Abstract
Poor fetal and neonatal nutrition interacts with the environment to produce subtle changes in the development, structure, and function of organs. Over the lifetime, these subtle changes alter an individual’s response to stressors and increase disease susceptibility. One mechanism by which early-life nutrition causes long-term effects is via epigenetic programming. The inherent complexities in epigenetic programming and subsequent modulation of gene expression in response to early life events are beginning to be understood. However, work remains to be done in identifying susceptible genes with phenotypic contributions, as well as in understanding how epigenetic modifications contribute to changes in gene expression during development.
Keywords
Fetal and Neonatal Nutrition, Development, Programming, Epigenetics
Early Nutrition and Adult Phenotype
An infant’s early nutrition affects the adult phenotype ( Table 11.1 ). As old and intuitive as this concept may be, it has re-entered the consciousness of the research community only in the last 20 years. As a result, the overall interest in the specific experiences and mechanisms through which different perinatal nutritional environments lead to adult-onset diseases has grown exponentially. This interest is an important research priority because it is an avenue for preventing adult diseases, such as diabetes, obesity, and hypertension, before they exact a direct toll.
| Attention deficit disorder | Insulin resistance |
| Chronic lung disease | Neurodevelopmental delay |
| Divorce | Neuroendocrine reprogramming |
| Dyslipidemia | Poor postnatal growth |
| Hypertension | Renal insufficiency |
| Immunodeficiency | Schizophrenia |
The continuum of early nutrition experienced by humans varies greatly within a single population, let alone between different populations. As a result, the majority of epidemiologic studies interested in understanding the adult consequences of fetal and neonatal nutrition have used poor growth, both in utero and in the early postnatal period, as a marker of poor nutrition. Therefore an important assumption of this chapter is that poor nutrition leads to poor growth in both the fetus and the neonate. Epidemiologic studies on this subject have historically been limited to infants who are small for gestational age (SGA), which is typically defined as weight <10th percentile. Infants included in this group may be small for multiple reasons, including normal genetic variation. Furthermore, these studies that focus only on SGA infants miss infants who are smaller than they should be but are still >10th percentile. More recently, epidemiologic studies have separated those infants who fail to grow to their genetic potential, that is, those who suffer intrauterine growth restriction (IUGR), from those who are SGA, but do not have IUGR. Generally, these studies include maternal and fetal parameters that contribute to poor growth, such as maternal uteroplacental insufficiency.
Despite these limitations, the epidemiology in this field has been vitally important and often elegant, which is certainly true of the three cohorts (and their respective studies) that have set the standard for understanding the adult consequences of neonatal and fetal nutrition. These cohorts involve the (1) Dutch famine of 1944 to 1945; (2) the early studies by Barker et al., hence the “Barker Hypothesis”; and (3) the more recent Nurses’ Health Study.
The first of these cohorts involves the Dutch famine of 1944 to 1945. This famine occurred as a result of German reprisal for a general railway strike intended to disrupt the transport of German reinforcements against Allied liberation movements. The famine lasted approximately 5 months. Daily rations in Amsterdam dropped from 1800 kcal/day in December 1943 to 400 to 800 kcal/day in April 1945. Although a goal was set for children under age 1 year and for pregnant or lactating women to receive supplemental rations, this was not possible at the height of the famine. After liberation in May 1945, caloric intake increased to >2000 kcal/day. The effects of the famine on the Dutch population have been examined via multiple sources, including population-based cohorts, military induction records, psychiatric registries, and self-reports. One of the more comprehensive studies is the Dutch famine birth cohort study, through which the investigators interviewed 912 individuals who were born at term between November 1, 1943, and February 28, 1947, and assessed socioeconomic factors, lifestyles, and medical histories. Babies exposed to famine conditions early in gestation were not smaller or lighter than nonexposed infants; however, they suffered from an increased incidence of coronary heart disease, hypertension, dyslipidemia, and obesity later in life. Although not quite statistically significant, the adults with coronary heart disease were also more likely to have had lower birth weights and smaller head circumferences. Similarly, a cohort born in the Wilhelmina Gasthuis of Amsterdam between November 1943 and February 1947 revealed an association between maternal experience of famine early in pregnancy and an atherogenic lipid profile (higher low-density lipoprotein/high-density lipoprotein [LDL/HDL] cholesterol ratios). Famine experienced early in gestation also altered the perceptions of affected individuals in that the proportion of people reporting self-perceived poor health was significantly higher in the group of famine exposure early during gestation compared with those who had not experienced the famine conditions in utero.
Babies exposed to the famine during gestation suffered from an increased incidence of obstructive airway disease as well as an increased incidence of microalbuminuria. Interestingly, both these findings were independent of size at birth. Babies exposed to the famine in late gestation were also more likely to exhibit impaired glucose tolerance compared with those who had not experienced the famine conditions in utero. Furthermore, affective disorders occurred more often in those individuals exposed to famine conditions in utero in mid-to-late gestation. Other findings from the Dutch famine cohort that are intriguing, and somewhat frightening, include the associations between experiencing the famine in utero and increased rates of schizophrenia, schizophrenia spectrum disorders, and antisocial personality disorder.
