Main

Multiple lines of evidence indicate that IGF-I has an important role in normal brain development(1). IGF-I and the type 1 IGF receptor are expressed early in the development of the CNS. As determined by in situ hybridization histochemistry, IGF-I expression in the developing rodent brain begins by embryonic d 14 and occurs predominantly in neurons(2,3). IGF-I overexpressing Tg mice exhibit early postnatal overgrowth of the brain that is due to increased myelin content and cell number(4–8). In contrast, IGFBP-1 Tg mice that ectopically express this protein in brain exhibit brain growth retardation beginning in the first weeks of postnatal life(6,9), the time in development when endogenous IGF-I gene expression is maximal in the rodent brain(10). Likewise, mice with homozygous deletions of the genes for either IGF-I or the type 1 IGF receptor have small, undermyelinated brains(11–13).

It has long been known that nutrition profoundly influences growth and maturation of the developing brain(14–16). Prenatal and/or postnatal malnutrition delays maturation of the brain, leading to decreased brain size, underdevelopment of neuronal elements, and a reduced number of synapses. Because the expression of IGF-I and some IGFBPs are markedly altered in undernutrition(17), we postulated that reduced brain IGF-I expression and/or actions mediate the brain pathophysiology of malnutrition and that brain IGF-I overexpression will ameliorate the brain growth retardation caused by undernutrition, whereas ectopic IGFBP-1 overexpression will exacerbate it.

We induced undernutrition in newborn IGF-I overexpressing Tg mice, IGFBP-1 Tg mice, and their non-Tg littermates by separating half of each litter from their dam during the suckling period, the time of greatest vulnerability to malnutrition. Early postnatal brain growth in undernourished IGF-I Tg mice was comparable to that in well-fed control mice, whereas the brain growth of undernourished IGFBP-1 Tg mice was more retarded than that of undernourished control mice. Our results indicate that IGF-I has protective actions against the effects of malnutrition in the developing brain.

METHODS

Mice. Studies were conducted in IGF-I and IGFBP-1 Tg mice carrying human IGF-I and IGFBP-1 transgenes driven by the mouse metallothionein-1 promoter(4,6,9). In each of these Tg mouse lines, expression of the transgene is high in brain and is first detectable at the time of birth. IGF-I Tg mice were bred as heterozygotes because we have been unable to generate homozygotes, and two lines of IGF-I Tg mice (32L and 52L) were studied(6,8). Both heterozygous and homozygous IGFBP-1 Tg mice were studied in different experiments as indicated. Non-Tg littermates of the heterozygous Tg mice served as normal controls. The genotypes of mice were confirmed by Northern blot hybridization analysis for the transgenes. The generation and care of mice as well as the protocol used were approved by the Institutional Review Committee of the University of North Carolina at Chapel Hill.

Experimental design. Undernutrition of mice was accomplished with a modification of the method described for rats by Fuller et al.(18,19). Newborn mice were adjusted to six to eight mice per litter by eliminating extra pups or by fostering pups from other appropriate litters delivered on the same day. Undernutrition was induced in newborn IGF-I and IGFBP-1 Tg mice and their normal littermates by separating half the pups in each litter from their lactating dams for 4 h on P 1, for 8 h on P 2, and for 12 h thereafter. All mice were housed in a temperature- and light-controlled room and weighed each day to ensure that control mice were subjected to the same degree of handling and stress. Pups were weighed and killed on P 21. Brain and liver were removed and weighed; then the brains were dissected into five regions and each was weighed.

