Main

Despite adequate food rehabilitation, failure of catch-up growth and low serum and liver IGF-I are observed during postnatal development in rat pups with IUGR caused by gestational protein malnutrition(1, 2) or by in utero ethanol exposure(3). Likewise, serum IGF-I remains low at 1 y of age in IUGR infants who did not undergo catch-up growth(4).

During the neonatal period in the rat, control of somatic growth switches from relative pituitary independence in the fetus to pituitary dependence in the adult(5–8). This pituitary regulation of somatic growth is mediated by GH, which stimulates serum IGF-I and promotes BW gain. Indeed, administration of IGF-I in normal neonatal rats causes increased BW, TL, and organ weight(9). Although somatic growth and serum levels of IGFs have been shown to be pituitary GH-dependent in both late fetal(10) and neonatal rats(6–8), GH injections in neonatal rats with growth retardation due to limited milk availability are not associated with any increase in growth parameters(9), thereby suggesting a relative resistance to the growth-promoting effects of exogenous GH.

The IGFBPs, which bind IGFs in serum and other extracellular fluids and modulate their biologic effects(11), are also regulated by both GH and nutrition in fetal and neonatal rats(6, 7, 10–13). However, we have previously reported that gestational protein malnutrition causes no obvious change in serum IGFBP concentrations, either at birth(1) or at later stages of postnatal development(2).

The present study was undertaken to determine and compare the effects of exogenous IGF-I and GH treatment on the postnatal growth of rat pups with IUGR due to gestational protein restriction, and placed after birth under conditions of adequate food rehabilitation. To this end, we have studied in these rats somatic growth (BW and TL), individual organ growth, serum IGFBPs, and liver IGF-I gene expression at weaning.

METHODS

Animals. Timed pregnant Wistar rats (vaginal plug = gestational d 0) were purchased from KUL laboratories (Leuven, Belgium). Upon arrival at d 0, the rats were randomly assigned to one of two dietary groups (U.A.R., Villemoisson-sur-Orge, France). One group was fed a normal protein powdered diet (20% casein; P20; control; n = 9), whereas a second group was fed a low protein diet (5% casein; P5; n = 8). These diets were isocaloric (325 cal/100 g) and contained similar amounts of fat (5% corn oil), cellulose (17%), and essential minerals and vitamins (8%); the difference in protein content being compensated by an increased amount of carbohydrates(corn starch and dextrose) in the P5 diet (65 versus 50% in the P20 diet). Both diets were given ad libitum throughout gestation. Animals were housed in individual metabolic cages under controlled conditions of lighting (0700-1900 h) and temperature (22 ± 2 °C) and had free access to tap water. Food intake and BW of the dams were recorded daily.

Experimental design. At birth (postnatal d 0), the number of pups, their sex, weight, and TL were noted. Litters of both dietary groups were immediately fostered to well nourished lactating dams having given birth 1 d before. To avoid competition within the litter between the suckling P20 and P5 pups, foster dams were randomly allocated to nurse littermates only from the same dietary group. On postnatal d 2, the size of each litter(n = 6/dietary group) was reduced to 6 pups per dam, to improve postnatal nutritional conditions(14, 15). Because there is no obvious gender-related difference in control or prenatally malnourished rats before they reach sexual maturity(2, 7), and because our preliminary work and analysis of the present study showed no significant sex-difference in the growth response to treatment with GH or IGF-I in P20 or P5 pups, male and female pups were considered together in our experiments. A 2 × 3 factorial design was used to analyze the effects of gestational protein restriction and hormonal therapy on postnatal growth. Within each P20 or P5 litter, pups (2 pups/litter) were randomized to receive twice daily (0900 and 1800 h) s.c. injections of recombinant bovine GH (2.5 μg/g BW/d), recombinant human IGF-I (1.8 μg/g BW/d) or an equivalent volume of saline from d 2 through postnatal d 23 (12 pups treated with each hormone or saline in each dietary group). Hormonal doses were selected according to previous substitution experiments in intact(9) and hypophysectomized(8) neonatal rats. The BW of each pup was recorded daily, whereas the TL (margin of the anus to tip of the tail) was determined weekly.

