Main

HI during the neonatal period can lead to severe brain damage, resulting in cerebral palsy and mental retardation(1). Over the past decade, HI has attracted renewed attention as our understanding of the neurochemical cascade of these lesions has increased. However, the precise molecular mechanisms behind HI brain damage are not yet fully understood.

A number of genes coding for growth factors and their associated binding proteins and receptors reportedly are induced in the area of brain damage after hypoxic ischemia(2–7). Several of these growth factors have also been implicated in the process of neuronal rescue after brain damage, including IGF-I, basic fibroblast growth factor (bFGF), and NGF(8–16).

IGF-I is a peptide growth factor synthesized in most tissues and regulated primarily by GH and nutritional status(17). IGF-I binds mainly to the IGF-I receptor (type I), although its bioavailability is also influenced by six binding proteins (IGFBP1-6)(18). IGF-I participates in the regulation of function and growth of multiple tissues, including the brain during development(19). In 21-d-old rats, IGF-I mRNA and protein, IGFBP2, IGFBP3, and IGFBP5 mRNA were found to be up-regulated after HI, with a peak 3-5 d after the insult, whereas expression of IGFBP4 mRNA decreased(20). A somewhat different pattern was found in 7-d-old rats, there being an initial general suppression of IGF-I, IGF-I receptor, IGFBP2, and IGFBP5 mRNA, followed by an increased abundance of IGF-I and IGFBP5 mRNA in astrocytes after 3-5 d of recovery(21). Treatment with IGF-I administered into cerebral ventricles beginning 2 h after HI has been shown to reduce brain injury in adult rat models of ischemia(14) and in fetal lambs subjected to HI(16).

It is still controversial regarding the extent to which GH crosses the blood brain barrier. However, the mRNA for GH receptor and binding protein are expressed throughout the rat CNS in both neuronal, glial, and endothelial cells(22). The expression is even more pronounced in certain areas of the CNS that co-express IGF-I and IGF-I receptor mRNA and protein(23). It has been reported that GH receptors are expressed in the choroid plexus, the hippocampus, the hypothalamus, and the pituitary gland of humans(24). Moreover, when GH-deficient adults were treated with GH for 1 month, there was a 10-fold increase in GH CSF levels compared with baseline, suggesting that GH does indeed pass the blood CSF barrier(25). There were also increments in the CSF levels of IGF-I and IGFBP-3.

Aims of the present study were to explore further the regulation of the components of the GH-IGF-I axis after HI in neonatal rats and to investigate the possibility that GH has a neuroprotective effect.

MATERIALS AND METHODS

Animals. In a neonatal model that has been described previously(26,27), 7-d-old inbred Wistar Fue rats were anesthetized with halothane (2% for induction and 1% for maintenance) in a mixture of nitrous oxide and oxygen, by means of a snout mask. The left common carotid artery was cut between two prolene sutures (9-0). After anesthesia and surgery, the animals were allowed to recover for 60 min. They were then exposed to 70 min of hypoxia in a humidified chamber at 36°C with 7.7 ± 0.001% oxygen in nitrogen. The pups were returned to and kept with their dams until they were killed, as described below. Animal studies were approved by the animal ethics committee in Göteborg.

Study protocol I. In the first series of experiments, 28 rats were subjected to HI, as previously described. Another 28 rats were left intact and served as controls. The animals were killed at 3 h, 24 h, 72 h, and 14 d after HI. After decapitation, the brains were removed from the skull, separated into two hemispheres, immediately frozen in isopentane on dry ice, and later analyzed with a solution hybridization assay with RNA probes for IGF-1, IGF-1R, and GHR.

Study protocol II. In the second series of experiment, rats were subjected to HI; immediately after the insult, half of the litter received daily s.c. injections of rhGH (Genotropin, which was a generous gift from Pharmacia & Upjohn, Stockholm, Sweden) in three different doses: 5 mg/kg body weight for 14 d (n = 22), 50 mg/kg for 7 d (n = 23), and 100 mg/kg for 7 d (n = 34). The other half of the litter served as controls and received daily s.c. injections with saline (i.e., separate litters and controls for each group). The pups were kept with their dams until killed at 21 d of age for evaluation of the brain damage. From each litter, two animals-one GH-treated and one control-were killed after 72 h and were processed for solution hybridization. The rectal temperature, which has been shown to correspond to the brain core temperature(28), was measured in two litters at 1, 3, 6, 12, 24, 36, 48, and 72 h after HI.

Evaluation of brain damage. The brain damage was evaluated by weighing the hemispheres 14 d after HI. Brain damage was expressed as the weight deficit of the damaged hemisphere as percentage of the contralateral hemisphere. Previously, it was shown that there is a close correlation between weight and histopathology with regard to evaluation of brain damage in this model(29,30). Both wet and dry weights were registered, and the water component of the brain was calculated.