Histology. To assess cytoarchitectural changes induced by early postnatal undernutrition, we evaluated the size of the PMBSF, a portion of layer IV of the CTX(20). We selected this anatomic region for several reasons: a) IGF-I is likely to stimulate its development because IGF-I mRNA is well-expressed in this region during the suckling period in rats(21); b) we have shown that IGF-I availability influences PMBSF growth because its size is increased in IGF-I Tg mice and decreased in IGFBP-1 Tg mice(8); c) the effects of undernutrition on PMBSF growth during the suckling period in rats are well documented and include decreases in size and increases in cell density(22); and d) its dimensions can be precisely delineated with histologic techniques, allowing comparisons among mice. The methods used were identical to those that we previously described(8). Briefly, mice were perfused with normal saline followed by 10% glycerol in phosphate buffer (0.1 M, pH 7.4), the brains were removed, and the cerebral hemispheres were dissected, flattened, and frozen. Cryostat tangential sections (20 µm) of the cortex were obtained and processed for cytochrome oxidase histochemistry. PMBSF two-dimensional maps were then traced from these sections at a final magnification ×40 with the aid of a camera lucida. Completed maps were digitized and PMBSF cross-sectional areas were measured by use of the Image-Pro imaging analysis system.

Data analysis. Statistical significance was assessed by oneway ANOVA or t test. Data are presented as the mean ± SD, and p < 0.05 was considered significant.

RESULTS

Brain and body weight. At P 21, the brain weights of IGF-I overexpressing Tg mice (0.53 g) were larger than those of control mice (0.43 g; p < 0.05), whereas the brains of IGFBP-1 Tg mice were smaller (0.39 g; p < 0.05) (Fig. 1, panel A), findings that are consistent with our previous reports(4,6,8,9). Undernourished control mice had smaller brains (0.35 g) than well-fed control mice (0.43 g; p < 0.05). The brain weights of undernourished IGF-I Tg mice (0.42 g) were similar to those of well-fed control mice (0.43 g; p = ns) and larger than those of undernourished control mice (0.35 g; p < 0.05), whereas the brains of undernourished IGFBP-1 mice (0.32 g) were more retarded than those of undernourished control mice (74.8 and 82% of well-fed non-Tg control mice, respectively, p < 0.05; see Fig. 1, panel B). Unlike in the brain, there were no significant differences in body weight among any of the three groups of well-fed mice (Table 1). This is consistent with our previous findings and is likely because the transgenes in these mice are not highly expressed in organs other than the brain and because these Tg mice have nearly normal circulating IGF-I concentrations [(6,8), and our unpublished data]. Marked decreases in body weights (54 to 59%) were apparent in all three groups of undernourished mice, and there were no differences among the groups. The differences in brain/body weight ratio (Fig. 1, panel C), therefore, are due solely to changes in brain weight and are not secondary to modification of overall somatic growth. Similarly, liver weights among the groups of well-fed mice did not differ and, although markedly reduced after dietary restriction, there were no differences among the groups of undernourished mice (Table 1).

Figure 1
figure 1

The effect of undernutrition on brain weight in 32L IGF-I Tg mice (solid bars), IGFBP-1 Tg mice (hatched bars), and their non-Tg littermate controls (open bars). Panel A shows whole brain weights in grams, whereas in panel B brain weights are calculated as a percentage of brain weights in well-fed controls. Panel C shows the brain/body ratio calculated as a percentage of that in well-fed controls. The broken horizontal line in panels B and C represents the mean (100%) in well-fed controls. All values represent the mean ± SD. *p < 0.01, compared with well-fed controls. †p < 0.05 and ‡p < 0.01, compared with undernourished controls.

Full size image
Table 1 The effects of undernutrition on body and liver weight (g, mean ± SD) in 32L IGF-I Tg mice, IGFBP-1 Tg mice, and their non-Tg (control) littermates
Full size table

As we have previously reported(4,6), brain weights do not differ between Tg and non-Tg mice at birth, because these transgenes are not significantly expressed before birth. To mean brain weight of 103 mg at birth, which is derived from the mean brain weights at birth of all groups of mice. The growth of undernourished IGF-I Tg mice was comparable to that of well-fed control mice (increases of 4.13- and 4.22-fold, respectively; p = ns) and greater than that of undernourished control mice (increase of 3.45-fold; p < 0.01; see Fig. 2), whereas undernourished IGFBP-1 Tg mice exhibited less growth than undernourished control mice (increases of 3.15-and 3.45-fold, respectively; p < 0.05).

