Molecular and CeNuiar Endocrinology, 84 (1992) 145-154 Q 1992 Elsevier Scientific Publishers Ireland, Ltd. 0303-7207/92/$05.00

145

MOLCEL 02717

Thyroid hormone and androgen regulation of nerve growth factor gene expression in the mouse submandibular gland Marsha A. Black, Fleur-Ange Lefebvre, Louise Pope, Yvonne A. Lefebvre and Peter Walker Endocrine Research Laboratories, Loeb Institute for Medical Research, Ottawa CiGc Hmpital and Department Faculty of medicines Unil,ersity of Ottawa, Ottawa Kl Y 4E9, Canada

of Medicine,

(Received 16 September 1991; accepted 2 December 1991)

Key words: Thyroid hormone; Androgen: Nerve growth factor gene expression; Submandibular gland; (Mouse)

Summary The nerve growth factor (NGF) content of the mouse submandibular gland (SMGI is under hormonal control and is modulated by both thyroid hormones (TH) and androgens. The sexual dimorphism of the gland is well documented. In the adult male mouse, the SMG contains 10 times more NGF compared to the female. Conversely, castration of male mice reduces the SMG NGF levels to those found in control females. In order to determine the locus at which androgens and TH exert their effect on NGF gene expression in the SMG, steady-state NGF mRNA levels were determined. Daily treatment of adult female mice with TH for 1 week increased NGF mRNA levels &fold. Androgen treatment produced a 20-fold increase in SMG NGF mRNA, which was comparable to levels detected in the control adult male SMG. The effect of TH on NGF mRNA levels was time-dependent and coincided with the increase in NGF protein concentrations. At 48 h after a single TH injection, NGF mRNA levels (measured in SMG total RNA) increased 2-4-fold, while heteronuclear (hn) RNA levels were increased 1.52-fold. The NGF gene transcription rate was determined by run-on assays following TH treatment. A small but significant 2-fold induction by TH of NGF gene transcription was found at 24-48 h. Gytoplasmic RNA prepared from the same SMGs used in the run-on experiments was tested by Sl nuclease protection; NGF cytoplasmic RNA was increased 7-fold in the SMGs of females treated with TH 48 h previously. These results demonstrate that the effect of TH on NGF gene expression is due in part to an induction of NGF gene transcription. The discrepancies observed between transcription rate and mRNA levels suggest that the major effect of TH is at the post-transcriptional level, possibly mRNA stabilization. The time required to observe an induction of TH on NGF gene transcription is suggestive of an indirect effect, possibly through the induction by TH of another protein which in turn activates the NGF gene,

Introduction Correspondence to: Peter Walker, MD, FRCPC, Endocrine Research Laboratories, Loeb Institute for Medical Research, Ottawa Civic Hospital, 1053 Carling Avenue, Ottawa KlY 4E9, Canada. Tel. (613) 761-4667; Fax (613) 7296770.

The submandibular gland (SMG) of the mature mouse contains high concentrations of severa1 biologically active peptide growth factors,

146

including nerve growth factor (NGF; Bradshaw, 1978) and epidermal growth factor (EGF; Carpenter et al., 19791, in addition to renin (Dickinson et al., 1984) and several members of the kallikrein gene family (Clements, 1989). NGF is a M, 26,000 homodimer comprising two p subunits (Greene et al., 1980) that is essential for survival of dorsal root ganglion neurons in the fetus (Levi-Montalcini et al., 1951) and maintenance and survival of adrenergic neurons in the peripheral nervous system in postnatal rodents (LeviMontalcini et al., 1960). More recently, NGF has been shown to increase acetylcholinesterase activity in specific areas of the central nervous system, notably in the forebrain (Martinez et al., 1985; Mobley et al., 1985, 1986; Hefti, 1986; Korsching, 1986; Fischer et al., 1987). In the SMG, moduIation of NGF levels is dependent upon the hormonal status of the animal. Androgens are responsible for the sexual dimorphism of the gland and its content of NGF (Byyny et al., 1974; Ishii et al., 1975). Levels of SMG NGF in the male mouse are an order of magnitude greater than those in female mice. Castration of male mice reduces the levels of both growth factors and androgen replacement restores NGF concentrations to control values. Similarly, androgen treatment of female mice induces an increase in SMG NGF to those observed in intact male mice. Thyroid hormones increases SMG NGF content and concentration in both the immature and mature mouse (Aloe et al., 3980a; Walker et al., 1980, 1981; Gresik et al., 1981). This effect has also been observed in androgen-resistant testicuIar feminized mice, thus excluding the androgen receptor as the mediator of the thyroid hormone action in the SMG (Aloe and Levi-Montalcini, 1980b; Hosoi et al., 1981). We have previously demonstrated the presence of specific nuclear T, receptors in the mouse SMG and shown that the SMG EGF response to triiodothyronine CT,) administration is both time- and dose-dependent (Walker et al., 1982). Furthermore, administration of T, to immature mice increased the numbers of mRNA species in the SMG coding for EGF as assessed by immunoprecipitation of translation products in an in vitro wheat germ translation system (Walker, 1986). We concluded

that thyroid hormones induce SMG EGF and NGF accumulation by increasing their relative rates of biosynthesis at a pretranslational level. To test this hypothesis directly, we examined the effect of thyroid hormones on expression of the NGF gene by measuring NGF gene transcription rates in nuclear run-on studies and the steadystate levels of NGF mRNA in the adult mouse SMG.