Probes. Antisense GH receptor 35S-UTP-labeled RNA was synthesized according to the instructions provided by the manufacturer (Boehringer Mannheim, Mannheim, Germany; Promega, Madison, WI), with T3 RNA polymerase (Pharmacia, Uppsala, Sweden); EcoRI linearized plasmid pT7T3 18U was used as template(31). Synthetic standard was generated, using HindII linearized plasmid and T7 RNA polymerase. The pT7T3 18U plasmid contains a 560-bp BamHI fragment of the rat GH receptor cDNA that encodes a part of the extracellular domain of the GH receptor. The probe allowed the detection of both the GH receptor and GH-binding protein. A 153-bp SmaI fragment of a genomic subclone of mouse IGF-I (exon 3 by analogy to human IGF-I) subcloned into a pSP64 plasmid was used as a template for probe synthesis(32). The plasmid was linearized with EcoRI and used as a template for the synthesis of 35S-UTP-labeled IGF-I cRNA antisense probe with SP6 RNA polymerase, according to the instructions of the manufacturer (Promega, Madison, WI). Synthetic standard was generated by using EcoRI linearized plasmid and SP6 RNA polymerase. The IGF-I-R RNA probe was synthetized from a 265-bp fragment of the rat IGF-I-R cDNA subcloned into a plasmid vector, pGEM-3(33). The vector was linearized with EcoRI, and the rat IGF-I-R antisense RNA was synthesized by use of a SP6 RNA polymerase and 35S-UTP. Synthesis of rat IGF-I-R sense RNA standard was achieved by using BamHI for linearization of the plasmid and by the addition of T7 RNA polymerase.

RNase protection assay. The Ribonuclease Protection Assay Kit (RPAII-kit, Ambion Intermedica) was used. Briefly, total RNA (40 µg) from brain was hybridized overnight at 45°C with 35S-α-UTP (Amersham, UK)-labeled r-IGF-I, r-IGF-IR, and rGH-R probes. RNase digestion was at 37°C for 30 min. As control, the probes were hybridized with 25-µg yeast tRNA or were only digested. 32P-labeled HAE III DNA (Promega) was used as a molecular size marker.

The Rnase-protected fragments were separated by electrophoresis through a 6% polyacrylamide gel. After the gels were dried, they were exposed on a PhosphoImager screen. The screens were developed on a Phosphor Imager (Molecular Dynamics Inc., Sunnyvale, CA), and densitometric analyses were performed.

Solution hybridization. Frozen tissue was homogenized with a Polytron in 1% SDS, 20 mM Tris-HCl (pH 7.5), and 4 mM EDTA. The homogenate was then treated with proteinase-K, and TNA was extracted with phenol-chloroform(31). A solution hybridized assay was used to quantify mRNA for IGF-I, IGF-I-R, and GH-R(34). TNA samples were assayed at 70°C for 24 h in 0.06 M NaCl, 20 mM Tris-HCl (pH 7.5), 4 mM EDTA, 0.1% SDS, 10 mM DTT, 25% formamid, and a 35S-labeled IGF-I RNA probe, IGF-I-R RNA probe, or GH-R RNA probe. After the addition of 100 µg herring sperm DNA, the samples were treated with 40 µg/mL RNase A and 2 µg/mL RNase T1 (Sigma Chemical Co., St. Louis, MO). Trichloroacetic acid-precipitated protected hybrids were then collected on glass-fiber filters (GF/C Whatman, Whatman International Ltd, Maidstone, UK) and were counted in a scintillation counter. The signal was then compared with a standard curve based on known amounts of IGF-I mRNA, IGF-I-R mRNA, or GH-R mRNA. The results were related to the DNA content(35) in the TNA sample.

Statistical analyses. Values are given as the mean ± SEM. Differences in levels of mRNA for IGF-I, IGF-IR, and GHR were compared by nonparametric Mann-Whitney test. The differences in weight reduction (protective effect) between GH-treated and control rats was evaluated by comparing animals from the same litter, and the results obtained from all litters were pooled by use of Mantel's test(36).

RESULTS

Study protocol I. There was a significant increase in IGF-I mRNA in the damaged left hemisphere 72 h and 14 d after HI, compared with the intact left hemisphere of control rats (p < 0.05) (Fig. 1A). Moreover, there was an increased level of IGF-I mRNA in the contralateral hemisphere 2 weeks after HI (p < 0.05) (Fig. 1B).

Figure 1
figure 1

Quantification of IGF-I mRNA by solution hybridization assay, comparing IGF-I mRNA in the left (A) and right (B) hemispheres after HI with that of intact controls. Levels of IGF-I mRNA are expressed as amol/µg DNA. Values are mean ± SEM. *p < 0.05 for HI vs controls.