Figure 2
figure 2

Brain growth from birth to P 21 in well-fed and undernourished 32L IGF-I Tg mice (solid bars), IGFBP-1 Tg mice (hatched bars), and their non-Tg littermate controls (open bars). Assuming a mean of 0.103 g for all mice (there are no differences in brain weight between Tg and control mice at birth), brain growth was calculated as the percent increase of birth brain weight. Values represent the mean ± SD. *p < 0.01, compared with well-fed controls. †p < 0.05 and ‡p < 0.01, compared with undernourished controls.

Full size image

To determine whether IGF-I availability modulated the effects of undernutrition differently in distinct brain regions, we assessed the size of the following five brain regions: BS, CB, CTX, DIE, and HIP in well-fed and undernourished mice from each group at P 21 (Table 2 and Fig. 3). In this experiment, a different line of IGF-I Tg mice (52L) was studied to be certain that the results obtained in the first experiment were because of IGF-I overexpression rather than an artifact of a mutation caused by the insertion of the transgene. In addition, homozygous rather than heterozygous IGFBP-1 Tg were studied in an attempt to maximize the expression of the transgene and possibly the consequences of its expression. The differences in whole brain growth among the three groups of mice were similar to the first experiment, both with and without adequate nutrition. Undernutrition, however, influenced the growth of different brain regions to different degrees. In non-Tg control mice, undernutrition affected the CB most and the DIE least (Fig. 3). The order of impact and degree of growth retardation in control mice, calculated as the percent reduction from that of well-fed control mice, was as follows: CB (24.2%) > HIP (20.5%) > BS (12.3%) ≅ CTX (11.2%) > DIE (4.6%). In IGF-I Tg mice, the brain regions that expressed the transgene best (CTX ≅ HIP > DIE > BS > CB)(6,23) also exhibited the most growth under both well-fed and undernourished conditions. In fact, HIP (117.2% of normal well-fed), CTX (110.5%), and DIE (112.3%) overgrew despite undernutrition. Further, in an additional experiment with a single litter of 32L IGF-I Tg, we found that the HIP, CTX, and DIE of these IGF-I Tg mice maintained nearly normal growth in the face of undernutrition.

Table 2 The effects of undernutrition on whole and regional brain weights (mg, mean ± SD) in 52L IGF-I Tg mice, IGFBP-1 Tg mice, and their non Tg littermates
Full size table
Figure 3
figure 3

The effects of undernutrition on whole and regional brain weight in 52L IGF-I Tg (solid bars), IGFBP-1 Tg mice (hatched bars), and their non-Tg littermate controls (open bars). Brain weight is calculated as a percentage of that in well-fed controls. The broken horizontal line represents the mean (100%) in well-fed controls. WB indicates whole brain. Values represent the mean ± SD. *p <0.05 and **p <0.01, compared with well-fed controls. †p <0.05, compared with undernourished controls.

Full size image

In well-fed IGFBP-1 Tg mice, all brain regions were undergrown to a similar degree (91.4 to 85.8% of well-fed controls). With restricted nutrition, the undergrowth of each brain region was worsened; however, HIP (61.6% of well-fed controls) and CB (57.1%) were more severely retarded than DIE (67.5%), CTX (74.1%), and BS (79.5%).

PMBSF. PMBSF areas were increased in well-fed IGF-I Tg mice compared with well-fed control mice (0.399 mm2 compared with 0.324 mm2, p < 0.01) (Fig. 4). Although the PMBSF areas in IGFBP-1 Tg mice were smaller than those of control mice (0.314 mm2), this difference was not significant. Our previous finding that PMBSF size is significantly smaller in adult IGFBP-1 Tg mice(8) suggests that a longer developmental period is necessary for the effects of IGFBP-1 to become statistically significant.

