Main

GH is an important hormone for growth and anabolism in many mammalian tissues. Its participation in fetal and early neonatal growth, however, has been controversial. Growth in the fetus is thought to be predominantly pituitary independent. Pituitary dependence begins some time after birth and increases after infancy (1). This apparent delay in GH response occurs even though GH plasma levels in the immediate postnatal period are high in both normal (2, 3) and transgenic mice (4). Moreover, the onset of growth retardation in models of GH deficiency is usually not seen until after d 15 (5–7). Mice transgenic for either GH-releasing hormone or GH genes exhibit supernormal growth but only after wk 3 of life (4). There is, however, evidence that GH is active in early life. Treatment with rat GH antiserum reduced growth in d-2 rats (8), and early inhibition of growth was noted after infant hypophysectomy (9). IGF-I mRNA, an important GH-responsive gene, is low in rats until after d 15 of life (4). Thus, it is not clear whether GH is active in promoting growth or gene expression in early neonatal life. We, therefore, set out to determine at what age and through what mechanisms gene expression moderated by GH occurs in neonatal rats.

The liver is a well-defined GH target organ in the adult rat. Two rat hepatic Spi, Spi 2.1 and 2.3, are GH responsive (10, 11) and developmentally regulated (10, 12). Their expression in adult animals is greatly reduced by hypophysectomy and can be restored to 40% of normal values by administration of GH alone. Their rapid induction by GH is direct and not mediated by IGF-I (13). In previous work (14, 15), we noted an increase and decrease in Spi 2.1 and 2.3 gene expression from fetal d 19 to neonatal d 11. The sum of hepatic content of Spi 2.1 and 2.3 mRNA of normal d-2 neonates (35% of the adult value) was considerably higher than that in d-11 normal neonates (5% of the adult value). This pattern parallels that of serum GH levels in infant rats in which GH is first detectable in fetal serum on d 19 of gestation, reaches a peak on d 1 of life, and declines thereafter (2). An active chromatin configuration in the Spi 2.1 promoter is also present in fetuses and neonates (15). We, therefore, hypothesized that GH was responsible for the changes in Spi 2.1 and 2.3 expression, that d-2 neonates are capable of responding to GH, and that monitoring the mRNA levels of Spi 2.1 and 2.3 and GHR would be a useful method for evaluation of neonatal GH action.

Recent advances in signal transduction research have broadened our knowledge of the GH signaling pathway. A general model has been proposed in which the GH signal cascade begins with GH binding to GHR on the cell surface, leading to GHR dimerization. Members of a family of transcription factors referred to as Stat proteins are recruited to the receptor complex and phosphorylated by associated activated Janus protein-tyrosine kinases (16). Translocation of phosphorylated Stat proteins to the nucleus is followed by binding to cognate DNA response elements of target genes such as Spi 2.1. We have previously characterized a GHRE extending from −147 to −103 in the 5′ flanking region of the Spi 2.1 gene that is responsible for its induction by GH (13) and have purified a GHINF from rat liver that binds to the GHRE in a state-specific manner (17). This complex contains Stat5 (18). Binding of GHINF to the GHRE is necessary to stimulate Spi 2.1 transcription. We anticipated that measuring the abundance of Stat5 protein and activated Stat5 in hepatic nuclear extracts from neonatal rats would indicate whether changes in the abundance of this transcription factor contributed to changes in GH-responsive gene expression.

In addition to Spi 2.1, IGF-I is also a well-characterized GH-responsive hepatic gene. The interaction between GH, GHR, and IGF-I is an important aspect of the somatotropic axis. Examining the ontogeny of GH responsiveness of IGF-I in early neonates was, therefore, also of interest.

To examine all of the above, we induced hypothyroidism in neonates with consequent GH deficiency, treated them with GH at d 2 and 7, and assayed their livers for abundance of activated Stat5 and expression of GHR, Spi 2.1 and 2.3, and IGF-I mRNA.

METHODS

Animals.