Materials and methods

Animals

Outbred mature female Swiss-Webster mice (Simonsen Laboratories, Gilroy, CA, USA) were housed in a temperature- (25 + 1°C) and lightcontrolled (14 h light) environment. They were fed standard mouse laboratory chow and had free access to water. In a first series of experiments, the effect of thyroid hormones on steady-state NGF mRNA levels was examined and compared to the effect of androgens. Adult female mice received daily se. injections of T4 (40 pg/lOO g body weight) for 6 days or a single injection of testosterone cypionate (10 mg/lOO g body weight). Control female and male mice were uninjected in this experiment. On day 7 (24 h after the last T4 injection), the mice were killed by CO, narcosis and blood was removed from the inferior vena cava. To examine the time course of the thyroid hormone effect, female mice received a single S.C. injection of T, (100 g pg/lOO g body weight). This dose of T, is based on a previous report indicating an ED,, of 60 pg T,,‘lOO g body weight for EGF and a maximal effect of T, at 100 pg/lOO g body weight (Walker et al., 1982). Control animals received equivalent volumes of saline vehicle. T,-treated and control mice were killed at 0, 12, 24, 36, 48, 60, 72 and 96 h after T, injection. SMGs were rapidly and aseptically removed and snap frozen on methanol/dry ice. A portion of the gland was removed for the measurement of NGF concentrations by radioimmunoassay (RIA). For the nuclear run-on experiments, mice were killed at 0, 12,24 and 48 h after T, injection.

147 RNA isolation

Total RNA was isolated from SMGs by the guanidine-HCl (Chirgwin et al., 1979) or the LiCl (Auffray and Rougeon, 1980) method. The yield of total RNA from the SMGs was approximately 4 pg/mg tissue. In other instances, cytoplasmic RNA was isolated (Armstrong et al., 1986). Integrity of the isolated RNA was assessed by visualization under ultraviolet light following electrophoresis on 1.2% agarose gels and staining with ethidium bromide (ratio of 28S/l8S RNA species r 2). Analysis of submandibular gland hnRNA

Submandibular gland nuclei were prepared by a modification of the citric acid procedure (Vaulont et al., 1986). Nuclear RNA was isolated by CsCl centrifugation as described by Narayan et al. (1984) or by proteinase K digestion and subsequent phenol and chloroform extractions as described for the isolation of cytoplasmic RNA (Armstrong et al., 1986) and subjected to Northern blot hybridization or Sl nuclease protection analysis. Nuclear run-on transcription assays

Submandibular gland nuclei were isolated according to the method described by Gurr and Kourides (1985). 4.5 ~1 nuclei (l-4 X 10’ nuclei/ml) were added to a 100 ~1 reaction mix containing 16% glycerol, 20 mM Tris-Cl, pH 7.9, 1 mM MnCl,, 2.5 mM MgCl,, 5 mM NaF, 12.5 mM NH&SO&, 1 mM ATP, GTP, and CTP, 10 FM creatine phosphate, 20 p-g/ml creatine phospho~nase, 2 mM dithiothreitol and 250 PCi [32PlUTP (3000 Ci/mmol; Arnersham) and 0.1 mg/ml heparin (Turcotte et al., 1985). Transcription reactions were incubated at 26°C for 30 min, followed by the addition of l/5 volume of 5 X DNase buffer (10 mM Hepes, pH 7.4, 2.5 mM MgCI,, 5 mM CaCl,, 5 mM MnCl,) and RNasefree DNase I (10 pg/ml; BRL) and the incubation continued for 10 min. The reactions were adjusted to 1 X SET (10 mM Tris-Cl, pH 7.5, 5 mM EDTA, 1% sodium dodecyl sulfate (SDS) and 200 pg tRNA and 10 pg/ml proteinase K (Sigma) were added and incubated for 30 min at 45°C. The reactions were extracted twice with phenoI/~hloroform/isoamyl alcohol (25 : 24 : 1)

and precipitated with 30% trichloroacetic acid (TCA), containing 30 mM Na,P,O,. The precipitate was collected on a filter (Millipore; HA 0.45 FM) and washed with 3% TCA containing 22 mM Na,P,O,. The filters were subsequently digested with DNase I(10 pg/ml), 100 mM Tris-Cl, pH 8, 25 mM Hepes, 5 mM MgCl,, 1 mM MnC12 and 1 mM CaCl, for 30 min at 37°C. The labeled RNA was eluted from the filters by adding 10 X SET and incubating for 10 min at 65”C, followed by a second incubation in 1 X SET to rinse the filters. Following a final phenol/~hloroform extraction, the labeled RNA was precipitated overnight with 2 volumes of ethanol and 120 mM NaCl and resuspended in 10 mM Tris-Cl pH 8, 100 mM NaCl, 10 mM EDTA and 0.2% SDS. The j2P-labeled RNA was hybridized to fulllength NGF (pmngf6) and tubulin (MAT1.l) cDNA sequences fixed to nitrocellulose. Hybridization of the 32P-labeIed transcripts