All rats were killed by decapitation 1 h after the last hormone injection. At the time of death, blood was collected from trunk vessels into glass tubes and centrifuged, and the serum was stored at -20 °C until assayed for hormonal measurements. Liver was quickly removed from each animal, weighed, and flash-frozen in liquid nitrogen for liver IGF-I mRNA studies. A number of other organs, including heart, kidneys, spleen, brain, and thymus were also removed, and their weight was determined. To compare the effects of gestational protein restriction and hormonal therapy on the growth of individual organs relative to overall body growth of the pups, the organ weight ratios [(organ weight/BW) × 100] were calculated as previously described(6–8). The protocol was approved by the Institutional Animal Care and Use Committee of the University of Louvain at Brussels.

Hormone preparation. Recombinant bovine GH (Monsanto, Europe SA, Brussels, Belgium; batch V206-001) used for treatment was dissolved in glycin-buffered saline (pH 9.6) and further diluted in sterile isotonic saline to stock solutions of 5 mg/mL. These solutions were stored at -20 °C until use. Recombinant human IGF-I (G080AB; 5 mg/mL) used for treatment was a generous gift of Genentech, Inc. (South San Francisco, CA) and was kept at 4°C until use. Pups were injected with GH or IGF-I solution freshly diluted with isotonic saline before use. Delivery volume was 10 μL/g BW (hormone concentrations, 125 and 90 μg/mL, respectively) in rats with BW < 25 g, and 5 μL/g (hormone concentrations, 250 and 180 μg/mL) in rats with BW≥ 25 g.

Serum IGF-I and IGFBP measurements. Serum IGF-I concentrations were measured by RIA using a nonequilibrium technique(16) after removal of IGFBPs by C18 cartridge (Waters Associates, Milford, MA) chromatography(17, 18). Using this technique, 68 ± 4% (mean ± SD, n = 10) of the IGF-I is recovered from neonatal rat serum, and more than 99% of the binding protein is removed(18). A parallelism between serial dilution curves of serum extracts and IGF-I standard was observed. The intra- and interassay coefficients of variation of this assay were both 6%. The human IGF-I antibody (rabbit polyclonal antibody; UBK 487) was a kind gift from Dr. L. E. Underwood (Chapel Hill, NC). Values are reported in nanograms/mL using recombinant human IGF-I (Amersham, Buckinghamshire, UK) as a standard.

To assess the abundance of serum IGFBPs, ligand blotting was performed using polyacrylamide gel electrophoresis as previously described(19). To reduce the number of samples processed, four pools of sera (three pups/pooled sample) were formed for each experimental group. Electrophoresed proteins were transferred to membranes (polyvinyl difluoride, Immobilon, Millipore), which were incubated for 24 h at 4 °C with 250 000 cpm/mL of a mixture of 125I-IGF-I and 125I-IGF-II(250 μCi/μg), washed, and subjected to autoradiography. The autoradiograms were scanned using an LKB Ultroscan XL laser densitometry. For each sample, density signals were determined for the 45-39-kD (glycosylated variants of IGFBP-3), 34-29-kD (IGFBP-1 and -2, and glycosylated IGFBP-4, -5 and -6), and 24-kD bands (nonglycosylated IGFBP-4). The 34-29-kD IGFBP cluster was analyzed together because they could not be reliably separated for quantification by scanning densitometry, and it was assumed to contain IGFBP-1 and -2 predominantly. The abundance of IGFBPs in a pool of sera from normally fed adult rats served as a control. The sizes of radioactive protein bands were estimated by comparison with standards from Life Technologies, Inc.(Gaithesburg, MD).

RNA preparation and Northern blot hybridization. Total RNA was isolated by the guanidine thiocyanate/cesium chloride method(20) from pooled liver tissue (three pups/pooled sample;n = 4 pools/experimental group). Total RNA (20 μg/lane) samples were size-fractionated, and transferred to nylon membranes (Hybond-N, Amersham, Buckinghamshire, UK) as described previously(1, 7). Blots were hybridized with a 194-bp exon 4-specific rat IGF-I riboprobe, using prehybridization, hybridization, and washing conditions reported previously(1, 7). Hybridization of blots with a chicken 32P-labeled β-actin c DNA, kindly provided by Dr. D. W. Cleveland, was subsequently performed to estimate the variation in RNA loading. Abundance of hybridized mRNA was quantified by scanning of autoradiograms using an LKB Ultroscan XL laser densitometry scanner (LKB, Bromma, Sweden). The results were corrected for β-actin mRNA and standardized by assigning a value of 100 arbitrary units to the mean value found in saline-injected control animals.