The present study was reviewed and approved by the University of Minnesota Institutional Animal Use and Care Committee. Timed pregnant Sprague-Dawley dams from Harlan Farms (Indianapolis, IN) were obtained on d 10 of gestation and given water containing methimazole (0.025%) from d 12 of gestation until their pups were killed. The development of the thyroid gland in the fetus, which usually begins at d 14, is prevented by this treatment (19). After birth, hypothyroid neonates do not secrete thyroid hormones T3 and T4. Because T3 is necessary for the production of GH, these neonates are rendered GH deficient. In our experiments, the day of delivery is considered d 1 of life. At either d 2 or 7, neonates were injected subcutaneously with either GH (0.2 μg/g body weight) or equal volumes of saline. One hour later, they were exsanguinated and their livers were harvested. The livers were either pooled directly for isolation of nuclei or frozen individually in liquid nitrogen for subsequent RNA extraction. This experiment was repeated with neonates from four similarly treated dams. Identical experiments were performed with normal age-matched neonates for preparation of hepatic nuclear extracts.

RNA extraction and Northern blots.

RNA extraction was performed by the guanidine-hydrochloride method as previously described (10). The RNA were separated by electrophoresis in a 1.5% agarose gel containing 2% formaldehyde. The running buffer contained 20 mM 3-(N- morpholino) propanesulfonic acid (MOPS), pH 7.0, 5 mM sodium acetate, 1 mM EDTA, and 2% formaldehyde. After electrophoresis, the integrity of the RNA samples was assessed by visualization of ribosomal RNA bands on the gel. The gel was electrotransferred to either ZetaBind membrane (CUNO Laboratory Products, Meriden, CT) or a positively charged nylon membrane (Boehringer Mannheim, Indianapolis, IN). After completion of the transfer, the extent of the transfer was assessed by UV visualization of both the membrane and the gel. Dot blots of RNA samples were prepared on BA-S supported nitrocellulose membrane (Schleicher and Schuell, Keene, NH) as described (20). All membranes were baked at 80°C for 2 h and stored in a vacuum desiccator for subsequent hybridizations.

Radiolabeled probes for Spi 2.3 cDNA and GHR cDNA were generated by the random primer method with the Rediprime random primer labeling kit (Amersham Life Science, Arlington Heights, IL) using 32[P]-dCTP as the labeled nucleotide. A riboprobe for IGF-I cRNA was generated as previously described using 32[P]-UTP (21). Hybridization conditions and washing for cDNA probes and riboprobes were similar to those previously described (21) except that 200 μg/mL salmon sperm DNA (ssDNA) was used in dot-blot hybridizations, and 670 μg/mL ssDNA in prehybridization and 100 μg/mL ssDNA were included in hybridization solutions for Northern blots.

Northern blots and dot blots were quantitated by phosphorscreen autoradiography (Molecular Dynamics Corp., Sunnyvale, CA). All quantitations of Spi 2.1 and 2.3 and GHR mRNA on Northern blots were normalized to those of glyceraldehyde-3-phosphate dehydrogenase mRNA. The data for each age group were expressed as a percentage of the value of a d-65 normal adult male rat and were shown as mean ± SEM. The significance of changes was determined using either t test or ANOVA at p < 0.05 significance. Post hoc testing in ANOVA was performed using Fisher least-mean differences at 95% significance. For quantitation of IGF-I dot blots, a series of samples containing increasing amounts of RNA were applied to the membrane, and linear regressions of the hybridization signals were performed and analyzed.

EMSA.

