Molecular and Cellular Endocrinology, 76 (1991)125-133 0 1991ElsevierScientificPublishersIreland,Ltd. 0303-7207/91/$03.50
125
MOLCEL 02467
Regulation of human growth hormone receptor gene expression by human growth hormone in a human hepatoma cell line Primus E. Mullis lv2,Torben Lund 1*3,Mukesh §. Pate1 *, Charles G.D. Brook ’ and Paul M. Brickell ’ ’ Medical Molecular Biology Unit, Department of Biochemistry, University College and Middlesex School of Medicine, London, U.K., ’ Endocrine Unit, Cobbold Laboratories, University College and Middlesex School of Medicine, The Middlesex Hospital, London, V. K., and ’ Department of Immunology, University College and Middlesex School of Medicine, London, U.K.
(Received 27 September 1990; accepted 19 December 1990)
Key worak Growth hormone; Growth hormone receptor; Transcription; Hepatoma cell line
We have investigated the effects of recombinant human growth hormone (r-hGH) on the expression of hGH-receptor in a human hepatoma cell line (HUH 7). Levels of hGH-receptor mRNA in HUH 7 cells treated with different doses of r-hGH were measured by means of an RNase protection assay. Treatment with r-hGH at physiological concentrations (12.5,25 and 50 rig/m.... resulted in an increase in hGH-receptor mRNA levels within 1 h of addition of the hormone. A steady state was reached after 3-4 h and maintained for at least 48 h. In contrast, treatment with supraphysiological r-hGH concentrations (150 and 500 ng/ml) led to a down-regulation of hGH-receptor mRNA levels during the first 3 h after hormone addition followed by an increase in hGH-receptor mRNA levels thereafter. Nuclear run-off assays demonstrated that these changes in hGH-receptor mRNA levels were a result of changes in the rate of transcription of the hGH-receptor gene. Cycloheximide (10 pg/ml) did not affect these changes in hGH-receptor gene transcription significantly, indicating that they are mediated by pre-existing factors and do not require new protein synthesis. These data demonstrate that r-hGH specifically regulates the rate of transcription of the hGH-receptor gene in a human hepatoma cell line.
Introduction The cellular effects of growth hormone (GH) are initiated by its binding to specific receptors on the plasma membrane of target cells. Depending on the cell type, this binding is followed by a series of rapid and delayed responses which can
Address for correspondence: Dr. Primus E. Mullis, University Children’s Hospital, Inselspital, CH-3010 Bern, Switzerland.
alter the metabolic state of the cell. While these metabolic effects of GH have been well studied, it has become clear only recently that GH is potent in stimulating changes in the expression of a number of different genes (Norstedt et al., 1990, for review) and that GH is involved in regulating both the behaviour of mature cell types and the induction of the differentiation of progenitor cells (Green et al., 1985). Ligand-binding studies have been used to investigate the regulatory effects of GM on GH-receptor levels, but the interpretation of these stud-
126
ies is difficult. Firstly, GH C~XI potentially bid to three classes of cell surface receptors for hormones of the GH/prolactin/ placental lac’logen family (Tanaka et al., 1980; Freemark et al., 1986). Sec-
ondly, although down-regulation of GH-receptor has been observed after GH treatment of human lM-9 lymphoblasts (Lesniak and Roth, 1976), and mouse fibroblasts (Murphy and La~anrs, 1984) in culture and in rat liver after a single injection of GH (Maiter et al., 1988a), a number of studies have indicated that GH can induce its own receptor in hepatocytes from rat (Baxter et al., 1982. 1984; Baxter and Zaltsman, 1984; Barash and Posner, 1989), sheep (Posner et al., 1980) and swine (Chung and Etherton, 1986), and in rat adipocytes (Gause and Eden, 1986; Grichting and Goodman, 1986). Studies of the effects of GH on cultured hepatocytes have also proved more difficult than anticipated. Although primary hepatocyte cultures can be maintained in serum-free hormonally-defined medium (Enat et al., 1984; Barash et al., 1988; Barash and Posner, 1989), it has been reported that established hepatoma cell lines do not respond to GH, possibly because of loss of GHreceptors or inactivation of other crucial factors during cultivation (Nontedt et al., 1990). Indeed, loss of prolactin-receptors (Barash et al,, 1988) and GH-receptors (Barash and Posner, 1989) from rat hepatocytes has been reported after the third day in culture while the cell number was not affected. The isolation of human, rabbit (Leung et al., 1987) and rat (Mathews et al., 1989) GH-receptor cDNA clones has made the analysis of GH-receptar gene expression possible and regulation of GH-receptor mRNA levels by GH has been reported recently in cultured rat epiphyseal chondrocytes (Nilsson et al., 1990). Differential regulation of GH-receptor gene expression has been found in different rat tissues (Mathews et al., 1989). In the present study we have analyzed the effect of recombinant human GH (r-hGH) on human GH-receptor expression in a human hepatoma cell line (HUH 7), which has been reported to retain differentiated functions in culture (Nakabayashi et al., 1982). Our intention was to define a human system in which to study the
effect of physiological
and supraphysiological
r-
hGH doses on hGH-receptor gene expression. Such a system wouId also allow the investigation of regulation of the hGH-receptor by other drugs used to treat the commonest problem in paediatric endocrine practice which is constitutional delay of growth and development. Using this system, we demonstrate for the first time that hGH regulation of the hGH-receptor involves changes in the rate of transcription of the hGH-receptor gene in a human hepatoma cell line. Materials and methods Culture of cells
HUH 7 is a human hepatoma cell line (Nakabayashi et al., 1982). The cells were grown in a monolayer culture in 80 cm2/260 ml culture flasks (Nunc, Copenhagen, Denmark) in a nutrient mixture (Dulbecco MEM, NUT mix F-12 (Ham), Gibco-BRL, Uxbridge, U.K.) medium enriched with 10% fetal calf serum (FCS, GibcoBRL) at 37 OC in a humidified atmosphere of 5% 02. Neither antibiotics nor fungicides were added to the culture medium. The cells were passaged twice a week. Cell counts
To count cells, the monolayer was washed with 25 mM Tris-HCl, pH 7.510 mM MgClz (binding buffer) and then trypsinized for 1 min with 0.5 ml 0.1% trypsin containing 0.023% EDTA in phosphate-buffered saline (PBS, pH 7.4; Sigma Chemicals, St. Louis, MO, U.S.A.). Experimental procedure
Before each experiment, cells (3.5-4 x PO’ viable cells/ml) were cultured in 10 ml medium in a 100 mm Petri dish (Nunc). An aliquot of cell suspension was diluted 1: 5 with trypan blue to determine cell viability, w’hich was greater than 75% at the onset of culture. At 70% ronfluency of the cells in monolayer culture, the medium was aspirated, the cells were washed twice with PBS and 10 ml of a serum-free hormonally defined medium (Barash et al., 1988) was added. After overnight incubation, variable concentrations (12.5 ng/ml, 25 ng/mI, 50 ng/mI, 150 ng/ml, 500 ng/ml) of recombinant human growth hormone
127
(r-hGH, 22K, Kabi-Vitrum, Stockholm, Sweden) were added and the cultures were further incubated for variable time periods (0, 1, 2, 3, 4, 6, 10, 12, 24, 48 h). As controls, samples were also incubated without any r-hGH. Each experiment' was repeated 4 times.
RNA extraction After exposure to r-hGH, medium was aspirated from the Petri dish and the cells were washed 3 times with PBS. Cells were scraped from culture dishes and total RNA extracted using the singlestep acid guanidium thiocyanate-phenol-chloroform method (Chomczynski and Sacchi, 1987). The integrity and accuracy of the quantification of the RNA were confirmed by gel electrophoresis of aliquots of each extract in 1% (w/v) agaroseMOPS-formaldehyde minigels stained with ethidium bromide. Only samples with intact 28S and 18S rRNA subunit bands were further processed. The RNA concentration was measured by absorbance at 260 nm using a double-beam spectrophotometer (UV 150-02, Shimadzu Corporation, Tokyo, Japan). Preparation of probes The human GH-receptor eDNA clone pGHR. 501.1 was a gift from Dr. W. Wood (Genentech, San Francisco, CA, U.S.A.). It contained an 847 base pair (bp) SacI fragment from the coding region of the hGH-receptor eDNA, in the vector pUCll9 (Leung et al., 1987). For Northern blotting, the SacI insert was labelled with [a-32p]dCTP (3000 Ci/mmol; New England Nuclear) by the random primer method (Feinberg and Vogelstein, 1984). Filters were reprobed with the human //-actin clone pHA4.1 (Khalili et al., 1983). For the RNase protection assay, a 481 bp EcoRI fragment of pGHR.501.1 (nucleotides 923-1404, Leung et al., 1987) was subcloned into the plasmid vector Bluescript SK + (Stratagene, San Diego, CA, U.S.A.), yidding plasmid pGHR.501.1-SK + . This plasmid was used to generate [a-32p]-labdled antiseI~se RNA by transcription with T3 RNA polymerase after linearization with Ncol, essentially as described by Devlin et al. (1988). The probe was 275 nucleotides long and the fragment protected by hGH-receptor mRNA was 207
nucleotides long. The human fl-tubulin eDNA clone Dfl-1 (Hall et al., 1983) was used as an internal control so that equal quantities of RNA were present in each reaction tube. A 264 bp Pstl fragment (nucleotides 128-392 of Dfl-1 eDNA clone, Hall et al., 1983) was subcloned into Bluescript SK + (Stratagene), yiel~!ing plasmid pDfl-I-SK + . The RNA probe was synthesized by transcription with T3 RNA polymerase after linearization with NcoI. The probe was 232 nucleotides long and the fragment protected by fl-tubulin mRNA 172 nucleotides long. After transcription and DNase digestion, the probes were purified by polyacrylamide-urea gel electrophoresis. Fulllength transcripts were cut from the gel and eluted at 37 °C in a buffer containing 0.5 M ammonium acetate, 1 mM EDTA, 0.1% (w/v) sodium dodecyl sulphate (SDS), and 10 mM Tris-HCl pH 7.0. Sizes of probes and protected fragments are shown in Fig. 1.