Statistical analysis. Data were analyzed by ANOVA using the Statistical Analysis System(21) to determine the respective influences of gestational diet and hormonal therapy as well as their possible interactions for the different variables studied. The differences in BW and TL of rats were analyzed by analysis of variance for repeated measures. The means of the different experimental groups of rats were compared by the Student-Newman-Keuls test. Data are shown as the mean ± SEM, and p < 0.05 was considered significant.

RESULTS

Although food consumption of protein-restricted dams during the first 2 wk of gestation was higher than that of controls and less during the last 5 d of gestation, mean daily food intake over the entire gestation was similar between the two groups (Table 1). However, the BW gain of protein-restricted dams at term was markedly impaired, being 51% less than that of the control dams (p < 0.001; Table 1). Gestational protein restriction caused significant growth retardation and early postnatal mortality in newborn pups without affecting litter size. Indeed, BW and TL of P5 pups at birth were reduced by 30%, compared with P20 pups (Table 1), and there was a 3-fold increase in the number of deaths occurring within 48 h after birth in the P5 pups compared with controls (24 versus 8%).

Table 1 Food intake and BW gain in dams fed ad libitum a normal protein diet (P20; controls) or a low protein diet (P5) throughout gestation, with litter size, birth weight, and TL
Full size table

Despite food rehabilitation during the suckling period, the BW and TL of the P5 pups on postnatal d 23 were still significantly reduced compared with those of age-matched P20 controls (-24 and -17%, respectively; p< 0.01; Fig. 1, top panels). The weights of liver, heart, kidney, brain, and thymus were also decreased (-13 to -35% changes;Table 2) as were serum IGF-I concentrations (-51%;p < 0.001; Table 4). The weight of the spleen was, however, similar to controls. When compared with saline-injected P20 controls, saline-injected P5 pups showed three patterns of organ sensitivities to prenatal malnutrition: a decreased organ weight ratio for the liver only (by 15%), an increased organ weight ratio for spleen and brain (by 32 and 16%), and a similar organ weight ratio for the other organs(Table 3).

Figure 1
figure 1

(Top panels) BW and TL in pups born from normal protein-fed dams (P20; controls, solid lines) or protein-restricted dams (P5; dotted lines), and injected twice daily from d 2 through d 23 of postnatal life with saline (filled squares), recombinant bovine GH (bGH) (open circles) or recombinant human IGF-I (filled triangles). Within the two dietary groups, values in GH- and IGF-I-injected rats are significantly different(p < 0.01) from those of the saline-injected animals from the day indicated by an arrow. (Bottom panels) Changes in BW and TL in the six groups of pups between postnatal d 2 and 23. Data are presented as the mean ± SEM of 12 rats/group. **, ***, p < 0.001, 0.001 vs saline (Sal)-injected P20 rats;▴▴▴, p < 0.001 vs saline-injected P5 rats.

Full size image
Table 2 Organ weight (g) in 23-d-old rats born from dams fed ad libitum a normal protein diet (P20; controls) or a low protein diet(P5) throughout gestation, and injected twice daily from postnatal d 2 through d 23 with saline, recombinant bovine GH (bGH), or recombinant IGF-I.
Full size table
Table 4 Serum IGF-I and IGFBPs in 23-d-old rats born from dams fed ad libitum a normal protein diet (P20; controls) or a low protein diet (P5) throughout gestation, and injected twice daily from postnatal d 2 through d 23 with saline, recombinant bovine GH (bGH), or recombinant IGF-I
Full size table
Table 3 Organ weight ratios in 23-d-old rats born from dams fed ad libitum a normal protein diet (P20; controls) or a low protein diet (P5) throughout gestation, and injected twice daily from postnatal d 2 through d 23 with saline, recombinant bovine GH (bGH), or recombinant IGF-I
Full size table