Hepatic nuclear extracts were prepared as previously reported (18). Eight micrograms of d-2 or -7 hepatic nuclear extracts were incubated with 32[P]-dCTP-radiolabeled GHRE from the Spi 2.1 gene, GHRE, (TCGACGCTTCTACTAATCCATGTTCTGAGAAATCATCCAGTCTGCCCAC). These reactions were carried out in a buffer containing 20 mM HEPES (pH 7.6), 10% glycerol, 2 mM MgCl2, 5 mM CaCl2, 0.1 mg/mL BSA, 4% Ficoll 400, 1 mM spermidine, 1 mM DTT, and 1 mM PMSF. Poly (dI-dC) (1.5 μg/6 μg protein) (Pharmacia Biotech, Inc., Piscataway, NJ) was also added. Reaction mixtures were incubated at 30°C for 30 min and loaded onto a nondenaturing 5% polyacrylamide gel (39.5:1 acrylamide:bis-acrylamide, 3.2% gycerol) in a Bio-Rad (Hercules, CA) Protean II system. Gels were electrophoresed in 0.5X TBE (45 mM Tris, 45 mM boric acid, and 1 mM EDTA, pH 8.3) at 150 V for 2.5 h and then dried and exposed to x-ray film.

Immunoprecipitations and Western blotting.

Two hundred micrograms of total hepatic nuclear proteins was incubated with an antibody to Stat5 (SC-835 C-terminus, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) at a 1:500 dilution in 1X RIPA containing 5% Nonidet P-40, 0.5% SDS, 750 mM NaCl, 250 mM Tris, 10 mM EDTA, 0.5% Na deoxycholate, 50 mM NaF, 1 mM PMSF, and 10 μg/mL aprotinin. After incubation at 4°C for 4 h, protein-A agarose (SC-2001, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was added and the final mixture was incubated for 16 h rotating end-on-end at 4°C. The agarose pellets collected after centrifuging were washed three times in 1X RIPA. The bound proteins were then released from the agarose pellets by boiling in SDS-sample loading buffer at 95°C for 3 min. After a brief centrifuge at 7500 ×g, the supernatants containing the proteins were resolved by SDS-PAGE (7.5% separating, 5% stacking) in a Bio-Rad Mini Protean II system and transferred electrophoretically to a BA-S supported nitrocellulose membrane. The membrane was then blocked in 5% milk in TBS-T (10 mM Tris pH 8.0, 150 mM NaCl, 0.05% Tween 20) for 16 h at 4°C and probed the following day with the same Stat5 antibody (1:500 dilution) for 2 h or with a Stat5a-specific antibody (Santa Cruz antibody #1081X). After washes, the membrane was incubated for 1 h with anti-rabbit IgG antibodies (Amersham Life Science, Arlington Heights, IL) conjugated to horseradish peroxidase (1:3000 dilution). Positive signals were detected with the ECL detection system (Amersham Life Science, Arlington Heights, IL) and quantitated using an imaging densitometer (Bio-Rad Model GS-700).

RESULTS

To determine whether Spi 2.1 and 2.3 mRNA were GH responsive in infant rats, Northern blots were prepared from total RNA extracted from the livers of d-2 and -7 hypothyroid rats that were or were not treated with GH. Figure 1 shows a representative blot that was probed concurrently with radiolabeled cDNA for Spi 2.3 and glyceraldehyde-3-phosphate dehydrogenase. Because the Spi 2.3 cDNA hybridizes to all three genes in the Spi 2 locus (Spi 2.1, 2.2, and 2.3), it can be used to measure the combined expression levels of Spi 2.1 and 2.3 mRNA. Spi 2.1 and 2.3 mRNA are both 1.8 kb in size and comigrate in a Northern gel. Spi 2.2 mRNA is 2.3 kb in length, not GH responsive, and migrates more slowly. Three different d-2 samples are shown for each treatment. Six different samples are shown for each d-7 treatment. The values are expressed as percentages of the adult value. There was no significant difference in quantities of Spi 2.1 and 2.3 mRNA between untreated hypothyroid (9 ± 2%) and GH-treated (4 ± 1%) neonates at d 2. At d 7, GH administration resulted in significant Spi 2.1 and 2.3 induction (19 ± 2%) when compared with untreated neonates (5 ± 1%).