Northern blotting Total RI',IA (40 lag per track) was electrophoresed o~ 1% (w/v) agarose MOPS-formaldeh:;de gels, blotted on to Hybor,d-N filters (Amersham, Little Chalfont, U.K.). Pre-hybridization was performed for 3 h at 42°C in 50% (v/v) formamide, 5 x SSPE (20 x SSPE: 3.6 M ~4aCl, 0.2 M sodium phosphate, pH 7.7, 0.002 M Na2 EDTA), 5 >~Denhardt's solution, 0.5% (w/v) ~;DS and 500 lag/ml denatured sonicated salmon sperm DNA, followed by hybridization overnight in pre-hybridization solution containing 106 cpm/ml [a-32p]-labelled probes (hGH-receptor followed by fl-actin). The filters were then washed in 250 ml 2 x SSPE, 0.1% (w/v) SDS and 250 ml 1 x SSPE, 0.1% (w/v) SDS at 50°C for 15 and 30 min respectively, then dried and autoradiographed. Solution hybridization / RNase protection assay 40 lag of total RNA were resuspended in 30 lal hybridization buffer (80% deionized formamide (v/v), 40 mM Pipes pH 6.7, 400 mM NaCI, 1 mM EDTA) containing 7.5 × 105 cpm [a-32p]-labelled hGH receptor riboprobe and 5 x 10 a cpm [a32p]-labelled fl-tubulin riboprobe. The mixture was heated to 85 °C for 4 min and incubated at 56 °C overnight. RNase A (40 lag/ml) and RNase T1 (2
128 /~g/ml) were added and the mixture incubated for 1 h at 30 ° C. SDS and proteinase K were then added to final concentrations of 0.6~o (w/v) and 150/~g/ml respectively. After 15 min at 37°C the mixture was extracted with phenol-chloroform. After addition of 5 /tg of yeast carrier transfer RNA, the labelled RNA hybrids were precipitated in 2 vols. ethanol at - 7 0 °C for 30 min. Pellets were dissolved in sequencing loading buffer (Sequenase, version 2.0 DNA, Sequencing Kit; United States Biochemical Corporation) and were separated on an 8% polyacrylamide-8 M urea denaturing gel. After electrophoresis, the gel was fixed in 10% methanol-10~ acetic acid, dried and exposed to Kodak X-Omat AR film at - 7 0 °C with an intensifying screen for 3-4 days.
Quantification of protected RNA bands To quantify the amounts of hGH-receptor mRNA present in each sample, the radiolabelled probe fragments protected by hGH-receptor mRNA and fl-tubulin mRNA were cut out of the gels, soaked in Ecoscint A (National Diagnostics, Manville, NJ, U.S.A.) and analyzed using the 32p window of an SL 30, liquid scintillation counter (Interteclmique, Lablogic, Sheffield, U.K.). The ratio of the counts corresponding to hOH-receptor mRNA//~-tubulin mRNA at time--0 was set arbitrarily as 1.0 unit, and other data were expressed as hGH-receptor mRNA units using this base-line./~-Tubulin mRNA levels in HuH 7 cells did not change as a result of r-hOH treatment and could, therefore, be used to control for small variations in the quantity of RNA present in each sample analyzed. The values obtained were checked against values obtained by scanning densitometry (Bio-Rad Model 620 Video Densitometer, Bio-Rad Laboratories, Hemel-Hempstead, U.K.) of the autoradiograph, and were found to be significantly correlated (r - 0.81, p < 0.005).
Nuclear run-off transcription assays Nuclei were isolated from fresh washed pellets of r-hGH-treated or untreated HuH 7 cells and initiated transcripts were allowed to extend in the presence of [a-32P]CTP (3000 Ci/mmol, New England Nuclear) according to Mason et al. (1986). Each run-off probe was hybridized with one of
four identical filters dot-blotted as described previously (Patel et al., 1990) with 10/zg of each of pGHR.501.1-SK + DNA, pD~-I-SK + DNA and plasmid SK + DNA as a negative control. After cross-linking of the DNA on to the filters (Hybond-C, Amersham) by baking for 2 h at 80 ° C, the filters were pre-hybridized overnight at 42°C in 5 ml of 4 × SSC, 50 mM sodium phosphate pH 7.0, 1 x Denhardt's solution, 0.2~ (w/v) SDS, 250 /~g/ml yeast tRNA and 50% (v/v) formamide. Run-off probe (6 × 107 cpm) synthesized from treated and untreated HuH ? nuclei was added, and the filters hybridized at 42°C for 72 h. The filters were washed 3 times for 30 rain in 2 x SSC, 0.15 (w/v) SDS at 65°C, and then for 1 h in 2 x SSC, 20/~g/ml boiled RNase A at room temperature. They were then dried and autoradiographed. The strength of each autoradiographic signal was determined by liquid scintillation spectrometry of excised filter pieces (see Quantification).
Statistical analysis Statistical differences among experimental groups were determined by Student's unpaired two tailed t-test (Snedecor and Cochran, 1980). Values of p < 0.05 were considered significant. Values are expressed as means (SEM) unless otherwise stated.
Results On Northern blot analysis of total RNA prepared from the human hepatoma cell line (HuH 7) cultured in Dulbecco M E M / N U T Mix F-12 (Ham) medium with 10~ FCS, the hGH-receptor cDNA clone pGHR.501.1 identified an mRNA of 4.2 kb (Fig. l a) which is similar in size to that previously reported in rabbit and rat liver (Leung et al., 1987; Mathews et al., 1989). This mRNA was absent from human colon carcinoma cells (Fig. la). The/~-actin probe hybridized to a 1.9 kb mRNA (Khalili et al., 1983) in both cell types (data not shown). Mouse (Smith et al., 1988, 1989), rat (Frick and Goodman, 1989) and rabbit liver (Tiong et al., 1989) have been shown to express a short (1.2-1.5 kb) mRNA in addition to the fulllength mRNA. The smaller mRNA which is thought to encode the GH-binding protein (GH-
129
Lane
A C G T
1
2
3
4
5
6
nucleotides 275
~
"--" • mmt~,-
28 S
232
~
207
~
"'~,
:'H,
4.2
a)
.,
b)
1
.....
Fig. 1. (a) Northern blot of totalR N A isolatedfrom H u H 7 cells(track 1) and from colon carcinoma cells(track2) hybridized with hGH-receptor R N A probe. The sizeof the m R N A is shown in kb, as estimated by comparison with R N A molecular weight markers (BCL, London, U.K.; not shown). (b) RNase protectionassay.The intactprobe for hGH-receptor m R N A (lane 1), and the intact probe for #-tubulin m R N A (lane 3) were digested with RNase afterhybridizationto totalR N A from colon carcinoma cells(lane 5) and H u H 7 cells(lane 6). Probes digestedwith RNase without hybridizationto R N A are shown in lanes 2 and 4. Sizesof protected fragments (shown in nucleotides)were determined by referenceto an M 13 m p 18 sequences ladder (lanesA, C, G, T).
