Main

Maternal protein deficiency throughout gestation causes fetal growth retardation that cannot be corrected by adequate postnatal nutrition(1). At birth, rat pups from protein-malnourished dams exhibit reduced body and organ weight and decreased cell number in liver, kidney, heart, brain, and thymus(2, 3). These changes persist until weaning or later(2, 4). The cellular and/or hormonal defects involved in the failure of catch-up growth after gestational protein malnutrition are unknown. IGF-I is a mitogenic peptide that is thought to play an important role in the control of fetal and early postnatal growth and development(3, 5–7). Besides its regulation by GH, IGF-I expression is also controlled by nutritional status. Serum IGF-I concentrations and hepatic IGF-I mRNA abundance are decreased by fasting(8) and protein(9, 10) or energy restriction(11). Furthermore, the IGFBP, which bind IGF in serum and other extracellular fluids and modulate the biologic effects of IGF(12), are also nutritionally regulated(13–15).

The goal of this study was to examine the long-term effects of dietary protein restriction during gestation on serum IGF-I, serum IGFBP, and liver IGF-I gene expression during postnatal development in the rat and to relate these results to growth indices known to be affected by prenatal protein malnutrition.

METHODS

Animals and experimental design. 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 (n = 6 rats per group). One group (P20) was fed a normal protein powdered diet (20% casein; control), whereas a second group(P5) was fed a low protein diet (5% casein). Both diets were given ad libitum throughout gestation. These diets were isocaloric (325 cal/100 g) and contained similar amounts of fat (5%), cellulose (17%), and essential minerals and vitamins (8%; U.A.R., Villemoisson-sur-Orge, France). Animals were housed in individual metabolic cages under controlled conditions of lighting (0600-1800 h) and temperature (22 ± 2°C) and had free access to tap water. Food intake and body weight of the dams were recorded daily. At birth (postnatal d 0), the number of pups and their sex and weight were noted. Litters of each dietary group were immediately fostered to well nourished lactating dams having given birth 1 d before. On postnatal d 2, the size of each litter was reduced to six pups, to improve postnatal nutrition. When possible, the litters (n = 5/diet group) contained equal numbers of males and females. P20 and P5 pups were weaned at 22 d of age, and each animal was placed in a metabolic cage and fed a normal protein dietad libitum. The weight of each pup was recorded daily from birth until d 63, as was the food intake in the postweaning period. The tail length(margin of the anus to tip of the tail) was determined only at the time of killing.

On d 8, pups from one P20 litter and one P5 litter were killed by decapitation between 0900 and 1100 h. The body weight of these pups fell within 1 SD of the mean weight of all pups belonging to the same dietary group. At weaning (d 22), two pups per litter were chosen randomly and killed. The remaining rats were killed at 63 d of age.

At the time of death, blood was collected from trunk vessels, and the serum was stored at -20°C until assayed for IGF-I, IGFBP, GH, and insulin. The liver, heart, brain, spleen, and kidneys were removed and weighed. One liver fragment was flash-frozen in liquid nitrogen and stored at -70°C for determination of IGF-I content and mRNA studies. The remaining portion of the liver was homogenized in 10% (wt/vol) ice-cold 0.3 M sucrose as previously described(16) and stored at -20°C until assayed for bovine GH binding. For the 8- and 22-d-old pups, the data from males and females were combined, because no significant sex difference is present at these ages(17, 18). The protocol was approved by the Institutional Animal Care and Use Commitee of the University of Louvain at Brussels.

RIAs and binding studies. Serum immunoreactive IGF-I was measured in a nonequilibrium RIA(16) after removal of IGFBP by C18 cartridge (Waters Associates, Milford, MA) chromatography(19, 20). Using this technique, 68% of the IGF-I in rat serum was recovered, and more than 99% of the binding proteins was removed. The intra- and interassay coefficients of variation of this assay were both 6%(20). Values are reported in nanograms/mL using recombinant human IGF-I as a standard (Amersham Corp., Buckinghamshire, UK). Serum rat GH was measured in RIA(15) using the reagents provided by the NIDDK Rat Pitutary Hormone Distribution Program, and values are expressed in terms of the National Institutes of Health RP-2 reference preparation. The RIA for serum insulin, using rat insulin as a standard, was performed as described previously(16, 19).