Figure 1
figure 1

Upper panel, Northern blot of total hepatic RNA from d-2 and -7 hypothyroid neonates that had been probed with a Spi 2.3 cDNA. Twenty micrograms of total hepatic RNA were loaded in each lane. GH treatment is denoted as ±. After probing, the membrane was exposed to x-ray film for 48 h at room temperature. The bar on the right denotes the migration of the 18S rRNA. Major RNA species for Spi 2.1 and 2.3 are present at 1.8 kb, and Spi 2.2 is found at 2.3 kb. A denotes the adult control. Lower panel, Data are represented in graphic form as percent of the adult value. Each column denotes the mean value for each age group and treatment. SD are indicated by bars, and significant differences are indicated by an ★.

Full size image

To evaluate if the expression of GHR plays a role in the ontogeny of the GH responsiveness of Spi 2.1 and 2.3, a second Northern blot similar to the one shown in Figure 1 was prepared and probed with a full-length GHR cDNA that recognizes both GHR and GHBP mRNA. The results, expressed as percentages of the adult value, are shown in Figure 2. There was uniform low expression of the GHR and GHBP in both groups of d-2 neonates (21 ± 4%, 21 ± 2%). This level of expression increased significantly and similarly in both groups of d-7 neonates (45 ± 9%, 46 ± 10%). The mRNA abundance of neither GHR nor GHBP was significantly affected by GH status at d 2 or 7.

Figure 2
figure 2

Left panel, Northern blot of total hepatic RNA from d-2 and -7 hypothyroid neonates that had been probed with a cDNA for GHR that recognizes both GHR and GHBP. Twenty micrograms of total hepatic RNA were loaded in each lane. GH treatment was denoted as ±. After probing, the membrane was exposed to x-ray film for 12 h at −80°C. Bars on the right denote the migration of 18S and 28S rRNA. A denotes the adult control. Right panel, Data are represented in graphic form as percent of the adult value. Each column denotes the mean value for each age group and treatment. SD are indicated by bars, and significant differences are indicated by an ★.

Full size image

To understand the difference in the GH response of Spi 2.1 and 2.3 between d-2 and -7 hypothyroid neonates, we investigated the abundance and activation status of Stat5 in hepatic nuclear extracts of these neonates. An EMSA performed with these extracts and a radiolabeled GHRE are shown in Figure 3. The following values are expressed as percentages of binding of extracts from a normal treated adult. There was an increase in the formation of the Stat5/GHRE complex in extracts from both d-7 normal [(nl) + GH, 11.96%] and hypothyroid [(ht) + GH, 12.78%] neonates that were treated with GH. Untreated d-7 normal (1.72%) or hypothyroid neonates (no detectable binding) manifested little Stat5 binding. No Stat5 binding was seen in extracts from either normal or hypothyroid untreated d-2 neonates. GH treatment did not lead to an increase of Stat5 binding in these neonates.

Figure 3
figure 3

EMSA of hepatic nuclear proteins from d-2 and -7 normal and hypothyroid pups untreated (−) and treated (+) with GH. Binding reactions were carried out with 8 μg of neonate extracts or 5 μg of adult extracts (A) and the radiolabeled Spi 2.1 GHRE. Normal neonates are denoted as nl, and hypothyroid neonates are denoted as ht.

Full size image

To determine whether the difference in Stat5/GHRE binding seen in hepatic nuclear extracts from d-2 and -7 neonates is correlated with the abundance of Stat5 in these extracts, we carried out immunoprecipitations of these extracts with a Stat5 antibody and probed the subsequent blot with a Stat5 antibody (Fig. 4). The following values are expressed as a percent of the signal of purified GHINF. At d 2, GH administration to either normal (nl − GH, 32.4%; nl + GH, 24.5%) or hypothyroid (ht − GH, 29.8%; ht + GH, 31.8%) neonates did not result in an increase in Stat5 abundance. In d-7 neonates, however, there was a noticeable increase of Stat5 in hepatic extracts from both normal (− GH, 47.6%; + GH, 74.9%) and hypothyroid (− GH, 50.4%; + GH, 68.2%) neonates in response to GH. In addition, all d-7 samples showed a greater abundance of Stat5 protein. Similar results were obtained on repeated immunoprecipitations from different neonates, and an identical pattern of changes was seen using a Stat5a-specific antibody (not shown).