1,8' m ,ore mm G
Z l
E
1.6'
.t. 1.4
.....--4l ~
0 U
x
CP
25 ng/mL
~
50 ng/mL
----O~ :
150 ng/mL 500 ng/mL
~ 1.0
12.5ng/mL
" " 0 ----0--12
K
0 ng/rnl
~
0.8 '
J=
0
. 0
6
~
4
8
12
time
24
36
48
(h)
Fig. 2. Time course of effects of r-hOH treatment on levels of human GH-receptor mRNA in cultured HuH 7 cells. The cells were cultured as described in Materials and Methods. Different r-hGH concentrations were added to the culture medium, and cells were harvested for RNA extractions at 0, 1, 2, 3, 4, 6, 10, 12, 24, or 48 h after addition of r-hGH. Values plotted are the means of values obtained from four individual cultures. SEM were: mean 0.015 (range 0.005-0.03; n = 212). Levels of human GH-receptor mRNA are given as hGH-receptor mRNA units, with the ratio hGH receptor mRNA/p-tubulin mRNA at time = 0 arbitrarily assigned as 1.0 unit. All values shown were significantly different from the control value obtained at time = 0 ( p < 0.001, unpaired Student's
t-test).
i30
r-hGH
0 ng/ mL
cycloheximide
50 ng/ mL
+
4'
o
500 ng/ mL 4"
:++
2
•
•
•
3
"l
m
•
• 3
Fig. 3. Nuclear run-off assay. Effect of r-hGH on human GH-receptor gene transcription. The cells were cultured as described in Materials and Methods. Where indicated, cycloheximide (10 pg/ml) was added 30 min before addition of r-hGH (0 ng/ml, 50 ng/ml, 500 ng/ml). 3 h later the cells were harvested. Nuclei were isolated immediately from washed pellets, and initiated transcripts were allowed to extend as described in Materials and Methods. The resulting run-off transcripts were hybridized to filters carrying 10 pg of target DNAs: (1) hGH-receptor clone pGHR.501.1-SK+; (2) fl-tubulin clone pD~-I-SK+; (3) plasmid vector SK+, as described in Materials and Methods.
2.0-
...I.
8 1,5.
i
l.o-
0.5 0
[j 0
50
500
ng/mLhGH
Fig. 4. Effect of r-hGH on hGH-receptor transcription in HuH 7 cells assessed by nuclear run-off assays. The autoradiographic signals shown in Fig. 3 were quantified by liquid scintillation counting of filter pieces as described in Materials and Methods. The ratio of signal obtained with hGH-receptor target DNA and the fl-tubulin target DNA was calculated for each set of culture conditions, This ratio was arbitraly set at 1 unit for untreated HuH 7 cells, and the other values were adjusted accordingly and plotted. Values plotted are means of values obtained in four parallel experiments from four individual cultures of the type shown in Fig. 3. The lines above the bars indicate SEM; * * p < 0.001, compared to controls; white boxes: without cycloheximide; hatched boxes: cycloheximide was added.
BP) (Smith et al., 1989), was not detectable in HuH 7 cells. Having thus determined that HuH 7 cells express hGH-receptor mRNA, we set up a more sensitive and sequence-specific RNase protection assay for analysis of the effects of r-hGH upon hGH-receptor mRNA levels in these cells. The sizes of the hGH-receptor and fi-tubulin (internal control) probes and the fragments protected in untreated HuH 7 cells are shown in Fig. lb. The effect of r-hGH treatment on hGH-receptor mRNA levels is depicted in Fig. 2. The data are given in hGH-receptor mRNA units, calculated from the ratio hGH-receptor mRNA/fi-tubulin mRNA which was arbitrarily assigned 1.0 unit at time = 0. Treatment with r-hGH concentrations of 12.5 ng/ml, 25 ng/ml and 50 ng/ml resulted in a significant increase of hGH-receptor mRNA levels 1 h after addition to the culture medium. Maximum hGH-receptor mRNA levels were reached between 2 h (50 ng/ml) and 4 h (12.5 and 25 ng/rnl) and were then maintained for at least 48 h. By contrast, r-hGH concentrations of 150 ng/ml or 500 ng/ml resulted in a transient decrease in hGH-receptor mRNA levels, the nadir being reached 3 h after additio+, of *,he hormone in both cases. By 4 h after r-hGH addition, hGH-receptor
131
mRNA levels were restored to the level found in untreated cells and, thereafter, there was a progressive increase in mRNA levels which was more marked at the higher r-hGH concentration. Control cultures without added r-hGH showed a constant decrease in hGH-receptor mRNA levels which could not be explained by a decrease in cell viability. To address the question of whether the changes in hGH-receptor mKNA levels resulted from an changed rate of transcription or from changed mRNA stability, nuclear run-off assays were perfoJmed. HuH 7 cells were cultured for 3 h in the presence of different r-hGH concentrations (0 rig/l, 50 ng/ml, 500 ng/ml), under the same conditions as the cells used for RNA isolation. These r-hGH concentrations and incubation times were chosen because they corresponded to the conditions which provoked the greatest changes in hGH-receptor mRNA levels (Fig. 2). Nuclei were isolated from freshly isolated cells and radiolabelled run-off probes were synthesized and hybridized to filters carrying hGH-receptor and /~tubulin cDNA fragments. The results are shown in Figs. 3 and 4. 50 ng/ml r-hGH resulted in a 2-fold increase ( p < 0.001) in hGH-receptor run-off transcripts, whereas run-off transcript levels decreased by 50~ (p < 0.001) after treatment with 500 ng/ml r-hGH. In order to analyze whether these changes in the rate of hGH-receptor gene transcription were protein synthesis dependent, cycloheximide (10/~g/ml) was added to the culture medium 30 min before the r-hGH. The levels of run-off transcripts were only slightly changed, as shown in Figs. 3 and 4, and these changes were not statistically significant. This indicates that the effects of r-hGH treatment on hGH-receptor gene transcription are dependent on pre-existing factors and do not require protein synthesis.
Discussion In the present study we have demonstrated that hGH regulation of the hGH-receptor involves changes in the rate of transcription of the hGH-recepwr gene in the human hepatoma cell line HuH 7. Northern blot analysis has previously shown
that GH-receptor probes detect two homologous transcripts in some tissues. For instance, in mouse liver, these are 3.9 kb and 1.2 kb in length (Smith et al., 1988, 1989), while in rat chondrocytes only the larger transcript of 4.2 kb (Nilsson et al., 1990) has been detected. The two transcripts are thought to be produced tissue-specifically by alternative splicing (Frick and Goodman, 1989; Smith et al., 1989), with the longer transcript encoding fulllength GH-receptor protein (Leung et al., 1987) and the smaller transcript encoding uz~ isc~lated extraceUular domain of the GH-receptor (Smi_th et al., 1989), which is thought to be the c;.ccala~mg GH-binding protein. Northern blotting of ~otal RNA, isolated from HuH 7 cells, consistently revealed only the 4.2 kb transcript. The RNase protection assay which we used was specific for the full-length hGH-receptor mRNA, in particular for the transmembrane and cytoplasmic domain (Leung et al., 1987; Boutin et al., 1988; Godowski et al., 1989). Dose-dependent up-regulation of hGH-receptor mRNA levels was evident at physiological r-hGH concentrations up to 50 ng/ml. This effect was detectable 1 h after treatment and maximal induction was established by 4 h and maintained for at least 48 h. By contrast, supraphysiological concentrations (150 ng/ml, 500 ng/ml) showed a transient decrease in hGH-receptor mRNA levels, with its nadir at 3 h after addition of r-hGH. The fact that hGH-receptor expression was dose-dependently regulated and progressively decreased in an r-hGH-depleted culture medium suggests that regulation of hGH-receptor mRNA is a specific effect of hGH. The rapidity of the change in hGH-receptor mRNA levels suggests that this is a direct effect of r-hGH treatment. A similar rapid effect has been observed in GH-treated chondrocytes (Nilsson et al., 1990). In addition, the constant decrease of hGHreceptor mRNA in the control group without rhGH added underlines the fact that GH is needed to have a certain level of hGH-receptor RNA transcribed. The down-regulation of hGH.receptor mRNA levels by higher concentrations of r-hGH contrasts with the response of GH-receptor mRNA levels in GH-treated rat chondrocytes (Nilsson et al., 1990) but is in line with data obtained from specific binding of homologous 125I-GH to primary rat
132
hepatocyte cultures, which showed maximal induction of receptor binding at GH doses of 100200 ng/ml, whilst higher dosages reduced the binding level (Barash and Posner, 1989). Up-regulatory or inductive effects of peptide ligands on their own receptors are widely accepted (SheltonEarp et al., 1985; Rouiller and Gorden, 1987), and ligand-induced decreases in receptor levels have been documented for insulin (Gavin et al., 1974) and for many other hormones (Kaplan, 1981, for review). Although the application of models derived from binding studies to transcription-related events has to be undertaken with caution, our results can be discussed in the context of the proposed model for homologous receptor regulation (Barash et al., 1988). According to this model the ligand-receptor interaction activates two opposing processes: receptor production and receptor loss. Which process prevails will depend on the physiological setting and the particular ligand-receptor system exaiained. The turnover of GH-receptors seems to be faster at higher extracellular GH concentrations (Roupas and Herington, 1987, 1988), implying that occupancy of the receptor leads to a shorter half-life and, therefore, to receptor loss as described in cultured rat hepatocytes by binding studies (Barash and Posner, 1989). It is possible that when the receptor loss reaches a certain low level, the down-regulating effect of supraphysiological extraceUular GH levels must be overridden and GH-receptor gene transcription stimulated. Similarly, in GH-deficient rats, fiver GH-binding sites are acutely down-regulated in a time- and dose-dependent manner after a single GH injection, whereas GHreceptors are up-regulated by prolonged continuous GH delivery (Maiter et al., 1988a, b). This dual effect of GH might be exerted in the fiver only, because both short-term and prolonged GH delivery up-regulate GH-binding sites in isolated adipocytes (Gause and Eden, 1986; Grichting and Goodman, 1986). The r-hGH-induced changes in hGH-receptor mRNA levels in HuH 7 cells could be either transcription or mRNA stability dependent or both. To characterize the mechanism by which r-hGH regulates hGH-receptor mRNA levels, we performed run-off assays on HuH 7 cells treated
or untreated with r-hGH in the presence or absence of cycloheximide. The run-off assays showed that both up-regulation of the hGH-receptor mRNA levels by physiological doses of r-hGH and their down-regulation by supraphysiological doses were the result of changes in the rate of transcription of the hGH-receptor gene. Cycloheximide treatment did not significantly affect these changes in the rate of hGH-receptor gene transcription, suggesting that protein synthesis is not required and, therefore, that regulation by r-hGH involves pre-existing factors. The nature of these factors and the mechanism by which hGH interacts with them remain to be determined. These data indicate that the regulation of hGH-receptor expression is more complex than previously thought. Firstly, in view of the pulsatile nature of GH secretion which is an important element in its action on target tissues (Clark et al., 1985) and secondly, in view of the likely role of the serum GH-binding protein (GH-BP) on pulsatile GH secretory patterns to maintain the availability of GH during the inter-peak period (Baumann et al., 1987). This is particularly relevant because GH-BP is encoded by an mRNA derived ~..~ ~ternative splicing of the GH-receptor mRNA (Sn~th et al., 1989) and there is evidence that the two mRNAs are not coordinately regulated (Frick and Goodman, 1989). It remains to speculate whether hGH might have some direct role to play in the regulation of the alternative splicing which leads to production of hGH-receptor and hGH-BP. In the light of this complexity, it is difficult at present to explain the physiological significance of the rapid alterations in levels of hGH-receptor mRNA, which we have documented; particularly, the short-term down-regulation during the first 3 h after addition of high hGH concentrations. However, the demonstration of hGH regulation of hGH-receptor gene transcription in a human hepatoma cell line will be useful for further studies of the cellular mechanisms which regulate hGH-receptor expression. In particular, the system makes possible the analysis of hGH-receptor gene expression in response to different therapeutic modalities used for the treatment of short stature in the paediatric endocrine clinic.
