & Baxter, J. D. (1976) 52nd Meeting of the American Thyroid Association, Toronto, Canada, Abstr. no. T-10. 27. Sokoloff, L., Roberts, P. A., Januska, M. M. & Kline, J. E. (1968) Proc. Natl. Acad. Sci. USA 60,652-659. 28. Yang, S. S. & Sanadi, D. R. (1969) J. Biol. Chem. 244, 50815082. 29. Mathews, R. W., Oronsky, A. & Haschemeyer, A. E. V. (1973)
J. Biol. Chem. 248,1329-1333.