Plasmid DNAs (6 pg/filter) were linearized by digestion with Hind111 (Kafatos et al., 1979) and loaded onto nitrocellulose filters using a dot-blot apparatus (Hybridot manifold, BRL) and baked for 2 h at 80°C in vacua. Following a prehybridization for at least 3 h at 65°C in 50 mM Hepes, pH 7.4, 0.3 M NaCl, 10 mM EDTA, 0.2% SDS, 1 mg/ml yeast tRNA, 1% Na,P,O,, 0.1% Ficol and 1% polyvinylpolypyrrolidone, the fluid was removed and the filters were hybridized with 5-10 x lo6 cpm of the 32P-labeled RNA for at least 36 h at 65°C in 50 mM Hepes, pH 7.4, 0.3 M NaCI, 10 mM EDTA, 0.2% SDS, 100 pg/ml yeast tRNA, 0.1% Na4P20,, 0.02% Ficol, 0.02% pol~inylpolypyrrolidone and 0.02% RNAse-free bovine serum albumin (BSA) (BRL). In order to monitor the hybridization efficiency, the appropriate [3H]cRNAs were included in the hybridization reaction. After completion of the hybridization, the filters were washed 3 times in 2 X SSC (1 X SSC =O.l M NaCl and 0.015 M sodium citrate1 at 65°C for 30 min and then digested with Tl RNase (5 U/ml> and 2.5 pug/ml pancreatic RNase in 2 x SSC at 37°C for 30 min. After washing twice with 2 X SSC at room temperature for 30 min, the filters were digested with 100 pg/ml proteinase K in 2 x SSC containing 0.5% SDS at 37°C for 30 min. The filters were

14X

washed twice in 2 x SSC at room temperature for 1.5 min. Levels of hybridization were determined firstly by autoradiography and densitometric scanning and secondly by liquid scintillation counting after elution of the RNA from the filters. Nonspecific hybridization was determined by measuring hybridization to the appropriate vectors fixed to nitrocellulose filters, i.e. pBR322 and PUC8 for pmngf6 and MAT1.l, respectively. Synthesis

of[“H]RNAs

The NGF (916 bp: kindly provided by Dr. Graeme I. Bell (University of Chicago)) and tubulin (1437 bp) inserts were excised from the plasmids pmngf6 (Scott et al., 1983) and MAT1.l by digestion with Pst I and purified by electrophoresis in low melting point agarose. Complementary [sH]RNAs were prepared as described by Gurr and Kourides (1985). Hybridization

The pmngf6 probe was either nick-translated (spec. act. = l-4 x 10’ cpm/wg DNA) using commercially available kits (Bethesda Research Laboratories, Bethesda, MD, USA) or singlestranded probes were prepared as follows: Pst I fragments of pmngf6 were subcloned into M13mp9, radiolabeled by primer extension with DNA polymerase I (Klenow fragment) and restricted with ScaI to yield a 248 b single-stranded probe that was purified by 4-6% polyacrylamide gel electrophoresis. For tubulin, the 1437b PstI insert of MAT1.l was subcloned into pTZ19R, labeled by primer extension and cut with BglI resulting in a “‘P-labeled 575 b probe. For Northern analysis, 10 /Lg total RNA or hnRNA were glyoxylated and electrophoresed on 1.2% agarose gels and transferred to nitrocellulose filters. The filters were prehybridized at 42°C for 18 h in 50% formamide, 4 X SSC, 50 mM NaH,PO, pH 7.0, 5 x Denhardt’s solution, 500 pg/ml sheared salmon sperm DNA and 0.1% SDS. Hybridization was carried out for a further 24 h at 42°C in 50% formamide, 4 X SSC, 50 mM NaH,PO, pH 7.0, 10% dextran sulphate, 1 x Denhardt’s solution, 100 pg/ml sheared salmon sperm DNA, 0.1% SDS and lo5 to 10h cpm/ml “‘P-labeled NGF or tubulin probes. The blots were washed stringently (two washes at 42°C for

20 min with 5 X SSPE (1 x SSPE = 1.15 M NaCl, 0.01 M NaH,PO,, 1 mM EDTA); two washes at 42°C for 20 min with 1 x SSPE, containing 0.1% SDS; one wash at room temperature for 30 min with 0.1 x SSPE, containing 0.1% SDS). The filters were exposed to Kodak XAR-2 X-ray film in the presence of Cronex intensifying screens at - 70°C. Sl nuclease protection analysis