Somatogenic (GH) binding sites were quantified in liver homogenates using125 I-labeled bovine GH as previously described(19). Before binding, these homogenates were treated with 4 M magnesium chloride (MgCl2) to remove endogenously bound hormone and provide a measure of total binding sites. Binding values were expressed as specific binding (B)/total radioactivity (T)× 100/mg of protein present in the receptor preparation. The protein content of homogenates and MgCl2-treated pellets was determined by the method of Lowry et al.(21), and DNA content was measured by fluorometry, as described by Karsten and Wollenberger(22).

Measurement of liver IGF-I peptide concentrations. To measure liver IGF-I content, 0.05-0.1 g of each liver sample was pulverized in a mortar and pestle under liquid nitrogen, mixed with 1 mL of ice-cold acetic acid, vortexed vigorously, and incubated on ice for 2 h(9, 23). After centrifugation at 20 000 ×g for 30 min at 4°C, the supernatant was transferred to a polypropylene tube, and the pellet was reextracted using the same procedure. Both supernatants were combined and desiccated in a speed-vac concentrator(Savant Instruments, Farmingdale, NY). The material was stored at -20°C and reconstituted in assay buffer 24 h before IGF-I RIA. Using this technique, 82% of pure IGF-I added to liver extracts was recovered(23). The Hb concentration was also measured by the cyanomethemoglobin method on aliquots of each liver extract reconstituted in Tris buffer, pH 7.4. The amount of IGF-I in the extract that was derived from blood contamination was subtracted from the total extractable IGF-I, as described previously(9). The serum content of the liver samples ranged between 7.4 and 16.6 μL/g liver and constituted 3.5-7.8% of the liver IGF-I content. The results are expressed as nanograms of IGF-I per g of tissue.

RNA preparation and Northern blot hybridization. Total RNA was isolated by the guanidine thiocyanate/cesium chloride method(24) from pools of livers from two 8-d-old pups (three pools of two pups per group), and from separate liver samples in the 22- and 63-d-old rats (three animals/group). Total RNA (20 μg/lane) samples were size-fractionated and transferred to nylon membranes (Hybond-N, Amersham) as described previously(3, 9). Blots were hybridized with a 194-bp exon 4-specific rat IGF-I riboprobe, using prehybridization, hybridization, and washing conditions reported previously(9). Hybridization of blots with a chicken32 P-labeled β-actin cDNA, 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 63-d-old male controls.

Measurement of IGFBP. To assess the abundance of serum IGFBP, ligand blotting was performed using PAGE as previously described(25). Electrophoresed proteins were transferred to nitrocellulose membranes which were incubated for 24 h with 200 000 cpm/mL of a mixture of 125I-IGF-I and 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 45 000-39 000 Mr (IGFBP-3), the 34 000-29 000 Mr (IGFBP-1 and- 2), and the 24 000 Mr(IGFBP-4) bands. The 34 000-29 000 Mr bands representing IGFBP-1 and-2 were analyzed together because they could not be reliably separated for quantification by scanning densitometry. The abundance of IGFBP in a pool of sera from normally fed adult rats served as control. The sizes of radioactive protein bands were estimated by comparison to standards from Life Technologies, Inc. (Gaithersburg, MD).

Statistical analysis. Statistical differences among treatment groups were analyzed by an unpaired t test. For body weight and weight velocity which were measured sequentially, the differences were analyzed by two-way ANOVA using the Statistical Analysis System(26) to determine the respective influences of the two main factors (gestational diet and age of the rats) and their interactions on the different variables studied. Because of variance heterogeneity, data were log-transformed before ANOVA. Results are shown as the mean ± SEM, andp < 0.05 was considered significant.