Administration of GH for 3 wk in P5 rats raised their serum IGF-I (1-fold) to levels observed in saline-injected P20 controls (Table 3), and restored normal BW and TL (94 and 98% of controls, respectively;Fig. 1, top panels), and organ weights (91-107% of those of saline-injected controls; Table 2). Administration of IGF-I for 3 wk in P5 rats increased serum IGF-I to levels that were, 1 h after the last injection, 6 times greater than those observed in saline-injected P5 rats, and 3 times greater than in saline-injected controls(Table 4). This treatment also resulted in normalization of BW and TL (94 and 96% of controls; Fig. 1, top panels), and organ weights (92-111% of controls; Table 2). The growth-promoting effects of GH and IGF-I became evident starting from the postnatal d 8 for BW and d 9 for TL. There was no temporal difference between GH and IGF-I in promoting body growth during the treatment period(p > 0.05; GH-treated versus IGF-I-treated rats;Fig. 1, top panels). The ability of GH- and IGF-I-injected P5 pups to experience catch-up growth was evidenced by the higher rate of body growth in these pups than in saline-injected P5 pups (Fig. 1, bottom panels). The organ weight ratios of most tissues were unaffected by the hormonal treatments, indicating that GH and IGF-I promoted body and organ growth to a similar extent. Weight ratio of kidneys was, however, higher in both GH- and IGF-I-treated P5 pups, but not in treated controls (Table 3).

In P20 control rats, GH and IGF-I injections also caused significant increases in serum IGF-I 1 h after injection (0.6- and 2-fold increases, respectively) compared with saline injections (Table 4). These GH- and IGF-I-injected controls also had increased BW (14 and 11%) and TL (12 and 11%) on postnatal d 23 (Fig. 1, top panels). These increases were significant by the 10th postnatal day for BW and by the 16th postnatal day for TL. GH injections were associated with a 10-27% increase in the weight of liver, heart, and brain, whereas IGF-I injections caused increases (10-30%) in the weight of all organs studied, except the thymus, in comparison with the saline-injected controls(Table 2). In both control and P5 rats, no significant(p > 0.05) interaction between the effects of gestational diet and response to GH or IGF-I treatment was noted on BW, TL, and organ weights on postnatal d 23. No further mortality was observed in any of the treatment groups beyond postnatal d 2, and hormone injections were not accompanied by any obvious morbidity.

There was no significant difference in total liver IGF-I mRNA levels between the saline-injected P5 and P20 rats at d 23 (Fig. 2). Although the abundance of hepatic IGF-I mRNA was reduced in rats treated with IGF-I, the difference was not significant. In contrast, there was an obvious 60-70% increase in the abundance of total liver IGF-I mRNA in GH-injected control and P5 rats when compared with saline- and IGF-I-injected animals (Fig. 2).

Figure 2
figure 2

Effects of GH and IGF-I administration on liver IGF-I mRNA concentrations in 23-d-old rats born from dams fed ad libitum a normal protein diet (P20; controls) or a low protein diet (P5) throughout gestation, and injected twice daily from postnatal d 2 through d 23 with saline, bovine GH (bGH) or IGF-I. (Top panel) Northern blot analysis of liver IGF-I mRNA. Each lane in the gel represents 20 μg of total liver RNA from pooled livers (3 rats/pool). (Bottom panel) Relative abundance of total liver IGF-I mRNA (sum of all bands for each lane). Autoradiogram of Northern blot shown in top panel was densitometrically scanned and expressed as arbitrary units corrected by the β-actin densitometric value. Values are shown as the mean ± SEM of three pooled samples (3 rats/pool) per group. ***, p < 0.001vs saline- and IGF-I-injected animals.

Full size image

Gestational protein restriction did not influence significantly the relative abundance of the various serum IGFBP species in the neonatal rats(p > 0.05; saline-injected P5 versus saline-injected P20 rats; Fig. 3 and Table 4). Also, serum IGFBP-1 to -4 concentrations in 23-d-old rats treated with GH were not significantly different from those treated with saline alone, although serum IGFBP-3 levels were reduced by 30% in GH-injected P20 and P5 rats. In contrast, exogenous IGF-I significantly increased serum IGFBP-3 on d 23 compared with both GH- and saline-treated groups, without affecting serum IGFBP-1, -2, and -4.