The ‘2P-labeled 248 b NGF fragments was hybridized to 20 pg hnRNA in a reaction mixture containing 0.4 M NaCl, 10 mM Pipes, pH 6.5, 1 mM EDTA and 80% formamide. After an overnight incubation at 42°C 100 units of Sl nuclease (BRL) were added (in buffer containing 30 mM NaOAc, pH 4.5, 0.25 M NaCl, 1 mM ZnSO,, 5% glycerol, diluting the original hybridization mix 8-fold). After Sl nuclease digestion at 37°C for 30 min, the reaction was diluted to a final concentration of 200 mM Tris-Cl, pH 8.0, and 20 mM EDTA. After precipitation with ethanol, the samples were resuspended in formamide loading dye, denatured by heating at 100°C for 4 min and subjected to electrophoresis through an 8 M urea-6% polyacrylamide gel. After electrophoresis, the gel was soaked in 10% acetic acid, 5% methanol for 10 min, dried and observed by autoradiography. The protected 219 b NGF fragment was quantitated by densitometry. RlAs

SMG NGF concentrations were measured by a sensitive and specific double-antibody RIA as previously described (Walker et al., 1980). [12’I]NGF was purchased from New England Nuclear; unlabeled NGF was purchased from Collaborative Research (Waltham, MA, USA). NGF concentrations were expressed relative to SMG protein levels as measured by the method of Lowry et al. (1951) using bovine serum albumin as standard. Statistical analysis

Statistical analysis was carried out using analysis of variance and Student’s t-test for unpaired observations.

149

Female

A

Results T3 increases SMG NGF mRNA levels

C

SMG NGF concentrations of T,-treated female mice exceeded those of intact female mice by 4-fold (p < 0.005, Table 0, whereas the mean increase induced by testosterone cypionate was approximately 20-fold (p < 0.001). Testosteronetreated female mice had SMG NGF concentrations similar to those of control male mice. Hybridization of 32P-labeled pmngf6 to total RNA extracted from SMGs of control, T4- and testosterone cypionate-treated female mice and control male mice revealed a single band of 1.3 kb (Fig. 1A) with a significantly greater intensity of the hybridization signal in SMG RNA isolated from T4- and testosterone cypionate-treated females and untreated males compared to control female mice (Fig. 1B). The increases in intensity of the pmngf6 hybridization signal correlated with increases in SMG NGF protein concentrations in T4- and testosterone-treated female and untreated male mice. The response to testosterone treatment was significantly greater than the effect of T4.

T4

Male T

T

B

Fold Increase

Female C T4 T Male

C pg

RNA

Time course of the effect of T3

Fig. 1. Northern blot (panel A) and dot-blot (panel B) hybridization of 32P-labeled pmngf6 (to total RNA from SMG of control, T,-treated and testosterone cypionate CT)-treated female mice compared to intact control male mice (0. Experimental details are as in Materials and methods.

In mature female mice, a single injection of T, led to a time-dependent increase in SMG NGF concentrations (Fig. 2B), similar to that previously reported for EGF (Walker et al., 1982) with an apparent 24-36 h delay before the accumulation of newly synthesized NGF. The maximum

mean increase in SMG NGF concentration was 5-fold compared to control mice. Levels of NGF reached a plateau after 72 h.

TABLE

1

SERUM TERONE

T., AND T, CONCENTRATIONS AND SMG CYPIONATE-TREATED FEMALE MICE

Numbers

in parentheses

Female mice Control (5) T4 (5) Testo (5) Male mice Control (5)

represent

the numbers

of animals.

NGF

Values

CONCENTRATIONS

comprise

IN CONTROL,

mean f SEM

T4

T3

&g/d0

(ng/dl)

NGF (ng/mg

protein)

3.6 * 0.3 12.2k1.4 ** 3.0 f 0.3

84.6+ 6.3 333.15 27.4 * * 78.8k 5.0

379+ 2165+ 8764*

108 321 * 462 **

3.6+0.3

* p < 0.005; * * p < 0.001 compared

72.0+ to control

female

9.0

mice.

T4- AND

Fold increase

9525+1472

**

1 6 23 25

TESTOS-

150

ment of the expected size of 219 bp was found following hybridization of the NGF 248 b probe with hnRNA and Sl nuclease digestion (Fig. 3A). T, induced a 1.8- and 15fold increase in the levels of NGF hnRNA at 24 and 48 h following T, injection, respectively (Fig. 38).

A Time (h)

0

24

36

46

60

T3 increases NGF gene transcription rate

In nuclear transcription run-on assays, [s2P]UTP elongated transcripts from SMG nuclei were hybridized to full-length pmngf6 (NGF) and MAT1.l (tubulin) cDNAs as well as the appropriate control vectors fixed to nitrocellulose filters. A 240 b

219 b time (h) O!