RESULTS

Effects of gestational protein restriction on dams and newborn pups. Mean daily food intake throughout gestation was similar between control and protein-restricted dams (19.6 ± 0.6 and 19.3 ± 1.2 g/d, respectively). However, the body weight gain of protein-restricted dams was markedly impaired (P5 at term: 58.8 ± 6.1 g versus P20 controls, 119.6 ± 5.1 g; p < 0.001). In newborn pups, gestational protein restriction caused a 28% decrement in body weight and a 4-fold increase in mortality within 48 h after birth (Table 1).

Table 1 Litter size, birth weight, and neonatal mortality(deaths within 48 h after birth) of pups born from dams fed a normal protein diet (P20) or a low protein diet (P5) throughout gestation
Full size table

Postnatal growth. Despite food rehabilitation during suckling, the mean body weight of the P5 pups remained significantly lower at 8 and 22 d than that of age-matched control pups (Table 2 and Fig. 1). Tail length was also reduced at 22 d, as was the weight of liver, heart, brain, and kidney (P5 pups compared with controls: -17 to -35%;p < 0.01). Only the weight of the spleen did not show significant change (Tables 2 and 3). The reduction in liver weight was due mainly to decreased cell number, as indicated by the decrease in the protein and DNA content, without significant change in the protein/DNA ratio (Table 2). Food intake was recorded after weaning and was similar between P20 and P5 rats (data not shown).

Table 2 Body weight, tail length, and indices of liver growth in rats born from dams fed a P20 or a P5 diet throughout gestation, and killed 8, 22, and 63 d after nutritional rehabilitation from birth onward
Full size table
Figure 1
figure 1

Body weight (upper panels) and weight velocity (bottom panels) in male and female rats born from dams fed throughout gestation a normal protein diet (P20; filled circles) or a low protein diet (P5; open circles). The inset in left upper panel contains pooled data from male and female rats, plotted regardless of sex, because there are no gender-related differences at this age(d 0-21). Data are presented as the mean ± SEM of four litters (d 0-21) or six rats (d 22-63) per group.

Full size image
Table 3 Organ weights in rats born from dams fed a P20 or a P5 diet throughout gestation, and killed 8, 22, and 63 d after nutritional rehabilitation from birth onward
Full size table

After long-term food rehabilitation from birth until d 63, the tail length of male and female P5 rats normalized, as did liver weight and protein and DNA content (Table 2). Also, the weight of heart, brain, spleen, and kidney of the P5 rats was restored to control values (94-109%), except for the kidney weight in female rats (Table 3). In contrast, the body weight of the P5 rats remained lower than normal(Table 2; Fig. 1, upper panels). The percentage difference in body weight between the two dietary groups progressively diminished with age, being 28% at birth (p < 0.001), 25% at 8 d (p < 0.001), 18% at 22 d (p < 0.01), and 10% at 63 d (p < 0.05), indicating that the P5 rats had resumed normal growth velocity but had not experienced catch-up growth(Fig. 1, bottom panels).

Serum IGF-I and liver IGF-I peptide and mRNA concentrations. In P20 controls, serum and liver IGF-I concentrations increased with age and were lower in females than in males at 63 d of age (p < 0.01, respectively; Fig. 2). Serum and liver IGF-I concentrations in P5 rats were reduced by 40% at 8 d of age (p < 0.001 versus controls), and by 28 and 42%, respectively, at 22 d of age (p < 0.01 and p < 0.001 versus controls; Fig. 2). By d 63, however, serum and liver IGF-I values were normal in P5 animals.

Figure 2
figure 2

Serum IGF-I (A) and liver IGF-I concentrations (B) in 8-, 22- and 63-d-old rats born from dams fed a normal protein diet (P20; controls; open bars) or a low protein diet(P5; hatched bars) throughout gestation, and killed 8, 22, and 63 d after nutritional rehabilitation from birth onward. Values are shown as the mean ± SEM of duplicate measurements obtained from 6-8 pups/group for the 8- and 22-d-old animals, and 6 rats/group in the 63-day age group.**p < 0.01 and ***p < 0.001 vs P20 controls.