Figure 3
figure 3

Autoradiogram of Western ligand blot of serum IGFBPs in 23-d-old rats born from dams fed ad libitum a normal protein diet(P20; controls) or a low protein diet (P5) throughout gestation, and injected twice daily from postnatal d 2 through d 23 with saline, recombinant bovine GH(bGH) or recombinant IGF-I. Each lane represents a pooled sample of sera collected from three individual animals in each experimental group. A pool of sera from normal-fed adult rats was used for comparison. The Mr of the IGFBPs is indicated on the right.

Full size image

DISCUSSION

Our study shows that, under conditions of adequate nutrition, both GH and IGF-I induce postnatal catch-up growth in neonatal rats with IUGR caused by maternal protein malnutrition. To our knowledge, this is the first in vivo demonstration of the growth-promoting effects of GH and IGF-I in neonatal rats that are prenatally growth-retarded. In addition, our results confirm that GH and IGF-I have growth-promoting properties in normal neonatal rats(9, 22).

A reduced ability of the P5 pups to suckle could be potentially responsible for their decreased serum IGF-I concentrations and lack of catch-up growth at weaning, although reduction of litter size has been proposed as a method to increase postnatal nutrition(14, 15). Using this method, the growth rates in P5 and control pups are approximately equal(2, 15), suggesting that the P5 pups receive an equivalent amount of milk, in proportion to their body size. Also supporting this, are our observations that liver IGF-I mRNA levels and serum insulin are not reduced in P5 animals at weaning compared with controls(2), as would be expected if they were malnourished(23).

Earlier reports have suggested that in most mammalian species including man and rat, fetal and early postnatal growth is not under significant pituitary GH control(5), despite its very high concentrations in the circulation at this age(24, 25). Studies have also failed to demonstrate any effect of exogenous GH on neonatal growth in normal rats(9). However, some pituitary dependence of growth at early stages of life has been evidenced recently in neonatal rat models of hypophysectomy(5–8) and genetically defective dwarfism(10). More recently, support for the involvement of GH in neonatal growth has been obtained from studies showing that passive immunization against rat GH(26, 27) or GH releasing hormone(22, 28) during the neonatal period permanently alters the growth rate and several indices of the somatotropic function. GH replacement therapy restored to normal the defective growth rate in the GH releasing hormone-deprived neonatal rats, whereas it also increased BW gain in normal neonatal rats(22). Our results showing that GH injections increase body growth in normal neonatal rats and promote growth in prenatally growth-retarded neonatal rats reinforce the view that somatic growth may be GH-dependent very early in the postnatal period.

In hypophysectomized neonatal rats, GH effects are greater on BW than on TL, whereas thyroxine alone stimulates TL without increasing BW. In contrast, combined GH and thyroxine treatment in these animals restored both growth parameters to normal values(5–7). In the present study, we show that GH treatment equally stimulated BW and TL in our neonatal rats with intact pituitary. Taken together, these findings suggest that inappropriate substitution with other pituitary hormones may contribute to differences in the GH responsiveness of somatic growth in hypophysectomized neonatal rats.

Growth induction by GH in young rats is possible with or without a concomitant rise in circulating IGF-I(29), and studies(8, 26) have suggested that circulating IGF-I is not essential for growth in the neonatal rat. Because GH may stimulate growth and tissue IGF-I levels without a corresponding increase in circulating IGF-I(29), and because IGF-I is synthesized at multiple sites(30, 31), an autocrine or paracrine action for this growth factor appears to mediate the early phase of GH action in these animals. However, in our study, GH-induced growth and liver IGF-I mRNA was accompanied by an increase in serum IGF-I concentrations. We also observed increased body and organ growth of neonatal rats in response to exogenous IGF-I. These results suggest that, in addition to its well accepted autocrine/paracrine function(29, 31, 32), IGF-I is capable of expressing an endocrine effect in the mediation of neonatal growth.