Full size image

There were no significant changes in GH-R mRNA expression in the damaged left hemisphere throughout the study period (Fig. 2A). However, the levels of GH-R mRNA were significantly elevated in the contralateral hemisphere 2 weeks after HI (p < 0.01) (Fig. 2B).

Figure 2
figure 2

Quantification of GH-R mRNA by solution hybridization assay, comparing GH-R mRNA in the left (A) and right (B) hemispheres after HI with that of intact controls. Levels of GH-R mRNA are expressed as amol/µg DNA. Values are mean ± SEM. **p < 0.001 for HI vs controls.

Full size image

There were no significant changes in IGF-I-R mRNA expression in either the left or the right hemisphere throughout the study period (data not shown).

Study protocol II. No apparent protective effect, with regard to brain damage, was obtained with a daily dose of 5 mg/kg GH (Fig. 3). The mean extent of brain injury was somewhat lower in the pups that received 50 mg/kg or 100 mg/kg of GH compared with controls, but the differences did not reach statistical significance (p = 0.0556 and p = 0.22, respectively). However, if we combined the results of the 50 and 100 mg/kg groups, we found that brain damage was significantly reduced (p < 0.05, Mantels test). We found no differences between the two groups with regard to body temperature (Fig. 4), body weight, or brain water content of the brains (data not shown). Moreover, there were no significant differences between male and female animals in this study.

Figure 3
figure 3

Neuroprotective effects of GH shown as weight reduction of damaged hemispheres after HI, comparing GH-treated and saline-treated animals. *p < 0.05 when combining groups treated with doses of 50 and 100 mg/kg were compared with those treated with saline. Values are means ± SEM.

Full size image
Figure 4
figure 4

Rectal temperature of GH (n = 7) and saline (n = 7)-treated rats at 1, 3, 6, 16, 28, 42, and 72 h after HI. There were no differences between the rectal temperatures in the two groups. Values are means ± SEM.

Full size image

There was a significantly higher overall expression of IGF-I mRNA in both hemispheres of rats treated with 50 mg/kg of GH, compared with the rodent controls. This was also the case when the right hemispheres of GH and control rats were compared (Fig. 5). However, there were no significant differences in IGF-I mRNA levels between the left hemispheres of treated and control rats Also, there was significant upregulation of GH-R after hGH treatment in both the damaged left hemisphere (p < 0.05) and the contralateral right hemisphere (p < 0.05) of treated rats compared with controls (Fig. 6).

Figure 5
figure 5

Quantification of brain IGF-I mRNA by solution hybridization assay, expressed as IGF-I mRNA amol/µg DNA in GH- and saline-treated rats after HI. *p < 0.05 GH vs saline treatment and **p < 0.01 GH vs saline treatment. Values are means ± SEM.

Full size image
Figure 6
figure 6

Quantification of brain GH-R mRNA by solution hybridization assay, expressed as GH-R mRNA amol/µg DNA in GH- and saline-treated rats after HI. There was significantly more GH-R mRNA in GH-treated rats than in the controls, in comparisons of both right-right and left-left hemispheres (*p < 0.05, ***p < 0.001). Values are means ± SEM.

Full size image

IGF-I mRNA expression was significantly elevated in total brain tissue (p < 0.01) and in the contralateral hemisphere of GH-treated rats, compared with controls (p < 0.05) (Fig. 7). However, in GH-treated and control rats, there was no significant difference in IGF-I-R mRNA expression in the left hemisphere.

Figure 7
figure 7

Quantification of brain IGF-I-R mRNA by solution hybridization assay, expressed as IGF-I-R mRNA amol/µg DNA in both GH- and saline-treated rats after HI. There was significantly more IGF-I-R mRNA in the GH-treated rats than in the controls, as determined by our comparison of both the total brain tissue and the right-right hemisphere (*p < 0.05, **p < 0.01). Values are means ± SEM.

Full size image

RNase protection. RNase-protected fragments were analyzed on denaturing polyacrylamide gels. For IGF-I (Fig. 8A), the protected fragment was approximately 150 bases long when a probe was hybridized to RNA from rat brain (lanes 6-8) corresponding to the 153-base insert. An additional band of 170 bases was also detected in lanes 1-8.