Full size image

Northern blot analysis of liver IGF-I mRNA revealed the four expected mRNA size-classes (7.5, 4.7, 1.7, and 0.7-1.2 kb), and the expected increase in IGF-I mRNA with age (Fig. 3). In contrast to liver and serum IGF-I peptide concentrations, there was no difference in liver IGF-I gene expression between the P5 and control rats at any age studied.

Figure 3
figure 3

Effect of prenatal protein malnutrition on liver IGF-I mRNA concentrations in the progeny. (A) Autoradiogram of a representative Northern blot analysis of liver IGF-I mRNA in 8-, 22- and 63-d-old rats born from dams fed a normal protein diet (P20; controls) or a low protein diet (P5) throughout gestation and killed 8, 22, and 63 d after nutritional rehabilitation from birth onward. Each lane represents 20 μg of total RNA from a pool of livers of two 8-d-old pups or from one rat in the 22- and 63-d-old rats. (B) Relative abundance of total liver IGF-I mRNA(the sum of all bands for each sample) in P20 (open bars) and P5(hatched bars) at different ages studied. Results are shown as the mean ± SEM of three rats per group, except for the 8-d-old pups where three pools of two pups were used. Each mRNA value is the mean calculated from two different blots run with the same liver RNA preparation.

Full size image

Serum IGFBP, insulin and GH concentrations, and liver GH binding capacity. Determination of serum IGFBP in control rats also revealed a developmental profile, with emergence of the 39 000-45 000 Mr band (IGFBP-3) that was similar in time course to the increase in serum IGF-I concentrations (Figs. 2 and 4). At d 8, the predominant IGFBP in serum were the 29 000-34 000 Mr proteins corresponding to IGFBP-1 and -2. At this age, minimal binding was observed at the 24 000 Mr band (IGFBP-4) and the 39 000-45 000Mr band was not detected. By d 63, the 39 000-45 000Mr band predominated and the 29 000-34 000Mr band had decreased. There were no significant differences between IGFBP of the control and P5 rats at any age studied (Fig. 4 and Table 4).

Figure 4
figure 4

Autoradiogram of Western ligand blot of serum IGFBP at 8, 22, and 63 d of age in rats born from dams fed a normal protein diet (P20; controls) or a low protein diet (P5) throughout gestation and killed at d 8, 22, and 63. A pool of sera from normal-fed adult rats (NL) was used for comparison. The Mr of the IGFBP is indicated on the right.

Full size image

Serum insulin concentrations at 8, 22, and 63 d were comparable in control and P5 animals, as were serum GH concentrations. Likewise, total liver somatogenic binding sites did not differ between the dietary groups(Table 5).

Table 5 Serum insulin and GH concentrations, and total somatogenic binding sites in rats born from dams fed a P20 or a P5 diet throughout gestation, and killed 8, 22, and 63 d after nutritional rehabilitation from birth onward
Full size table

DISCUSSION

Previous studies have shown that the decreases in body and organ weight observed in offspring of protein-restricted dams persist throughout the suckling period or into maturity, and that these growth deficits result from reduced cellularity(2, 4). We confirm that despite food rehabilitation during the suckling period, pups born from protein-malnourished dams do not undergo immediate catch-up growth. Long-term rehabilitation, however, is associated with complete catch-up growth of tail length and organ weight, but not of body weight. Similar results were found in rats that were milk-deprived during the first 21 d of life(17).

The adverse effects of dietary protein restriction on postnatal growth and development of the pups were accompanied by parallel changes in liver and serum IGF-I concentrations. During the weaning period, body and organ growth of the P5 pups were impaired. At the same time, serum and liver IGF-I concentrations were low. By d 63, tail length and organ weights had recovered, and serum and liver IGF-I concentrations were normalized. These results suggest that failure of somatic growth during postnatal development may result from decreased serum IGF-I. They also support the role for IGF-I in the control of postnatal growth and development in the rat(15, 17, 27). The lack of complete normalization of body weight in the P5 rats despite normalization of their serum IGF-I levels might be due to a state of partial resistance to the growth-promoting effects of IGF-I at the tissue level. Indeed, young protein-restricted rats do not undergo somatic growth in response to exogenous IGF-I(13). Normalization of skeletal and organ growth after prolonged food rehabilitation, however, mitigates against IGF-I resistance.