In the present study, GH and IGF-I promotion of organ growth in control pups was commensurate with BW, a finding in agreement with data obtained in normal neonatal rats treated with IGF-I(9). Our study also shows similar organ responses to the growth-promoting effects of GH and IGF-I, as evidenced by similar organ weight ratios between GH- and IGF-I-treated pups in each tissue studied. However, a disproportionately greater growth of some organs (i.e. spleen, kidney) after IGF-I treatment compared with a relatively proportional GH-induced organ growth has been observed by others in GH-deficient neonatal(8) and mature(33) rats, suggesting that IGF-I may have a differential effect on organ growth. Using a model of neonatal rat malnutrition induced by maternal food restriction during lactation, Zhao et al.(34) also reported similar results in intact neonatal rats after short-term (4 d) nutritional repletion. However, in agreement with our findings, a more prolonged period of nutritional repletion and hormonal therapy (24 d) did not affect the organ weight ratios of most tissues(35). This might suggest a different time course in organ growth response to IGF-I treatment and nutritional rehabilitation in neonatal rats.

Serum IGF-I concentrations 1 h after hormone injection were more elevated in the IGF-I-injected rats than in those injected with GH. In contrast, in rats given two daily s.c. injections of GH or IGF-I for the first 2 wk of postnatal life, Philipps et al.(9) did not observe any difference in serum IGF-I concentrations between groups 16 h after the last injection, using similar weight-adjusted hormonal doses as those used in the present study. Kinetics of serum IGF-I concentrations after s.c. hormone injections were not addressed in our study. It is, however, likely that serum IGF-I reached a peak value 1 h after injection in the IGF-I-injected rats, but not in the GH-treated animals. Because of the low IGF-I half-life(36), the mean 24-h IGF-I concentrations could in fact be similar between both GH- and IGF-I treated groups.

Modulation of the serum IGFBP profile resulting from hormonal or nutritional influences could be a key regulator of IGF-I actions(11, 37). Despite the significantly improved growth of the P5 pups treated with GH compared with saline alone, the patterns of serum IGFBPs in these two groups were nearly identical. We also found similar growth-promoting effects between GH and IGF-I, whereas only the pups receiving IGF-I showed significant increased expression of serum IGFBP-3 without alteration in serum IGFBP-1 to -4. Thus, our data fail to support a major role of serum IGFBPs in the regulation of somatic growth in prenatally growth-retarded neonatal rats. However, a fine study of the kinetics of IGFBPs after s.c. IGF-I or GH injection would have been required to characterize rapid serum changes, in particular of IGFBP-1 and -2, in our neonates in which the serum acid-labile subunit necessary for formation of the 150-kD complex(38, 39) is low(40, 41). The lack of GH effect on IGFBP-3 in our study is consistent with studies in GH-deficient rats(8, 38) and during recovery from neonatal malnutrition(34, 35). Because only the treatment groups with the highest serum IGF-I levels demonstrated increased serum IGFBP-3 in our study, and because our GH-treated animals did not exhibit the same degree of IGF-I elevation, it remains possible that higher doses of GH would have induced serum IGFBP-3.

It has been reported that treatment with IGF-I in both neonatal(8) and mature(39) hypoxic rats was not as growth-promoting as GH itself. Differences between IGF-I and GH effects might be explained by a reduction in these animals of serum acid-labile subunit. In our study, although we show that IGF-I injections increase serum IGFBP-3, it remains to be determined whether serum ALS is also modified in our model. The restoration of the 150-kD complex by IGF-I infusion in diabetic(42) but not in hypophysectomized rats(38, 42) suggests that GH is necessary to accomplish formation of this complex. We anticipate that GH secretion in our P5 pups, as in diabetic rats, persists sufficiently for IGF-I treatment to induce formation of this large binding protein complex. This would explain why GH and IGF-I treatments are equally effective in stimulating catch-up growth in our study.

In conclusion, our results show that low serum IGF-I concentrations in neonatal rats with IUGR are responsible, in part, for the failure of postnatal catch-up growth, and that both adequate nutritional support and normal GH-IGF-I axis are essential for optimal postnatal growth. These effects are likely mediated via endocrine and autocrine/paracrine actions of IGF-I.