133
Acknowledgments The GH receptor cDNA clone pGHR.501.1 was a generous gift from Dr. W. Wood, Genentech, San Francisco, CA, U.S.A. The fl-tubulin cDNA clone Dfl-1 was a generous gift from Dr. N.J. Cowan, University Medical Center, New York. Dr. Primus E. Mullis was supported by Schweizedsche Stiftung ftir medizinischbiologische Stipendien. Dr. Mukesh S. Patel is supported by the Leukaemia Research Fund. References Barash, I. and Posner, B.i. (1989) Mol. Cell. Endocrinol. 62, 281-286. Barash, I., Cromlish, W. and Posner, B.I. (1988) Endocrinology 122, 1151-1158. Baumann, G., Amburn, K.D. and Buchanan, T.A. (1987) J. Clin. Endocrinol. Metab. 64, 657-660. Baxter, R.C. and Zaltsman, Z. (1984) Endocrinology 115, 2009-2014. Baxter, R.C., Zaltsman, Z. and Turtle, J.R. (1982) Endocrinology 111, 1020-1022. Baxter, R.C., Zaltsman, Z. and Turtle, J.R. (1984) Endocrinology 114, 1893-1901. Boutin, J.M., Jolicoeur, C., Okamura, H., Cagnon, J., Edery, M., Shirota, M., Banville, D., Dusanter-Fort, I., Djiane, J. and Kelly, P.A. (1988) Cell. 53, 69-77. Chomczynski, P. and Sacchi, N. (1987) Anal. Biochem. 162, 156-159. Chung, C.S. and Etherton, T.D. (1986) Endocrinology 119, 780-786. Clark, R.G., Jansson, J.O., lsaksson, O. and Robinson, I.C.A.F. (1988) J. Endocrinol. 104, 53-61. Devlin, C.J., Taylor, E., Hormbruch, A., Brickell, P.M., Craig, R.K. and Wolpert, L. (1988) Development 103, 111-118. Enat, R., Jefferson, D.M., Ruiz-Opaz, N., Gatmaitan, Z., Lenwand, L.A. and Reid, L.M. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 1411-1415. Feinberg, A.P. and Vogelstein, B. (1984) Anal. Biochem. 137, 266-267. Freemark, M., Comer, M. and Handwerger, S. (1986) Am. J. Physiol. 251, E328-E333. Frick, P. and Goodman, H.M. (1989) in 71st Annual Meeting Endocrine Society, 21-24 June 1989, Seattle, WA, The Endocrine Society, abstract 1648. Gause, I. and Eden, S. (1986) Endocrinology 118, 119-124. Gavin, Ill, J.R., Roth, J., Neville, Jr., D.M., De Meyts, P. and Buell, D.N. (1974) Proc. Natl. Acad. Sci. U.S.A. 71, 84-88. Godowski, P.J., Leung, D.W., Meacham, L.R., Galgani, J.P., Hellmiss, R., Keret, R., Rotwein, P.S., Parks, J.S., Laron, Z. and Wood, W.I. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 8083-8087. Green, H., Morikawa, M. and Nixon, T. (1985) Differentiation 29, 195-198.
Grichting, G. and Goodman, H.M. (1986) Endocrinology 119, 847-854. Guzelian, P.S., Li, D., Schuetz, E.G., Thomas, P., Levin, W., Mode, A. and Gustafsson, J.A. (1998~ Proc. Natl. Acad. Sci. U.S.A. 85, 9783-9787. Hall, J.L., Dudley, P.R., Dobner, P.R., Lewis, S.A. and Cowan, N.J. (1983) Mol. Ceil. Biol. 3, 854-862. Hung, C.H. and Moore, W.W. (1984) Mol. Cell. Endocrinol. 35, 151-157. Kaplan, J. (1981) Science 212, 14-20. Khalili, K., Salas, C. and Weinmann, R. (1983) Gene 21, 9-17. Lesniak, M.A. and Roth, J. (1976) J. Biol. Chem. 251, 37203729. Leung, D.W., Spencer, S.A., Cachianes, G., Hammonds, R.G., Collins, C., Henzel, W.J., Barnard, R., Waters, M.J. and Wood, W.I. (1987) Nature 330, 537-543. Maiter, D., Underwood, L.E., Maes, M. and Ketelslegers, J.M. (1988a) Endocrinology 122, 1291-1296. Maiter, D., Underwood, L.E., Maes, M., Davenport, M.L. and Ketelslegers, M. (1988b) Endocrinology 123, 1053-1059. Mason, I.J., Murphy, D., Munke, M., Francke, U., Elliot, R.W. and Hogan, B.L.M. (1986) EMBO J. 5, 1831-1837. Mathews, L.S., Enberg, B. and Norstedt, G. (1989) J. Biol. Chem. 17, 9905-9910. Murphy, L.J. and Lazarus, L. (1984) Endocrinology 115, 16251632. Nakabayashi, N., Taketa, K., Miyano, K., Yamane, T. and Sato, K. (1982) Cancer Res. 42, 3858-3863. Nilsson, A., Carlsson, B., Mathews, L. and Isaksson, O.G.P. (1990) Mol. Cell. Endocrinol. 70, 237-246. Norstedt, G., Enberg, B., Moiler, C. and Mathews, L.S. (1990) Acta Paediatr. Scand. (Suppl.) 366, 79-83. Patel, M., Leevers, S.J. and Brickell, P.M. (1990) Int. J. Cancer 45, 342-346. Posner, B.I., Patel, B., Vezinhet, A. and Charrier, J. (1980) Endocrinology 107, 1954-1958. Rouiller, D.G. and Gorden, P. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 126-130. Roupas, P. and Herington, A. (1987) Endocrinology 121, 1521-1530. Roupas, P. and Herington, A.C. (1988) Mol. Cell. Endocrinol. 57, 93-99. Shelton-Earp, E.S., Austin, K.S., Blaisdel, J., Rubin, R.A., Nelson, K.G., Lee, L.W. and Grisham, J.W. (1985) J. Biol. Chem. 261, 4777-4780. Smith, W.C., Linzer, D.I.H. and Talamantes, F. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 9576-9579. Smith, W.C., Kuniyoshi, J. and Talamantes, F. (1989) Mol. Endocrinol. 3, 984-990. Snedecor, G.W. and Cochran, W.G. (1980) Statistical Methods, 7th edn., pp. 59-61, Iowa State University Press, Ames, IA. Tanaka, T., Shiu, R.P.C., Gout, P.W., Beer, B.T., Noble, R.L. and Friesen, H.G. (1980) J. Clin. Endocrinol. Metab. 51, 1058-1063. Tiong, T.S., Freed, K.S. and Herington, A.C. (1989) Biochem. Biophys. Res. Commun. 158, 141-148.