0

I

12

I

24

I

36

I

46

I

60

I

72

,

96

Time (h) after T3 injection

Fig. 2. Time course of increase of SMG NGF protein panel B) and mRNA (U n ; panel A and (0 -0; B) levels following a single injection of T, (100 kg/100 g body weight) at t = 0 h. Results arc expressed as fold increases over control levels ( + SE) obtained at the same time points. n 2 4 per time point.

xlo-

200.

Steady-state NGF mRNA levels were stable for 24 h after T, injection and then rose progressively through 60-72 h post-injection (Fig. 2B). Levels of NGF mRNA declined after 72 h. By 48 h after T, injection, NGF mRNA levels were 2to 4-fold above control (Fig. 2A and B). The time-dependent increase in SMG NGF mRNA levels preceded the increase in NGF protein concentrations. T3 increases NGF hnRiVA levels

Hybridization of 32P-labeled pmngf6 to SMG hnRNA revealed the presence of three molecular weight species of NGF RNA. Two predominant species of 1.3 and 1.8 kb and a minor band of 3.2 kb were noted (data not shown). All responded similarly to T, administration. A protected frag-

IW-

0

T3 Treatment

Fig. 3. Effect of T, on NGF hnRNA levels at 0, 12, 24 and 48 h as determined by Sl nuclease protection analysis. Adult female mice were injected with T, at t = 0 h. 20 yg SMG hnRNA was hybridized overnight to the 248 b [s*PlNGF probe. After Sl nuclease digestion, the products were fractionated through an 8 M urea-6% polyacrylamide gel and exposed to autoradiography for 62 h. A representative autoradiogram is shown in panel A. The results of Sl nuclease protection analysis of several preparations of hnRNA’are expressed graphically in panel B. The results are presented as the mean + SE (n = 6 nuclei preparations). * p < 0.025; ** p < 0.005 compared to t = 0 h.

151

Control hybridization experiments, using filters to which single-stranded sense NGF sequences were fixed, showed no hybridization above background levels indicating that transcription originating from the nontemplate strand had not occurred (data not shown). In some experiments, 1 pg/ml cu-amanitin was included in order to assess transcription by RNA polymerase II (Lindell et al., 1970). cu-Amanitin reduced total, NGF and tubulin gene transcription by 52%, 85% and 90%, respectively, suggesting that the NGF and tubulin genes were primarily transcribed by RNA polymerase II.

NGF

A

rime(h)

0

12

24

Tu~lin

46

0

12

24

NGF

A

C

NGF

Tubulin

1.3 kb

Time (h)

0

Time(h)

0

1.6 kb

4% Tubulin

Time (h)

46

48

0 12 24 48

0 12 24 48

Fig. 5. Effect of TX on NGF and tub&n gene transcription at 0, 12, 24 and 48 h after T3 administration. Panel A: Representative autoradiograms obtained after hybridization of the ‘*P-labeled transcripts to plasmids containing NGF and tubulin cDNA sequences. Panel B: The rates of NGF and tubulin gene transcription were calculated as described in the legend to Fig. 48. The results are expressed as mean k SE. * p < 0.005 compared to t = 0 h. Panel C: Effect of Ts on cytoplasmic NGF and tubulin mRNA levels after T, administration. After sedimentation of nuclei for run-on transcription assays, cytoplasmic RNA was isolated from the supernatant. 20 pg RNA was size fractionated on a 1% agarose gel containing 2.2 M formaldehyde, transferred to nylon membranes and hybridized sequentially to the 248 b NGF and 575 b tubulin probes. The membrane was subjected to autoradiography for 10 and 66 h for NGF and tubulin, respectively.

TUBULIN NGF Fig. 4. Effect of Ts on NGF and tubulin gene transcription. Panel A: Representative autoradiograms of NGF cDNA- and tubulin cDNA-containing filters after hybridization of the 32P-labeled transcripts. Panel B: The effect of T, treatment on the rates of NGF and tubulin gene transcription. The rate of transcription was calculated as the specific counts hybridized (cpm bound to NGF or tubulin vector minus cpm bound to control vector) divided by the input 13’P]RNA cpm and corrected for the preparations). * p < 0.05 compared to controls.