Figure 8
figure 8

RNase protection assay. Rat brain RNA was hybridized with 35S-UTP-labeled mouse IGF-I (A), rat GH-R (B), and rat IGF-I-R (C) probes. 32P-labeled HAE III DNA was used as a molecular size marker. (A) lane 1, undigested mouse IGF-I probe; lane 2, RNase digested probe; lane 3, RNase digested probes and tRNA; lanes 4 and 5, probe hybridized to sense RNA, synthesized in vitro; lanes 6-8, probe hybridized to rat brain RNA; lane 9 (left empty); and lane 10, end-labeled ØX174/HaeIII DNA marker (Promega) used as a molecular size marker. (B) lane 1, undigested rat GH-R probe; lane 2, RNase digested probe; lane 3, RNase digested probe and tRNA; lanes 4 and 5, probe hybridized to sense RNA, synthesized in vitro; lanes 6-8, probe hybridized to rat brain RNA; lane 9, end-labeled ØX174/HaeIII DNA marker (Promega) used as a molecular size marker. (C) lane 1, undigested rat IGF-I-R probe; lane 2, RNase digested probe; lanes 3 and 5, probe hybridized to sense RNA synthesized in vitro; lane 4, RNase digested probe and tRNA; lanes 6-8, probe hybridized to rat brain RNA; and lane 9, end-labeled ØX174/HaeIII DNA marker (Promega) used as a molecular size marker.

Full size image

For the GH-R (Fig. 8B), the undigested probe was found to be 560 bases long (lane 1), corresponding to the insert and protected when hybridized to RNA from rat brain (lanes 6-8).

The IGF-I-R-protected band (Fig. 8C) was 265 bases long when the probe was hybridized to rat brain RNA.

DISCUSSION

To our knowledge, the role of growth hormone in cerebral hypoxia ischemia has not been addressed previously. The main findings of the present study were that IGF-I mRNA expression was up-regulated 72 h and 14 d after injury in the damaged hemisphere, and that mRNA for IGF-I and GH receptor was significantly increased in the contralateral hemisphere 14 d after injury. There were no significant changes in IGF-I receptor mRNA expression throughout the study period. Finally, we observed that high doses of GH (50 and 100 mg/kg/d, s.c.) provided a moderate degree of neuroprotection.

The pattern of IGF-I mRNA expression in the present study correlates well with previous data reported by other groups, with an increase already apparent after 72 h that was maintained until 10 d after HI(20,21). However, increased levels of IGF-I mRNA expression after 14 d have not previously been demonstrated. The significance of the additional finding that IGF-I mRNA was also up-regulated in the contralateral hemisphere is unclear. However, it could be speculated that the induction of IGF-I gene expression is of importance for adaptation of the contralateral hemisphere to hypoxia.

GH receptor and binding protein expression reportedly are distributed widely in the CNS(22). However, so far, there have been no reports of studies on the regulation of GH receptor expression in the brain. In other tissues, there are indications that local demand may regulate regional expression of the GH receptor mRNA. In the heart, there is up-regulation of the GH receptor during a hemodynamic load and development of cardiac hypertrophy(37), occurring, also, in skeletal muscle after ischemic injury(38). Thus, in the brain, an increased expression of GH receptor mRNA expression may constitute a link in the activation of the GH-IGF-I axis during adaptation after HI.

IGF-I receptor mRNA expression pattern after HI has not previously been reported, although the binding capacity of I125-IGF-I has been shown to increase in an adult ischemic model(39). However, whether this was due to increased binding capacity or, rather, to an increased number of IGF-I receptors, was not revealed. In our study, we could not detect any significant changes in IGF-I receptor mRNA expression throughout the study period.

Previous studies have shown that IGF-I provides a considerable degree of neuroprotection in fetal lambs and 21-d-old rats(9,16),) whereas IGF-I was ineffective in an adult rat model of global ischemia(39). Furthermore, IGF-I reduces both infarction size and apoptosis in a model of myocardial infarction(40).

It has been demonstrated that both GH and IGF-I interact with neural tissue, and both agents reportedly stimulate regeneration of the rat sciatic nerve(41,42). Mechanisms of GH action might be the result of local induction of IGF-I. It has been shown that IGF-I decreases in both serum and CNS after hypophysectomy and that the levels can be restored after the administration of GH(43). GH injections into cerebral ventricles reportedly also increase the expression of IGF-I mRNA in the brain of adult rats(44). However, the possibility cannot be excluded that GH exerts direct effects in the CNS, independent of IGF-I.

Our finding that the protective effects of GH were obtained only with pharmacologic rather than physiologic doses might be due to the incomplete passage of GH over the BBB. The vascular permeability for GH across the BBB is unknown. although it has been shown that, in humans, CSF-GH increases after the S.C. administration of GH(25). A focal spinal cord injury, especially in younger rats, increased the vascular permeability for GH into CSF(45). Therefore, it seems reasonable to assume that GH crosses the BBB to some extent in our model of HI in 7-d-old rats.

In conclusion, HI induces the expression of IGF-I mRNA, and HI also evoked an up-regulation of GH-R mRNA 14 d after the insult. GH treatment enhanced the expression of IGF-I mRNA, GH receptor mRNA, and IGF-I receptor mRNA 72 h after the insult, and it also provided a moderate degree of cerebral protection in the neonatal rat brain.