125
MOLCEL 02467
Regulation of human growth hormone receptor gene expression by human growth hormone in a human hepatoma cell line Primus E. Mullis lv2,Torben Lund 1*3,Mukesh §. Pate1 *, Charles G.D. Brook ’ and Paul M. Brickell ’ ’ Medical Molecular Biology Unit, Department of Biochemistry, University College and Middlesex School of Medicine, London, U.K., ’ Endocrine Unit, Cobbold Laboratories, University College and Middlesex School of Medicine, The Middlesex Hospital, London, V. K., and ’ Department of Immunology, University College and Middlesex School of Medicine, London, U.K.
(Received 27 September 1990; accepted 19 December 1990)
Key worak Growth hormone; Growth hormone receptor; Transcription; Hepatoma cell line
We have investigated the effects of recombinant human growth hormone (r-hGH) on the expression of hGH-receptor in a human hepatoma cell line (HUH 7). Levels of hGH-receptor mRNA in HUH 7 cells treated with different doses of r-hGH were measured by means of an RNase protection assay. Treatment with r-hGH at physiological concentrations (12.5,25 and 50 rig/m.... resulted in an increase in hGH-receptor mRNA levels within 1 h of addition of the hormone. A steady state was reached after 3-4 h and maintained for at least 48 h. In contrast, treatment with supraphysiological r-hGH concentrations (150 and 500 ng/ml) led to a down-regulation of hGH-receptor mRNA levels during the first 3 h after hormone addition followed by an increase in hGH-receptor mRNA levels thereafter. Nuclear run-off assays demonstrated that these changes in hGH-receptor mRNA levels were a result of changes in the rate of transcription of the hGH-receptor gene. Cycloheximide (10 pg/ml) did not affect these changes in hGH-receptor gene transcription significantly, indicating that they are mediated by pre-existing factors and do not require new protein synthesis. These data demonstrate that r-hGH specifically regulates the rate of transcription of the hGH-receptor gene in a human hepatoma cell line.
Introduction The cellular effects of growth hormone (GH) are initiated by its binding to specific receptors on the plasma membrane of target cells. Depending on the cell type, this binding is followed by a series of rapid and delayed responses which can
Address for correspondence: Dr. Primus E. Mullis, University Children’s Hospital, Inselspital, CH-3010 Bern, Switzerland.
alter the metabolic state of the cell. While these metabolic effects of GH have been well studied, it has become clear only recently that GH is potent in stimulating changes in the expression of a number of different genes (Norstedt et al., 1990, for review) and that GH is involved in regulating both the behaviour of mature cell types and the induction of the differentiation of progenitor cells (Green et al., 1985). Ligand-binding studies have been used to investigate the regulatory effects of GM on GH-receptor levels, but the interpretation of these stud-
126
ies is difficult. Firstly, GH C~XI potentially bid to three classes of cell surface receptors for hormones of the GH/prolactin/ placental lac’logen family (Tanaka et al., 1980; Freemark et al., 1986). Sec-
ondly, although down-regulation of GH-receptor has been observed after GH treatment of human lM-9 lymphoblasts (Lesniak and Roth, 1976), and mouse fibroblasts (Murphy and La~anrs, 1984) in culture and in rat liver after a single injection of GH (Maiter et al., 1988a), a number of studies have indicated that GH can induce its own receptor in hepatocytes from rat (Baxter et al., 1982. 1984; Baxter and Zaltsman, 1984; Barash and Posner, 1989), sheep (Posner et al., 1980) and swine (Chung and Etherton, 1986), and in rat adipocytes (Gause and Eden, 1986; Grichting and Goodman, 1986). Studies of the effects of GH on cultured hepatocytes have also proved more difficult than anticipated. Although primary hepatocyte cultures can be maintained in serum-free hormonally-defined medium (Enat et al., 1984; Barash et al., 1988; Barash and Posner, 1989), it has been reported that established hepatoma cell lines do not respond to GH, possibly because of loss of GHreceptors or inactivation of other crucial factors during cultivation (Nontedt et al., 1990). Indeed, loss of prolactin-receptors (Barash et al,, 1988) and GH-receptors (Barash and Posner, 1989) from rat hepatocytes has been reported after the third day in culture while the cell number was not affected. The isolation of human, rabbit (Leung et al., 1987) and rat (Mathews et al., 1989) GH-receptor cDNA clones has made the analysis of GH-receptar gene expression possible and regulation of GH-receptor mRNA levels by GH has been reported recently in cultured rat epiphyseal chondrocytes (Nilsson et al., 1990). Differential regulation of GH-receptor gene expression has been found in different rat tissues (Mathews et al., 1989). In the present study we have analyzed the effect of recombinant human GH (r-hGH) on human GH-receptor expression in a human hepatoma cell line (HUH 7), which has been reported to retain differentiated functions in culture (Nakabayashi et al., 1982). Our intention was to define a human system in which to study the
effect of physiological
and supraphysiological
r-
hGH doses on hGH-receptor gene expression. Such a system wouId also allow the investigation of regulation of the hGH-receptor by other drugs used to treat the commonest problem in paediatric endocrine practice which is constitutional delay of growth and development. Using this system, we demonstrate for the first time that hGH regulation of the hGH-receptor involves changes in the rate of transcription of the hGH-receptor gene in a human hepatoma cell line. Materials and methods Culture of cells
HUH 7 is a human hepatoma cell line (Nakabayashi et al., 1982). The cells were grown in a monolayer culture in 80 cm2/260 ml culture flasks (Nunc, Copenhagen, Denmark) in a nutrient mixture (Dulbecco MEM, NUT mix F-12 (Ham), Gibco-BRL, Uxbridge, U.K.) medium enriched with 10% fetal calf serum (FCS, GibcoBRL) at 37 OC in a humidified atmosphere of 5% 02. Neither antibiotics nor fungicides were added to the culture medium. The cells were passaged twice a week. Cell counts
To count cells, the monolayer was washed with 25 mM Tris-HCl, pH 7.510 mM MgClz (binding buffer) and then trypsinized for 1 min with 0.5 ml 0.1% trypsin containing 0.023% EDTA in phosphate-buffered saline (PBS, pH 7.4; Sigma Chemicals, St. Louis, MO, U.S.A.). Experimental procedure
Before each experiment, cells (3.5-4 x PO’ viable cells/ml) were cultured in 10 ml medium in a 100 mm Petri dish (Nunc). An aliquot of cell suspension was diluted 1: 5 with trypan blue to determine cell viability, w’hich was greater than 75% at the onset of culture. At 70% ronfluency of the cells in monolayer culture, the medium was aspirated, the cells were washed twice with PBS and 10 ml of a serum-free hormonally defined medium (Barash et al., 1988) was added. After overnight incubation, variable concentrations (12.5 ng/ml, 25 ng/mI, 50 ng/mI, 150 ng/ml, 500 ng/ml) of recombinant human growth hormone
127
(r-hGH, 22K, Kabi-Vitrum, Stockholm, Sweden) were added and the cultures were further incubated for variable time periods (0, 1, 2, 3, 4, 6, 10, 12, 24, 48 h). As controls, samples were also incubated without any r-hGH. Each experiment' was repeated 4 times.
RNA extraction After exposure to r-hGH, medium was aspirated from the Petri dish and the cells were washed 3 times with PBS. Cells were scraped from culture dishes and total RNA extracted using the singlestep acid guanidium thiocyanate-phenol-chloroform method (Chomczynski and Sacchi, 1987). The integrity and accuracy of the quantification of the RNA were confirmed by gel electrophoresis of aliquots of each extract in 1% (w/v) agaroseMOPS-formaldehyde minigels stained with ethidium bromide. Only samples with intact 28S and 18S rRNA subunit bands were further processed. The RNA concentration was measured by absorbance at 260 nm using a double-beam spectrophotometer (UV 150-02, Shimadzu Corporation, Tokyo, Japan). Preparation of probes The human GH-receptor eDNA clone pGHR. 501.1 was a gift from Dr. W. Wood (Genentech, San Francisco, CA, U.S.A.). It contained an 847 base pair (bp) SacI fragment from the coding region of the hGH-receptor eDNA, in the vector pUCll9 (Leung et al., 1987). For Northern blotting, the SacI insert was labelled with [a-32p]dCTP (3000 Ci/mmol; New England Nuclear) by the random primer method (Feinberg and Vogelstein, 1984). Filters were reprobed with the human //-actin clone pHA4.1 (Khalili et al., 1983). For the RNase protection assay, a 481 bp EcoRI fragment of pGHR.501.1 (nucleotides 923-1404, Leung et al., 1987) was subcloned into the plasmid vector Bluescript SK + (Stratagene, San Diego, CA, U.S.A.), yidding plasmid pGHR.501.1-SK + . This plasmid was used to generate [a-32p]-labdled antiseI~se RNA by transcription with T3 RNA polymerase after linearization with Ncol, essentially as described by Devlin et al. (1988). The probe was 275 nucleotides long and the fragment protected by hGH-receptor mRNA was 207
nucleotides long. The human fl-tubulin eDNA clone Dfl-1 (Hall et al., 1983) was used as an internal control so that equal quantities of RNA were present in each reaction tube. A 264 bp Pstl fragment (nucleotides 128-392 of Dfl-1 eDNA clone, Hall et al., 1983) was subcloned into Bluescript SK + (Stratagene), yiel~!ing plasmid pDfl-I-SK + . The RNA probe was synthesized by transcription with T3 RNA polymerase after linearization with NcoI. The probe was 232 nucleotides long and the fragment protected by fl-tubulin mRNA 172 nucleotides long. After transcription and DNase digestion, the probes were purified by polyacrylamide-urea gel electrophoresis. Fulllength transcripts were cut from the gel and eluted at 37 °C in a buffer containing 0.5 M ammonium acetate, 1 mM EDTA, 0.1% (w/v) sodium dodecyl sulphate (SDS), and 10 mM Tris-HCl pH 7.0. Sizes of probes and protected fragments are shown in Fig. 1.