Representative autoradiograms of the filters obtained after hybridization to the 32P-labeled transcripts are shown in Fig. 4A. T3 administration significantly increased the transcription rate of the NGF gene to l.4-fold (p < 0.025) above the values obtained at t = 0 h. Results were similar when expressed as absolute transcription rates or as a function of paired tubulin transcription rates which did not significantly change following T3 injection (Fig. 4B). In order to study the time course of thyroid hormone action on NGF gene transcription in the

SMG, nuclear run-on transcription assays were carried out at 0, 12, 24 and 48 h following T, injection. T, increased the rate of NGF gene transcription by - 2-fold at both 24 and 48 h after T, (Fig. 5A and B), while no significant increase in NGF gene transcription was observed at 12 h. There was no significant change in tubulin gene transcription at any time. Cytoplasmic mRNA isolated from the supernatant retained after the nuclei were sedimented for the run-on transcription assays was hybridized against both the 248 b NGF and 575 b tubuIin probes. In contrast to the - 2-fold increase in NGF gene transcription seen at both 24 and 48 h, T, increased the steady-state levels of NGF cytoplasmic mRNA by 2.5- and 7.3-fold at 24 and 48 h, respectively (Fig. 50. Although T, did not alter tubuhn gene transcription, thyroid hormone treatment increased tubulin mRNA levels by - 1.5 and - I.&fold at 24 and 48 h, respectively. Discussion Our results support and extend previous data which demonstrate that thyroid hormones induce NGF gene expression in the SMG of the female mouse (Aloe and Levi-Montalcini, 1980a; Gresik et al., 1981; Walker et al., 1981). The presence of specific nuclear T3 receptors in SMG nuclei of adult female mice supports the hypothesis that thyroid hormones modulate NGF biosynthesis in the SMG, presumably through the T, receptor (Walker et al., 1982). Thus, the data, taken together, support the conclusion that thyroid hormones increase NGF gene expression at a pretranslational level and argue for a physiologic role for thyroid hormones in regulation of SMG NGF gene expression. However, the precise mechanism of thyroid hormone action in the induction of NGF gene expression is unclear. The present studies indicate that T4 treatment of intact adult female mice results in a significant increase in SMG NGF mRNA levels. A single injection of T, induces a time-dependent increase in the steady-state levels of NGF mRNA species, followed by a detectable accumulation of newly synthesized NGF protein by 24 h post-injection. The time course of the T, effect observed in the

present study is, therefore, in excellent agreement with that previously reported for EGF protein (Walker et al., 1982). Since the primary mechanism of thyroid hormone action is generally thought to involve the interaction of the hormone-receptor complex with a specific DNA sequence resulting in the altered transcription of the target gene, we sought to determine whether thyroid hormones increased NGF gene transcription. The results from nuclear run-on assays showed that T, treatment consistently increased the rate of NGF gene transcription in the female mouse SMG by 1.5-2-foId at 48 h after administration of the hormone. This increase correlated well with the effect of T, on NGF hnRNA levels, where a 1.5-2-fold effect of T, was observed. The effect of T, on NGF gene transcription and hnRNA levels, however, was consistently lower than its effect on NGF total mRNA levels, where a 3-4-fold increase was observed at 48 h. The discrepancy between the T,-induced concentration of NGF mRNA and the rate of transcription of the NGF gene is compatible with an additional pretranslational control affecting the stability of the nuclear or cytoplasmic NGF transcripts or the rate of processing of the primary NGF transcript, as suggested for T, induction of the liver malic enzyme (Dozin et al., 1986; Song et al., 1988) and Spot 14 (Jump, 1989) genes. Since NGF hnRNA levels are not significantly increased beyond the rate of NGF gene transcription, a significant effect on nuclear RNA stability or processing is unlikely. Rather, the observations point to the possibility of a thyroid hormone-mediated increase in cytoplasmic NGF transcript stability. Posttranscriptional regulation by thyroid hormones has also been demonstrated for the rat hepatic 3-hydrop-3-methylgluta~l coenzyme A reductase, where the reductase mRNA was shown to be 4-&fold more stable in thyroid hormone-treated animals (Simonet et al., 1988, 1989). Therefore, in view of the discrepant effects of T, on NGF mRNA levels and the rate of NGF transcription, it is probable that, in the mature mouse, the major locus of thyroid hormone action is posttranscriptional, more specifically at the level of cytoplasmic NGF transcript StabiIization.

153

No effect of T, on the rate of NGF gene transcription was observed before 24 h. The long lag time observed before an increase in NGF gene transcription and mRNA levels is apparent makes the assumption of a direct transcriptional effect of T, difficult. For example, GH gene transcription in the pituitary is activated within minutes of exposure to thyroid hormone (Yaffe et al., 1984). Unlike thyroid hormone responsive genes such as GH (Glass et al., 1987; Koenig et al., 1987; Wight et al., 1988; Ye et al., 1988), hepatic malic enzyme (Petty et al., 1990) and Spot 14 (Zilz et al., 19901, no thyroid hormone response elements have been reported in the regulatory portions of the NGF gene to date (Zheng et al., 1988). Although 5 kb of the mouse NGF promoter have been cloned, only 1 kb has been sequenced. Although it is possible that thyroid hormone response elements may lie further upstream (Zilz et al., 1990), the present observations suggest that the effects of thyroid hormones in increasing NGF gene expression are indirect and may involve the synthesis of one or more intermediate factors that would, in turn, activate NGF gene transcription and/or promote NGF mRNA stability. The indirect effects of thyroid hormone on NGF gene expression contrast with the more direct action of this hormone on renin gene expression in the GCT cells of the female mouse SMG, where a stimulatory effect of thyroxine on renin gene transcription has been observed as early as 2 h (Tronik et al., 1988). Moreover, the presence of conserved motifs in the 5’ flanking region of the renin genes, similar to the thyroid hormone responsive elements of the rat GH promoter, supports the suggestion that thyroid hormones act directly on the SMG renin gene to induce transcription (Tronik and Rougeon, 1988). How do thyroid hormones induce the expression of the NGF gene? NGF gene expression is restricted to the granular convoluted tubule (GCT) cells of the mouse SMG (Barka, 1980). The mouse SMG is sexually dimorphic, which is reflected by the androgen-induced increase in the number of GCT cells in female mice and their disappearance in castrated male mice (Chretien, 1977; Barka, 19801. In the female mouse, thyroid hormones and androgens induce the differentia-