Most of the weight gained in older rats is due to increasing adipose mass(28). The size of the adipose depot in man and animals is ultimately dependent upon the number and size of adipocytes(29, 30). Williams et al..(28) suggested that the failure of complete recovery of body weight in rats that are growth-retarded during the suckling period is due to underdevelopment of adipose tissue. Anguita et al..(31) reported decreased body weight and fat in 53-d-old rats born from dams that were food-restricted during the first 2 wk of gestation. The offspring of rats protein-restricted throughout gestation have permanently stunted body weight and reduced adipose tissue weight, cell number, and size(4). Finally, the capacity of adipose tissue for hyperplasia resides in a group of fibroblast-like preadipocytes whose proliferation and differentiation is regulated by IGF-I(32). Although we did not quantify changes in adipose tissue, it is tempting to speculate that at least part of the body weight deficit in rats that have been malnourished in utero might be due to an IGF-I-related deficit in adipose tissue development. However, other effects of intrauterine malnutrition, including reduced insulin secretion, may have an important impact on adipose development.

Alterations in GH secretion could have contributed to decreased serum IGF-I concentrations and growth deficiency in our protein-deprived animals, as IGF-I is partially linked to GH at this stage of life(18, 33). There was, however, no significant difference between serum GH concentrations of prenatally well fed and underfed pups, making impaired GH secretion an unlikely mechanism. Because GH does not appear to exhibit pulsatile secretion until after 25-30 d of age(34), a single determination of serum GH concentrations seems to be sufficient to assess GH secretion in neonatal rats. Liver GH binding was not reduced in the growth-retarded pups after gestational protein restriction. Furthermore, we have reported that changes in binding of GH to its receptors are not critical for reduction of serum IGF-I during protein restriction of postweanling rats(19).

We were surprised that no reduction in liver IGF-I mRNA concentration was observed postnatally in pups born from protein-malnourished dams, thus excluding persistence of the attenuation in IGF-I gene transcription observed at birth in pups that are growth-retarded by maternal protein restriction(3). Therefore, the reduction observed in serum and liver IGF-I concentration in the P5 rats might result from decreased translation of the IGF-I mRNA. An example of such a phenomenon is albumin, for which steady state levels of liver albumin mRNA are normal during starvation, whereas albumin synthesis is reduced because the mRNA is dissociated from the ribosomes(35). Increased clearance of serum IGF-I and/or decreased stability of IGF-I peptide could also account for the reduction observed in serum IGF-I concentrations in the P5 rats. A secretory defect of IGF-I from liver cells seems unlikely, because there was no accumulation of IGF-I in livers of P5 rats.

There is evidence that the IGFBP influence clearance and degradation of serum IGF-I(36, 37). Changes in serum IGFBP concentrations therefore might have contributed to decreased serum IGF-I in neonatal rats after gestational dietary protein restriction. Our results, however, show that neither the relative abundance of the various IGFBP, nor the pattern of transition from predominant IGFBP-1 and -2 in neonatal serum to IGFBP-3 as the major form in adult serum(38), were altered by gestational protein malnutrition. Insulin is the major hormonal and metabolic regulator of IGFBP-1 and -2 expression(39), and serum insulin levels were not different between growth-retarded and control rats in our study.

In conclusion, our study shows that the long-term adverse effects of gestational protein malnutrition on the postnatal growth of the progeny are accompanied by parallel changes in circulating and liver concentrations of IGF-I. This observation further supports a role for IGF-I in the control of catch-up growth after intrauterine growth retardation. There is, however, no concomitant alteration in liver IGF-I mRNA, serum IGFBP, insulin, and GH concentrations, suggesting that reduced IGF-I in these rats results mainly from a translational defect in IGF-I synthesis.

Table 4 Serum IGFBP in rats born from dams fed a P20 or a P5 diet throughout gestation, and killed 8, 22, and 63 d after nutritional rehabilitation from birth onward
Full size table