Figure 4
figure 4

The effects of undernutrition on the PMBSF in 52L IGF-I Tg mice, IGFBP-1 Tg mice, and non-Tg control mice. The left panel shows representative reconstructed barrel fields drawn with use of a camera lucida from tangential cerebral cortical sections after cytochrome oxidase staining. The right panel shows the sum of the PMBSF barrel areas (mm2; mean ± SD) derived from camera lucida drawings of well-fed (closed bars) and undernourished (open bars) mice in each group. n = number of mice studied in each group. *p < 0.05 and ***p < 0.01, compared with well-fed control mice.

Full size image

Undernutrition significantly reduced PMBSF size in non-Tg control mice (0.263 mm2, as compared with 0.324 mm2 in well-fed controls, p < 0.05) and IGFBP-1 Tg mice (0.273 mm2, as compared with 0.314 mm2 in well-fed IGFBP-1 Tg mice, p < 0.05). Although undernutrition reduced PMBSF size in IGF-I Tg mice, these differences were not significant (0.336 mm2 and 0.399 mm2 in undernourished and well-fed IGF-I Tg mice, p = 0.14). Further, undernourished IGF-I Tg mice had a similar PMBSF size to those of well-fed normal mice (0.336 and 0.324 mm2, respectively; p = ns) and a larger PMBSF size than those of undernourished normal mice (0.336 and 0.263 mm2, respectively; p < 0.01). Each of the latter findings indicates that IGF-I protects the PMBSF from the growth-retarding effects of undernutrition.

DISCUSSION

In previous studies of Tg mice with altered brain IGF-I availability, we showed that IGF-I stimulates brain growth postnatally(4,6,9,24). Specifically, IGF-I overexpressing Tg mice, created with a transgene that is expressed in brain beginning at approximately the time of birth, exhibit marked postnatal brain overgrowth, whereas Tg mice that express IGFBP-1, an inhibitor of IGF-I action, ectopically in brain, exhibit postnatal brain growth retardation. The results reported here confirm these findings and extend them by showing that IGF-I can stimulate nearly normal brain growth in the face of undernutrition. Because IGFBP-1 decreases IGF-I availability and actions, the exacerbation of undernutrition-induced brain growth retardation in IGFBP-1 Tg mice also supports a role for IGF-I in brain growth. Further, this finding suggests that the relatively high endogenous brain IGF-I expression in early postnatal life(10,25) may contribute to the relative preservation of brain growth compared with somatic growth during early postnatal undernutrition. The relative sparing of brain growth during undernutrition is evident in the present study as it is in other such studies(18). For example, brain weights in our undernourished control mice are 82% of those in well-fed mice, whereas body weights are only 42% of those in well-fed controls.

Although the brains of IGF-I Tg mice maintain an essentially normal rate of growth during undernutrition, they nonetheless are affected by nutritional deprivation. Despite their normal size, brains of undernourished IGF-I Tg mice are 75-79% the size of well-fed IGF-I Tg mice (depending upon the line of Tg mice studied). In other words, the relative brain growth retardation imposed by undernutrition in IGF-I Tg mice is similar or greater than that imposed on the brain of non-Tg control mice (81 and 86% of well-fed controls in two different experiments), but undernourished IGF-I Tg mice maintain a normal rate of growth. Undernutrition also reduces the brain size of IGFBP-1 Tg mice to 82% of well-fed IGFBP-1 Tg mice, but because the brain growth rate is reduced, undernourished IGFBP-1 Tg mice exhibit a disproportionate reduction in brain weight. IGF-I, therefore, seems to ameliorate the brain growth retarding effects of early postnatal undernutrition by stimulating the rate of brain growth rather than by directly antagonizing the brain growth retardation caused by nutritional deprivation.