Figure 4
figure 4

Immunoblot of Stat5 immunoprecipitates of hepatic nuclear extracts probed with antibody to Stat5. GH treatment is denoted as ±. The arrow on the left denotes the location of Stat5 with an apparent molecular mass of ∼93 kD. Affinity purified GHINF was used as a positive control.

Full size image

We also evaluated the response of IGF-I, another hepatic gene regulated by GH in adult life, in our hypothyroid neonates. Total RNA from d-2 and -7 hypothyroid neonates and those treated with GH were applied to dot blots and probed with an IGF-I cRNA probe. The results, expressed as percentages of the adult value, are shown in Figure 5. IGF-I mRNA increased slightly from d 2 (− GH, 21.2 ± 2%; + GH, 24.1 ± 4%) to 7 (− GH, 55.1 ± 9%; + GH, 50.6 ± 7%), but the abundance on either day did not correlate with the GH state of the neonates. These data are representative of two separate experiments, each involving two dams.

Figure 5
figure 5

Quantitation of hybridizations of dot blots of total hepatic RNA probed with an IGF-I cRNA. Values (y axis) are expressed as percent of adult (d 65) value. Days (x axis) are d 2 and 7 of life. GH treatment is denoted as ±. Values are expressed as mean ± SE. NS denotes a statistically insignificant difference between treatments.

Full size image

DISCUSSION

Pregnant rat dams treated with methimazole show an 83% decrease in maternal plasma T3 levels and a 70% decrease of T3 in fetal levels by d 21 of gestation (19). A further decrease of T3 by d 2 after birth prevents much of the production of GH in treated neonates. If the normal neonatal rise in Spi 2.1 and 2.3 mRNA expression is due to the physiologic increase in GH, Spi 2.1 and 2.3 mRNA expression in hypothyroid neonates should be decreased. Spi 2.1 and 2.3 mRNA expression was only 9% of that observed in normal adult animals rather than 35% seen in age-matched normals. To determine whether this decrease was due to reduced GH in these neonates, we treated them with GH. Notably, this did not result in increased expression of Spi 2.1 and 2.3 mRNA at d 2 of life. A relative lack of GHR expression at d 2 is not the only factor limiting expression, because GHR mRNA levels are comparable in normal d-2 neonates (15) and their hypothyroid counterparts. This raises the intriguing possibility that the increased expression of Spi 2.1 and 2.3 in normal d-2 neonates is a response to an unidentified ligand. In addition to GH, the GHRE of the Spi 2.1 gene promoter can be activated by other ligands (22). The decrease in Spi 2.1 and 2.3 mRNA expression in normal neonates from d 2 to 7 then could be due to the decreasing levels of this unidentified ligand. Because growth in the fetus is primarily pituitary independent, growth in d-2 neonates may be similarly pituitary independent. In addition to the lack of Spi 2.1 and 2.3 mRNA response, we found that GH treatment in d-2 hypothyroid neonates did not result in an increased accumulation in tyrosine-phosphorylated Stat5. This indicates that the GH signaling pathway active in adult rats is not yet fully functional at d 2 of life. Alternatively, hypothyroidism may have secondary effects that limit the ability of d-2 neonates to respond to GH, leading to reduced Spi 2.1 and 2.3 mRNA expression at this age.

In contrast with d-2 hypothyroid neonates, we found that d-7 hypothyroid neonates responded to GH treatment as evidenced by the presence of tyrosine-phosphorylated Stat5 into the nucleus and its binding to the GHRE on EMSA, and increased Spi 2.1 and 2.3 mRNA expression. These events are coincident with increases in expression of GHR in these neonates. GHR and GHBP mRNA levels in the juvenile rat are GH independent and are not affected by GH or IGF-I treatment or hypophysectomy (23). Their expression is low in prenatal and infant livers (15, 24–26) and increases rapidly to adult levels with the onset of puberty (27–30). We have shown that at d 7, GHR expression in both normal and hypothyroid neonates was significantly higher than at d 2. This level, 45% of the adult value, is likely to be physiologically significant. In young rats, development of GH-dependent growth correlates with maturity of hepatic GHR (29). Transgenic mice carrying GH genes do not show accelerated growth until they are 3 wk old (4), also coinciding with the maturation of GHR. Between d 2 and 7, a critical threshold in the amount of GHR is established, resulting in effective transcription of Spi 2.1 and 2.3 mRNA using the GH/Stat5 signaling pathway. This observation suggests that GH is active by d 7 of life and that some of its actions are modulated by the amount of GHR expression. The correlation of Spi 2.1 and 2.3 mRNA expression with GHR expression continues as the infant rat matures (12).