Northern blotting Total RI',IA (40 lag per track) was electrophoresed o~ 1% (w/v) agarose MOPS-formaldeh:;de gels, blotted on to Hybor,d-N filters (Amersham, Little Chalfont, U.K.). Pre-hybridization was performed for 3 h at 42°C in 50% (v/v) formamide, 5 x SSPE (20 x SSPE: 3.6 M ~4aCl, 0.2 M sodium phosphate, pH 7.7, 0.002 M Na2 EDTA), 5 >~Denhardt's solution, 0.5% (w/v) ~;DS and 500 lag/ml denatured sonicated salmon sperm DNA, followed by hybridization overnight in pre-hybridization solution containing 106 cpm/ml [a-32p]-labelled probes (hGH-receptor followed by fl-actin). The filters were then washed in 250 ml 2 x SSPE, 0.1% (w/v) SDS and 250 ml 1 x SSPE, 0.1% (w/v) SDS at 50°C for 15 and 30 min respectively, then dried and autoradiographed. Solution hybridization / RNase protection assay 40 lag of total RNA were resuspended in 30 lal hybridization buffer (80% deionized formamide (v/v), 40 mM Pipes pH 6.7, 400 mM NaCI, 1 mM EDTA) containing 7.5 × 105 cpm [a-32p]-labelled hGH receptor riboprobe and 5 x 10 a cpm [a32p]-labelled fl-tubulin riboprobe. The mixture was heated to 85 °C for 4 min and incubated at 56 °C overnight. RNase A (40 lag/ml) and RNase T1 (2
128 /~g/ml) were added and the mixture incubated for 1 h at 30 ° C. SDS and proteinase K were then added to final concentrations of 0.6~o (w/v) and 150/~g/ml respectively. After 15 min at 37°C the mixture was extracted with phenol-chloroform. After addition of 5 /tg of yeast carrier transfer RNA, the labelled RNA hybrids were precipitated in 2 vols. ethanol at - 7 0 °C for 30 min. Pellets were dissolved in sequencing loading buffer (Sequenase, version 2.0 DNA, Sequencing Kit; United States Biochemical Corporation) and were separated on an 8% polyacrylamide-8 M urea denaturing gel. After electrophoresis, the gel was fixed in 10% methanol-10~ acetic acid, dried and exposed to Kodak X-Omat AR film at - 7 0 °C with an intensifying screen for 3-4 days.
Quantification of protected RNA bands To quantify the amounts of hGH-receptor mRNA present in each sample, the radiolabelled probe fragments protected by hGH-receptor mRNA and fl-tubulin mRNA were cut out of the gels, soaked in Ecoscint A (National Diagnostics, Manville, NJ, U.S.A.) and analyzed using the 32p window of an SL 30, liquid scintillation counter (Interteclmique, Lablogic, Sheffield, U.K.). The ratio of the counts corresponding to hOH-receptor mRNA//~-tubulin mRNA at time--0 was set arbitrarily as 1.0 unit, and other data were expressed as hGH-receptor mRNA units using this base-line./~-Tubulin mRNA levels in HuH 7 cells did not change as a result of r-hOH treatment and could, therefore, be used to control for small variations in the quantity of RNA present in each sample analyzed. The values obtained were checked against values obtained by scanning densitometry (Bio-Rad Model 620 Video Densitometer, Bio-Rad Laboratories, Hemel-Hempstead, U.K.) of the autoradiograph, and were found to be significantly correlated (r - 0.81, p < 0.005).
Nuclear run-off transcription assays Nuclei were isolated from fresh washed pellets of r-hGH-treated or untreated HuH 7 cells and initiated transcripts were allowed to extend in the presence of [a-32P]CTP (3000 Ci/mmol, New England Nuclear) according to Mason et al. (1986). Each run-off probe was hybridized with one of
four identical filters dot-blotted as described previously (Patel et al., 1990) with 10/zg of each of pGHR.501.1-SK + DNA, pD~-I-SK + DNA and plasmid SK + DNA as a negative control. After cross-linking of the DNA on to the filters (Hybond-C, Amersham) by baking for 2 h at 80 ° C, the filters were pre-hybridized overnight at 42°C in 5 ml of 4 × SSC, 50 mM sodium phosphate pH 7.0, 1 x Denhardt's solution, 0.2~ (w/v) SDS, 250 /~g/ml yeast tRNA and 50% (v/v) formamide. Run-off probe (6 × 107 cpm) synthesized from treated and untreated HuH ? nuclei was added, and the filters hybridized at 42°C for 72 h. The filters were washed 3 times for 30 rain in 2 x SSC, 0.15 (w/v) SDS at 65°C, and then for 1 h in 2 x SSC, 20/~g/ml boiled RNase A at room temperature. They were then dried and autoradiographed. The strength of each autoradiographic signal was determined by liquid scintillation spectrometry of excised filter pieces (see Quantification).
Statistical analysis Statistical differences among experimental groups were determined by Student's unpaired two tailed t-test (Snedecor and Cochran, 1980). Values of p < 0.05 were considered significant. Values are expressed as means (SEM) unless otherwise stated.
Results On Northern blot analysis of total RNA prepared from the human hepatoma cell line (HuH 7) cultured in Dulbecco M E M / N U T Mix F-12 (Ham) medium with 10~ FCS, the hGH-receptor cDNA clone pGHR.501.1 identified an mRNA of 4.2 kb (Fig. l a) which is similar in size to that previously reported in rabbit and rat liver (Leung et al., 1987; Mathews et al., 1989). This mRNA was absent from human colon carcinoma cells (Fig. la). The/~-actin probe hybridized to a 1.9 kb mRNA (Khalili et al., 1983) in both cell types (data not shown). Mouse (Smith et al., 1988, 1989), rat (Frick and Goodman, 1989) and rabbit liver (Tiong et al., 1989) have been shown to express a short (1.2-1.5 kb) mRNA in addition to the fulllength mRNA. The smaller mRNA which is thought to encode the GH-binding protein (GH-
129
Lane
A C G T
1
2
3
4
5
6
nucleotides 275
~
"--" • mmt~,-
28 S
232
~
207
~
"'~,
:'H,
4.2
a)
.,
b)
1
.....
Fig. 1. (a) Northern blot of totalR N A isolatedfrom H u H 7 cells(track 1) and from colon carcinoma cells(track2) hybridized with hGH-receptor R N A probe. The sizeof the m R N A is shown in kb, as estimated by comparison with R N A molecular weight markers (BCL, London, U.K.; not shown). (b) RNase protectionassay.The intactprobe for hGH-receptor m R N A (lane 1), and the intact probe for #-tubulin m R N A (lane 3) were digested with RNase afterhybridizationto totalR N A from colon carcinoma cells(lane 5) and H u H 7 cells(lane 6). Probes digestedwith RNase without hybridizationto R N A are shown in lanes 2 and 4. Sizesof protected fragments (shown in nucleotides)were determined by referenceto an M 13 m p 18 sequences ladder (lanesA, C, G, T).