The brain overgrowth in IGF-I Tg mice is because of a marked increase in myelination(5,6), a modest increase in cell number as evidenced by increased DNA content(4), and, where studied, increased neuron(8) and oligodendrocyte number(6). The increase in cell number is likely due in large part to increased neuronal survival (i.e. decreased apoptosis) rather than cell proliferation because the transgenes in our Tg mice are first expressed after birth(6,9), after the time when most neurogenesis occurs but before major waves of apoptosis. As judged from our detailed histologic studies of the PMBSF in Tg mice(8), IGF-I also appears to stimulate neuritic outgrowth. In IGF-I Tg mice, we found a marked increase in the PMBSF volume occupied by neuropil, the area composed of dendrites and axons (as well as their myelin sheaths), rather than by cell bodies. The increased neuropil volume in IGF-I Tg mice also results in a decrease in neuron density despite the increased neuron number. These changes are in marked contrast with the deficits in myelination and increased brain cell density found in IGFBP-1 Tg mice(8) and observed after undernutrition imposed during the suckling period(15,22). The findings in the PMBSF reported here confirm our previous observations that IGF-I stimulates an increase in PMBSF size(8) and that undernutrition retards PMBSF growth(22). Further, we found that undernourished IGF-I Tg mice had similar or slightly larger PMBSF sizes compared with well-fed controls. This finding further supports our conclusion that IGF-I ameliorates the effects of undernutrition on brain development. Because PMBSF size reflects the magnitude of neuritic outgrowth and, to a lesser extent, the degree of myelination(22), these findings also indicate that IGF-I most likely acts by lessening the impact of undernutrition on neuron growth and myelination.

Undernutrition during the suckling period differentially influences the growth of distinct brain regions, and the magnitude of these effects differs somewhat between control and Tg mice. Although the reasons for these differences among brain regions are not clear, two factors appear to be important: 1) the degree of development and growth occurring in a specific region during the period of undernutrition, and 2) the magnitude of transgene expression in the region. Significant cell proliferation in the CB and dentate gyrus of the HIP, predominantly of granule cells(26–28), occurs during the suckling period, and this appears to render these regions especially vulnerable to undernutrition. When the degree of growth retardation imposed by undernutrition is assessed, CB and HIP are the most affected regions in controls (24.2 and 20.5% reductions in CB and HIP weight, respectively), in IGF-I Tg mice (33.0 and 24.5% reduced, respectively, compared with well-fed IGF-I Tg), and in IGFBP-1 Tg mice (34.1 and 28.3% reduced, respectively, compared with well-fed IGFBP-1 Tg). The CTX, which also undergoes significant postnatal development, is affected intermediately in each group of mice.

The magnitude of the IGF-I transgene expression clearly has a major influence on both regional brain growth and the impact of undernutrition on regional growth. The regions with the greatest transgene expression [CTX ≅ HIP > DIE > BS > CB; 30, 25, and 6%, respectively, of CTX and/or HIP(6,23)] are nearly the same as the regions that are most overgrown in well-fed 52L and 32L IGF-I Tg mice (HIP ≅ CTX > DIE > CB ≅ BS) and most preserved in size after undernutrition in both lines (HIP ≅ CTX ≅ DIE > BS > CB). Many other factors undoubtedly influence the impact of undernutrition on regional brain growth, including the magnitude of endogenous IGF-I expression and the sensitivity of the region to IGG-I actions. The regions most affected by IGFBP-1 transgene expression (CB ≅ HIP > CTX > DIE > BS) may reflect the latter factors because the most affected regions are those with relatively high endogenous IGF-I expression during early postnatal life(10,25). In other words, regions whose growth is most influenced by endogenous IGF-I gene expression may be most affected by a blunting of IGF-I action.

Taken together, our findings suggest that a decrease in IGF-I action, caused either by decreased IGF-I expression or resistance to its actions, contributes to the brain abnormalities induced by undernutrition. There are few reports, however, on the effects of undernutrition on brain IGF-I expression and action. The findings that IGF-I mRNA is decreased in whole brain and hypothalamic nuclei of adult rats fasted for short intervals(29,30) and in the CB of undernourished suckling rats(31) suggest that reduced IGF-I expression may be a mechanism mediating undernutrition-induced brain growth retardation. Regardless of the mechanism(s), our data suggest that IGF-I plays a role in protecting the brain from the full deleterious consequence of undernutrition and point to the important actions of IGF-I in stimulating postnatal brain growth and in protecting the brain against injury.