Many somatogenic and anabolic effects of GH in neonates are mediated by IGF-I (31, 32). It has been suggested that the decreased levels of IGF-I that accompany neonatal hyposomatotrophism are the ultimate cause of growth retardation (1, 8). In addition to GH, thyroid hormone plays a role in both expression and release of IGF-I from liver and other tissues (33–35). Administration of GH to hypothyroid rats did not restore IGF-I levels, indicating that thyroid hormones exert an effect on IGF-I independently of GH (36, 37). In agreement with these observations, our dot-blot data show that IGF-I mRNA expression was not responsive to GH administration in either d-2 or -7 hypothyroid neonates. It is also likely that IGF-I mRNA regulation is thus different from that of Spi 2.1 and 2.3 mRNA. Moreover, maturation of the GH/Stat5 pathway by d 7 has no effect on IGF-I mRNA expression. To date, no Stat5 binding sites have been identified in the IGF-I promoter or intron regions. We did find, however, that IGF-I mRNA levels increased significantly from d 2 to 7 in hypothyroid neonates, in agreement with Hoyt et al. (38). This GH-independent increase supports the observation that GH immunoneutralization had no effect on plasma IGF-I levels in d-7 neonates (39). The dependence of IGF-I on pituitary factors in the normal rat is present by d 8 (9), but the onset of GH dependence before that day has not been demonstrated. Beginning at d 8, the postnatal increase in hepatic IGF-I mRNA is highly correlated with age-related increases in the expression of GHR mRNA (29, 30) and is, by inference, GH dependent.

Upon GH administration, both normal and hypothyroid d-7 neonates accumulated twice as much Stat5 protein in their nuclei as did d-2 neonates. This increase is coincident with the 2-fold increase in GHR expression in d-7 hypothyroid neonates when compared with that of d-2 neonates. There is also a corresponding increase in the amount of tyrosine-phosphorylated Stat5 in the d-7 normal and hypothyroid GH-treated nuclear extracts as evidenced by increased binding to the GHRE in EMSA. Because only phosphorylated Stat5 binds to the GHRE, its increased binding seen in the d-7 extracts is a strong indication that the GH/Stat5 signaling pathway is active. The lack of binding seen in the d-7 untreated normal samples is not uncommon, as the amount of Stat5 in the nucleus is dependent on the timing of normal GH pulses. The extracts from d-7 normal or hypothyroid neonates treated with GH manifested similar amounts of activated Stat5 and Stat5/GHRE binding. This finding indicates that GH alone is sufficient for the initiation of the signaling pathway that leads to the accumulation of activated Stat5 by d 7 of life. This accumulation is likely a direct reflection of increasing GHR expression. In d-2 neonates, low expression of GHR may prevent the accumulation of sufficient tyrosine-phosphorylated Stat5 in the nucleus to initiate Spi 2.1 and 2.3 transcription in response to GH.

Our work on the mechanism and kinetics of GH action in the regulation of Spi 2.1 and 2.3 expression indicates that GH action in d-2 hypothyroid neonates is limited and corroborates the observation by others that growth in early neonatal life is primarily pituitary independent. In contrast, GH responses in d-7 hypothyroid neonates are similar to those seen in adults. They include the accumulation of activated Stat5 in the nucleus with a subsequent increase in Spi 2.1 and 2.3 mRNA levels. These responses correlate well with increased GHR expression.