tion of these cells (Aloe and Levi-Montalcini, 1980a). In the case of thyroid hormones, this effect is particularly evident in the immature mouse (Gresik et al., 1980; Chabot et al., 1987). In the adult mouse, androgens appear to induce the differentiation of GCT cells by catalyzing the transition between the (uninducedl striated duct cells to the (induced) GCT cells (van Leeuwen et al., 1987). It has been suggested that androgens (Okamoto et al., 1985) and thyroid hormones stimulate the proliferation of these and other cells in the adult female mouse SMG (Gresik et al., 1981). However, whether proliferation in response to thyroid hormones is an important factor in increasing the number of GCT cells remains a matter of dispute, since it has also been reported that, in contrast to androgens, thyroid hormones induce only the differentiation of these cells (Okamoto et al., 1985). The present observations indicate that the effect of androgens is more robust compared to thyroid hormones, suggesting additional mechanisms for androgens in increasing NGF gene expression in the SMG. In contrast to renin, where a direct transcriptional effect of the hormone is likely, thyroid hormones may increase NGF gene expression by inducing the expression of other transcriptional factors in the SMG. The existence of an APl consensus sequence (Angel et al., 1987) in the first intron of the NGF gene (Zheng and Heinrich, 1988) suggests that the immediate-early response genes which encode transcription factors, such as c-&n and c-fos, could play a role in regulating NGF gene expression. Whether these transcription factors play a role in mediating the effect of T, on SMG NGF gene expression requires further study. In summary, T, increases NGF gene expression in the mouse submandibular gland, at least in part, by increasing NGF gene transcription. However, in view of the consistent discrepancies we have observed in comparing the rate of NGF gene transcription and the mRNA levels, it is possible that T, also increases NGF gene expression in a manner that is consistent with enhanced processing and/or stabilization of the cytoplasmic NGF transcripts. Our observations also suggest that a direct transcriptional effect of T, on NGF gene expression is unlikely. Rather, it is

1.54

more likely that T, induces the expression of another protein(s) which, in turn, transcriptionally activate(s) the NGF gene. References Aloe, L. and Levi-Montalcini, R. (1980a) Exp. Cell Res. 125, 15-22. Aloe, L. and Levi-Montalcini, R. (1980bl Cell Tissue Res. 205, 19-29. Angel, P., Imagawa, M., Chiu, R., Stein, B., Imbra, R.J., Rahmsdorf, H.J., Jonat, C., Herrlich, P. and Karin, M. (1987) Cell 49, 729-739. Armstrong, R.D., Lewis, M., Stern, S.G. and Cadman, E.C. (19861 J. Biol. Chem. 261, 7366-7371. Auffray, C. and Rougeon, F. (1980) Eur. J. Biochem. 107, 303-314. Barka, T. (1980) J. Histochem. Cytochem. 28, 836-859. Bradshaw, R.A. (1978) Annu. Rev. Biochem. 47, 191-216. Byyny, R.L., Orth, D.N., Cohen, S. and Doyne, ES. (1974) Endocrinology 95, 776-782. Carpenter, G. and Cohen, S. (19791 Annu. Rev. Biochem. 48, 193-216. Chabot, J.G., Walker, P. and Pelletier, G. (1987) Cell Tissue Res. 248, 351-358. Chirgwin, J.M., Przbyla, A.E., MacDonald, R.J. and Rutter, W.J. (1979) J. Biol. Chem. 18, 5294-5299. ChrCtien, M. (1977) Int. Rev. Cytol. 50, 333-396. Clements, J.A. (1989) Endocr. Rev. 10, 393-419. Dickinson, D.P., Gross, K.W., Piccini, N. and Wilson, C.M. (19841 Genetics 108, 651-667. Dozin, B., Magnuson, M.A. and Nikodem, V.M. (1986) J. Biol. Chem. 261, 10290-10292. Fischer, W., Wictorin, K., Bjorklund, A., Williams, L.R., Varon, S. and Gage, F.H. (1987) Nature 329, 65-68. Glass, C.K., France, R., Weinberger, C., Albert, V.R., Evans, R.M. and Rosenfeld, M.G. (1987) Nature 329, 738-741. Greene, L.A. and Shooter, E.M. (19801 Annu. Rev. Neurosci. 3, 353-402. Gresik, E.W. and Barka, T. (1980) Am. J. Anat. 159, 177-185. Gresik, E.W., Schenkein, I., van der Noen, H. and Barka, T. (19811 Endocrinology 109, 924-929. Gurr, J.A. and Kourides, I.A. (1985) DNA 4, 301-307. Hefti, F. (19861 J. Neurosci. 6, 2155-2162. Hosoi, K., Tanaka, I. and Veha, T. (1981) J. Biochem. 90, 267-270. Ishii, D.N. and Shooter, E.M. (1975) J. Neurochem. 25, 843851. Jump, D.B. (1989) J. Biol. Chem. 264, 4698-4703. Kafatos, F.C., Jones, C.W. and Efstratiadis, A. (1979) Nucleic Acids Res. 7, 1542-1552. Koenig, R.J., Brent, G.A., Warne, R.L., Larsen, P.R. and Moore, D.D. (1987) Proc. Natl. Acad. Sci. USA 84, 56705674.