1,8' m ,ore mm G
Z l
E
1.6'
.t. 1.4
.....--4l ~
0 U
x
CP
25 ng/mL
~
50 ng/mL
----O~ :
150 ng/mL 500 ng/mL
~ 1.0
12.5ng/mL
" " 0 ----0--12
K
0 ng/rnl
~
0.8 '
J=
0
. 0
6
~
4
8
12
time
24
36
48
(h)
Fig. 2. Time course of effects of r-hOH treatment on levels of human GH-receptor mRNA in cultured HuH 7 cells. The cells were cultured as described in Materials and Methods. Different r-hGH concentrations were added to the culture medium, and cells were harvested for RNA extractions at 0, 1, 2, 3, 4, 6, 10, 12, 24, or 48 h after addition of r-hGH. Values plotted are the means of values obtained from four individual cultures. SEM were: mean 0.015 (range 0.005-0.03; n = 212). Levels of human GH-receptor mRNA are given as hGH-receptor mRNA units, with the ratio hGH receptor mRNA/p-tubulin mRNA at time = 0 arbitrarily assigned as 1.0 unit. All values shown were significantly different from the control value obtained at time = 0 ( p < 0.001, unpaired Student's
t-test).
i30
r-hGH
0 ng/ mL
cycloheximide
50 ng/ mL
+
4'
o
500 ng/ mL 4"
:++
2
•
•
•
3
"l
m
•
• 3
Fig. 3. Nuclear run-off assay. Effect of r-hGH on human GH-receptor gene transcription. The cells were cultured as described in Materials and Methods. Where indicated, cycloheximide (10 pg/ml) was added 30 min before addition of r-hGH (0 ng/ml, 50 ng/ml, 500 ng/ml). 3 h later the cells were harvested. Nuclei were isolated immediately from washed pellets, and initiated transcripts were allowed to extend as described in Materials and Methods. The resulting run-off transcripts were hybridized to filters carrying 10 pg of target DNAs: (1) hGH-receptor clone pGHR.501.1-SK+; (2) fl-tubulin clone pD~-I-SK+; (3) plasmid vector SK+, as described in Materials and Methods.
2.0-
...I.
8 1,5.
i
l.o-
0.5 0
[j 0
50
500
ng/mLhGH
Fig. 4. Effect of r-hGH on hGH-receptor transcription in HuH 7 cells assessed by nuclear run-off assays. The autoradiographic signals shown in Fig. 3 were quantified by liquid scintillation counting of filter pieces as described in Materials and Methods. The ratio of signal obtained with hGH-receptor target DNA and the fl-tubulin target DNA was calculated for each set of culture conditions, This ratio was arbitraly set at 1 unit for untreated HuH 7 cells, and the other values were adjusted accordingly and plotted. Values plotted are means of values obtained in four parallel experiments from four individual cultures of the type shown in Fig. 3. The lines above the bars indicate SEM; * * p < 0.001, compared to controls; white boxes: without cycloheximide; hatched boxes: cycloheximide was added.
BP) (Smith et al., 1989), was not detectable in HuH 7 cells. Having thus determined that HuH 7 cells express hGH-receptor mRNA, we set up a more sensitive and sequence-specific RNase protection assay for analysis of the effects of r-hGH upon hGH-receptor mRNA levels in these cells. The sizes of the hGH-receptor and fi-tubulin (internal control) probes and the fragments protected in untreated HuH 7 cells are shown in Fig. lb. The effect of r-hGH treatment on hGH-receptor mRNA levels is depicted in Fig. 2. The data are given in hGH-receptor mRNA units, calculated from the ratio hGH-receptor mRNA/fi-tubulin mRNA which was arbitrarily assigned 1.0 unit at time = 0. Treatment with r-hGH concentrations of 12.5 ng/ml, 25 ng/ml and 50 ng/ml resulted in a significant increase of hGH-receptor mRNA levels 1 h after addition to the culture medium. Maximum hGH-receptor mRNA levels were reached between 2 h (50 ng/ml) and 4 h (12.5 and 25 ng/rnl) and were then maintained for at least 48 h. By contrast, r-hGH concentrations of 150 ng/ml or 500 ng/ml resulted in a transient decrease in hGH-receptor mRNA levels, the nadir being reached 3 h after additio+, of *,he hormone in both cases. By 4 h after r-hGH addition, hGH-receptor
131
mRNA levels were restored to the level found in untreated cells and, thereafter, there was a progressive increase in mRNA levels which was more marked at the higher r-hGH concentration. Control cultures without added r-hGH showed a constant decrease in hGH-receptor mRNA levels which could not be explained by a decrease in cell viability. To address the question of whether the changes in hGH-receptor mKNA levels resulted from an changed rate of transcription or from changed mRNA stability, nuclear run-off assays were perfoJmed. HuH 7 cells were cultured for 3 h in the presence of different r-hGH concentrations (0 rig/l, 50 ng/ml, 500 ng/ml), under the same conditions as the cells used for RNA isolation. These r-hGH concentrations and incubation times were chosen because they corresponded to the conditions which provoked the greatest changes in hGH-receptor mRNA levels (Fig. 2). Nuclei were isolated from freshly isolated cells and radiolabelled run-off probes were synthesized and hybridized to filters carrying hGH-receptor and /~tubulin cDNA fragments. The results are shown in Figs. 3 and 4. 50 ng/ml r-hGH resulted in a 2-fold increase ( p < 0.001) in hGH-receptor run-off transcripts, whereas run-off transcript levels decreased by 50~ (p < 0.001) after treatment with 500 ng/ml r-hGH. In order to analyze whether these changes in the rate of hGH-receptor gene transcription were protein synthesis dependent, cycloheximide (10/~g/ml) was added to the culture medium 30 min before the r-hGH. The levels of run-off transcripts were only slightly changed, as shown in Figs. 3 and 4, and these changes were not statistically significant. This indicates that the effects of r-hGH treatment on hGH-receptor gene transcription are dependent on pre-existing factors and do not require protein synthesis.
Discussion In the present study we have demonstrated that hGH regulation of the hGH-receptor involves changes in the rate of transcription of the hGH-recepwr gene in the human hepatoma cell line HuH 7. Northern blot analysis has previously shown
that GH-receptor probes detect two homologous transcripts in some tissues. For instance, in mouse liver, these are 3.9 kb and 1.2 kb in length (Smith et al., 1988, 1989), while in rat chondrocytes only the larger transcript of 4.2 kb (Nilsson et al., 1990) has been detected. The two transcripts are thought to be produced tissue-specifically by alternative splicing (Frick and Goodman, 1989; Smith et al., 1989), with the longer transcript encoding fulllength GH-receptor protein (Leung et al., 1987) and the smaller transcript encoding uz~ isc~lated extraceUular domain of the GH-receptor (Smi_th et al., 1989), which is thought to be the c;.ccala~mg GH-binding protein. Northern blotting of ~otal RNA, isolated from HuH 7 cells, consistently revealed only the 4.2 kb transcript. The RNase protection assay which we used was specific for the full-length hGH-receptor mRNA, in particular for the transmembrane and cytoplasmic domain (Leung et al., 1987; Boutin et al., 1988; Godowski et al., 1989). Dose-dependent up-regulation of hGH-receptor mRNA levels was evident at physiological r-hGH concentrations up to 50 ng/ml. This effect was detectable 1 h after treatment and maximal induction was established by 4 h and maintained for at least 48 h. By contrast, supraphysiological concentrations (150 ng/ml, 500 ng/ml) showed a transient decrease in hGH-receptor mRNA levels, with its nadir at 3 h after addition of r-hGH. The fact that hGH-receptor expression was dose-dependently regulated and progressively decreased in an r-hGH-depleted culture medium suggests that regulation of hGH-receptor mRNA is a specific effect of hGH. The rapidity of the change in hGH-receptor mRNA levels suggests that this is a direct effect of r-hGH treatment. A similar rapid effect has been observed in GH-treated chondrocytes (Nilsson et al., 1990). In addition, the constant decrease of hGHreceptor mRNA in the control group without rhGH added underlines the fact that GH is needed to have a certain level of hGH-receptor RNA transcribed. The down-regulation of hGH.receptor mRNA levels by higher concentrations of r-hGH contrasts with the response of GH-receptor mRNA levels in GH-treated rat chondrocytes (Nilsson et al., 1990) but is in line with data obtained from specific binding of homologous 125I-GH to primary rat
132
hepatocyte cultures, which showed maximal induction of receptor binding at GH doses of 100200 ng/ml, whilst higher dosages reduced the binding level (Barash and Posner, 1989). Up-regulatory or inductive effects of peptide ligands on their own receptors are widely accepted (SheltonEarp et al., 1985; Rouiller and Gorden, 1987), and ligand-induced decreases in receptor levels have been documented for insulin (Gavin et al., 1974) and for many other hormones (Kaplan, 1981, for review). Although the application of models derived from binding studies to transcription-related events has to be undertaken with caution, our results can be discussed in the context of the proposed model for homologous receptor regulation (Barash et al., 1988). According to this model the ligand-receptor interaction activates two opposing processes: receptor production and receptor loss. Which process prevails will depend on the physiological setting and the particular ligand-receptor system exaiained. The turnover of GH-receptors seems to be faster at higher extracellular GH concentrations (Roupas and Herington, 1987, 1988), implying that occupancy of the receptor leads to a shorter half-life and, therefore, to receptor loss as described in cultured rat hepatocytes by binding studies (Barash and Posner, 1989). It is possible that when the receptor loss reaches a certain low level, the down-regulating effect of supraphysiological extraceUular GH levels must be overridden and GH-receptor gene transcription stimulated. Similarly, in GH-deficient rats, fiver GH-binding sites are acutely down-regulated in a time- and dose-dependent manner after a single GH injection, whereas GHreceptors are up-regulated by prolonged continuous GH delivery (Maiter et al., 1988a, b). This dual effect of GH might be exerted in the fiver only, because both short-term and prolonged GH delivery up-regulate GH-binding sites in isolated adipocytes (Gause and Eden, 1986; Grichting and Goodman, 1986). The r-hGH-induced changes in hGH-receptor mRNA levels in HuH 7 cells could be either transcription or mRNA stability dependent or both. To characterize the mechanism by which r-hGH regulates hGH-receptor mRNA levels, we performed run-off assays on HuH 7 cells treated
or untreated with r-hGH in the presence or absence of cycloheximide. The run-off assays showed that both up-regulation of the hGH-receptor mRNA levels by physiological doses of r-hGH and their down-regulation by supraphysiological doses were the result of changes in the rate of transcription of the hGH-receptor gene. Cycloheximide treatment did not significantly affect these changes in the rate of hGH-receptor gene transcription, suggesting that protein synthesis is not required and, therefore, that regulation by r-hGH involves pre-existing factors. The nature of these factors and the mechanism by which hGH interacts with them remain to be determined. These data indicate that the regulation of hGH-receptor expression is more complex than previously thought. Firstly, in view of the pulsatile nature of GH secretion which is an important element in its action on target tissues (Clark et al., 1985) and secondly, in view of the likely role of the serum GH-binding protein (GH-BP) on pulsatile GH secretory patterns to maintain the availability of GH during the inter-peak period (Baumann et al., 1987). This is particularly relevant because GH-BP is encoded by an mRNA derived ~..~ ~ternative splicing of the GH-receptor mRNA (Sn~th et al., 1989) and there is evidence that the two mRNAs are not coordinately regulated (Frick and Goodman, 1989). It remains to speculate whether hGH might have some direct role to play in the regulation of the alternative splicing which leads to production of hGH-receptor and hGH-BP. In the light of this complexity, it is difficult at present to explain the physiological significance of the rapid alterations in levels of hGH-receptor mRNA, which we have documented; particularly, the short-term down-regulation during the first 3 h after addition of high hGH concentrations. However, the demonstration of hGH regulation of hGH-receptor gene transcription in a human hepatoma cell line will be useful for further studies of the cellular mechanisms which regulate hGH-receptor expression. In particular, the system makes possible the analysis of hGH-receptor gene expression in response to different therapeutic modalities used for the treatment of short stature in the paediatric endocrine clinic.
133
Acknowledgments The GH receptor cDNA clone pGHR.501.1 was a generous gift from Dr. W. Wood, Genentech, San Francisco, CA, U.S.A. The fl-tubulin cDNA clone Dfl-1 was a generous gift from Dr. N.J. Cowan, University Medical Center, New York. Dr. Primus E. Mullis was supported by Schweizedsche Stiftung ftir medizinischbiologische Stipendien. Dr. Mukesh S. Patel is supported by the Leukaemia Research Fund. References Barash, I. and Posner, B.i. (1989) Mol. Cell. Endocrinol. 62, 281-286. Barash, I., Cromlish, W. and Posner, B.I. (1988) Endocrinology 122, 1151-1158. Baumann, G., Amburn, K.D. and Buchanan, T.A. (1987) J. Clin. Endocrinol. Metab. 64, 657-660. Baxter, R.C. and Zaltsman, Z. (1984) Endocrinology 115, 2009-2014. Baxter, R.C., Zaltsman, Z. and Turtle, J.R. (1982) Endocrinology 111, 1020-1022. Baxter, R.C., Zaltsman, Z. and Turtle, J.R. (1984) Endocrinology 114, 1893-1901. Boutin, J.M., Jolicoeur, C., Okamura, H., Cagnon, J., Edery, M., Shirota, M., Banville, D., Dusanter-Fort, I., Djiane, J. and Kelly, P.A. (1988) Cell. 53, 69-77. Chomczynski, P. and Sacchi, N. (1987) Anal. Biochem. 162, 156-159. Chung, C.S. and Etherton, T.D. (1986) Endocrinology 119, 780-786. Clark, R.G., Jansson, J.O., lsaksson, O. and Robinson, I.C.A.F. (1988) J. Endocrinol. 104, 53-61. Devlin, C.J., Taylor, E., Hormbruch, A., Brickell, P.M., Craig, R.K. and Wolpert, L. (1988) Development 103, 111-118. Enat, R., Jefferson, D.M., Ruiz-Opaz, N., Gatmaitan, Z., Lenwand, L.A. and Reid, L.M. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 1411-1415. Feinberg, A.P. and Vogelstein, B. (1984) Anal. Biochem. 137, 266-267. Freemark, M., Comer, M. and Handwerger, S. (1986) Am. J. Physiol. 251, E328-E333. Frick, P. and Goodman, H.M. (1989) in 71st Annual Meeting Endocrine Society, 21-24 June 1989, Seattle, WA, The Endocrine Society, abstract 1648. Gause, I. and Eden, S. (1986) Endocrinology 118, 119-124. Gavin, Ill, J.R., Roth, J., Neville, Jr., D.M., De Meyts, P. and Buell, D.N. (1974) Proc. Natl. Acad. Sci. U.S.A. 71, 84-88. Godowski, P.J., Leung, D.W., Meacham, L.R., Galgani, J.P., Hellmiss, R., Keret, R., Rotwein, P.S., Parks, J.S., Laron, Z. and Wood, W.I. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 8083-8087. Green, H., Morikawa, M. and Nixon, T. (1985) Differentiation 29, 195-198.
Grichting, G. and Goodman, H.M. (1986) Endocrinology 119, 847-854. Guzelian, P.S., Li, D., Schuetz, E.G., Thomas, P., Levin, W., Mode, A. and Gustafsson, J.A. (1998~ Proc. Natl. Acad. Sci. U.S.A. 85, 9783-9787. Hall, J.L., Dudley, P.R., Dobner, P.R., Lewis, S.A. and Cowan, N.J. (1983) Mol. Ceil. Biol. 3, 854-862. Hung, C.H. and Moore, W.W. (1984) Mol. Cell. Endocrinol. 35, 151-157. Kaplan, J. (1981) Science 212, 14-20. Khalili, K., Salas, C. and Weinmann, R. (1983) Gene 21, 9-17. Lesniak, M.A. and Roth, J. (1976) J. Biol. Chem. 251, 37203729. Leung, D.W., Spencer, S.A., Cachianes, G., Hammonds, R.G., Collins, C., Henzel, W.J., Barnard, R., Waters, M.J. and Wood, W.I. (1987) Nature 330, 537-543. Maiter, D., Underwood, L.E., Maes, M. and Ketelslegers, J.M. (1988a) Endocrinology 122, 1291-1296. Maiter, D., Underwood, L.E., Maes, M., Davenport, M.L. and Ketelslegers, M. (1988b) Endocrinology 123, 1053-1059. Mason, I.J., Murphy, D., Munke, M., Francke, U., Elliot, R.W. and Hogan, B.L.M. (1986) EMBO J. 5, 1831-1837. Mathews, L.S., Enberg, B. and Norstedt, G. (1989) J. Biol. Chem. 17, 9905-9910. Murphy, L.J. and Lazarus, L. (1984) Endocrinology 115, 16251632. Nakabayashi, N., Taketa, K., Miyano, K., Yamane, T. and Sato, K. (1982) Cancer Res. 42, 3858-3863. Nilsson, A., Carlsson, B., Mathews, L. and Isaksson, O.G.P. (1990) Mol. Cell. Endocrinol. 70, 237-246. Norstedt, G., Enberg, B., Moiler, C. and Mathews, L.S. (1990) Acta Paediatr. Scand. (Suppl.) 366, 79-83. Patel, M., Leevers, S.J. and Brickell, P.M. (1990) Int. J. Cancer 45, 342-346. Posner, B.I., Patel, B., Vezinhet, A. and Charrier, J. (1980) Endocrinology 107, 1954-1958. Rouiller, D.G. and Gorden, P. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 126-130. Roupas, P. and Herington, A. (1987) Endocrinology 121, 1521-1530. Roupas, P. and Herington, A.C. (1988) Mol. Cell. Endocrinol. 57, 93-99. Shelton-Earp, E.S., Austin, K.S., Blaisdel, J., Rubin, R.A., Nelson, K.G., Lee, L.W. and Grisham, J.W. (1985) J. Biol. Chem. 261, 4777-4780. Smith, W.C., Linzer, D.I.H. and Talamantes, F. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 9576-9579. Smith, W.C., Kuniyoshi, J. and Talamantes, F. (1989) Mol. Endocrinol. 3, 984-990. Snedecor, G.W. and Cochran, W.G. (1980) Statistical Methods, 7th edn., pp. 59-61, Iowa State University Press, Ames, IA. Tanaka, T., Shiu, R.P.C., Gout, P.W., Beer, B.T., Noble, R.L. and Friesen, H.G. (1980) J. Clin. Endocrinol. Metab. 51, 1058-1063. Tiong, T.S., Freed, K.S. and Herington, A.C. (1989) Biochem. Biophys. Res. Commun. 158, 141-148.