Korsching, S. (19861 Trends Neurosci. 9, 570-573. Levi-Montalcini, R. and Booker, B. (1960) Proc. Natl. Acad. Sci. USA 49, 384-391. Levi-Montalcini, R. and Hamburger, V. (1951) J. Exp. Zool. 116, 323-361. Lindell, T.J., Weinberg, F., Morris, P.W., Roeder, R.G. and Rutter, W.J. (1970) Science 170, 447-449. Lowry, G.H., Rosebrough, N.Y., Farr, A.L. and Randall, R.L. (1951) J. Biol. Chem. 193, 265-275. Martinez, H.J., Dreyfus, C.F., Jonakait, G.M. and Black, I.B. (19851 Proc. Natl. Acad. Sci. USA 82, 7777-7781. Mobley, W.C., Rutkowski, J.L., Tennekoon, G.I., Buchanan, K. and Johnston, M.V. (1985) Science 229, 284-286. Mobley, W.C., Rutkowski, J.L., Tennekoon, G.I., Gemski, J., Buchanan, J. and Johnston, M.V. (1986) Mol. Brain Res. 1, 53-62. Narayan, P., Liaw, C.W. and Towle, H.C. (1984) Proc. Natl. Acad. Sci. USA 81, 4687-4691. Okamoto, S., Ito, M., Kitamura, Y. and Matsumato, K. (1985) Cell Tissue Kinet. 18, 583-588. Petty, K.J., Desvergne, B., Mitsuhashi, T. and Nikodem, V.M. (19901 J. Biol. Chem. 265, 7395-7400. Scott, J., Selby, M.J., Urdea, M., Quiraga, M., Bell, G.I. and Rutter, W.J. (1983) Nature 303, 538-541. Simonet, W.S. and Ness, G.C. (19881 J. Biol. Chem. 263, 12448-12453. Simonet, W.S. and Ness, G.C. (1989) J. Biol. Chem. 264, 569-573. Song, M.-K.H., Dozin, B., Grieco, D., Rail, J.E. and Nikodem, V.M. (1988) J. Biol. Chem. 263, 17970-17974. Tronik, D. and Rougeon, F. (19881 FEBS Lett. 234, 336-340. Turcotte, B., Guertin, M., Chevrette, M. and Belanger, L. (1985) Nucleic Acids Res 13, 2387-2398. van Leeuwen, B.H., Penschow, J.D., Coghlan, J.P. and Richards, R.I. (1987) EMBO J. 6, 1705-1713. Vaulont, S., Munnich, A., Decaux, J.-F. and Kahn, A. (19861 J. Biol. Chem. 261, 7621-7625. Walker, P. (1986) Biochem. Cell Biol. 64, 290-296. Walker, P., Weichsel, M.E.J., Guo, S.M., Fisher, D.A. and Fisher, D.A. (19801 Brain Res. 186, 331-341. Walker, P., Weichsel, M.E.J., Hoath, S.B., Poland, R.E. and Fisher, D.A. (19811 Endocrinology 109, 582-587. Walker, P., Coulombe, P. and Dussault, J.H. (1982) Endocrinology 111, 1133-1139. Wight, P.A., Crew, M.D. and Spindler, S.R. (19881 Mol. Endocrinol. 2, 536-542. Yaffe, B.M. and Samuels, H.H. (1984) J. Biol. Chem. 259, 6284-6291. Ye, Z.-S., Forman, B.M., Aranda, A., Pascual, A., Park, H.-Y., Casanova, J. and Samuels, H.H. (1988) J. Biol. Chem. 263, 7821-7829. Zheng, M. and Heinrich, G. (1988) Mol. Brain Res. 3, 133-140. Zilz, N.D., Murray, M.B. and Towle, H.C. (1990) J. Biol. Chem. 265. 8136-8143.