For all three groups of infants with early, mid-, or late gestation exposure to the Dutch famine, the rate of mortality at age 50 years was higher. Group-specific mortality at age 50 years for early, mid-, or late gestation exposure was 11.5%, 11.2%, and 15.2%, respectively. In contrast, mortality at age 50 years among those born after the famine was 7.2%.
The second set of studies is the work by Barker, who, along with his research group, pioneered the epidemiology of the adult consequences of neonatal and fetal nutrition, also known as the Developmental Origins of Health and Disease. As early as 1986, Barker and Osmond noted, in an article in the Lancet , a geographic association between ischemic heart disease in 1968 to 1978 and infant mortality in 1921 to 1925. These authors astutely speculated that “poor nutrition in early life increases susceptibility to the effects of an affluent diet.”
The observation linking infant mortality and adult disease was continued by the Hertfordshire studies, which are still the standard in this area. An initial focus of these studies was a cohort of 5654 men born between 1911 and 1930 in Hertfordshire, England. The 1989 manuscript revealed that men who had the lowest weights at birth and at age 1 year had the highest death rates from cardiovascular disease. A subset of these men was further used to evaluate the relationship between insulin sensitivity and birth weight. Specifically, 468 men born, raised, and living in east Hertfordshire were studied after ingesting a 75-g glucose drink. Men with impaired glucose sensitivity and non-insulin-dependent diabetes were characterized by lower weight at birth and at age 1 year. Furthermore, the percentage of men with impaired insulin sensitivity decreased as weight increased at age 1 year; this progression was statistically significant and independent of the adult body mass. This concept—that reduced growth in early life leads to impaired glucose tolerance—has impacted the way numerous clinicians and investigators now approach perinatal metabolism and nutrition.
Barker’s study and the findings on the people of Hertfordshire continue to contribute significantly to our understanding of the relationship between early nutrition and growth and adult diseases. By 2005, the cohort included 37,615 men and women born in Hertfordshire. Low birth weight in men from this population increases the risk of cardiovascular disease, whereas low birth weight in women predisposes the affected adults to cardiovascular and musculoskeletal disease, as well as pneumonia and diabetes. Although there are gender differences, data suggest that an increase in birth weight by 1 standard deviation (SD) would reduce mortality by 0.86% in both genders at age 75 years.
The third historically significant cohort was the Nurses’ Health Study cohort. This study was established in 1976 when approximately 122,000 married female registered nurses, ages 30 to 55 years, responded to a mailed questionnaire about their life histories. The study has continued through follow-up questionnaires every 2 years, thus eliciting updated histories and medical information. The validity of self-reported birth weights was assessed via the Nurses’ Health Study II, which compared weights recorded on birth certificates to the reported weights. The Spearman correlation coefficient between the two weights was 0.74. This study added weight to the observations of Barker et al. by noting a significantly increased relative risk for non-insulin-dependent diabetes in women who had low birth weight (<5.5 lb) compared with those women of median birth weight (7.1-8.5 lb). The relative risk for women whose birthweight was <5.5 lb and 5.5 lb was 1.88 (1.59-2.21) and 1.55 (1.32-1.83), respectively. Adjusting for age, body mass index (BMI), and maternal history of diabetes strengthened the association between low birth weight and non-insulin-dependent diabetes. No significant effects on relative risk were noted after adjustment for prematurity, multiple births, maternal age at birth, participant breastfeeding, ethnicity, parental occupation at age 16 years, paternal diabetes, participant height, parity, cigarette smoking, and physical activity.
The Nurses’ Health Study findings have also been used to investigate the relationship between birth weight and cardiovascular disease in women. . Nonfatal myocardial infarctions were included as an endpoint of the study if the criteria of the World Health Organization were met. Nonfatal strokes were included as another endpoint if the criteria of the national survey of stroke were met. For every 454-g increase in birth weight, a 5% decrease in the risk of nonfatal myocardial infarction was noted, as was an 11% decrease in the risk of nonfatal stroke. As with the Nurses’ Health Study focusing on insulin resistance, these findings were largely independent of other key factors, such as adult body weight, hypertension, diabetes, lifestyle, and childhood socioeconomic class.
The investigations involving these three cohorts have provided important and seminal insights into the relationship between early growth and nutrition and adult diseases. They have provided an impetus to further studies that identify possible physiologic and molecular mechanisms, which may lead to either in utero interventions or postnatal therapies to moderate the impending risks. The following section will focus on recently identified important biologic targets.
Developmental Biology of the Adult Consequences of Neonatal and Fetal Nutrition
This section is presented in two parts. The first part discusses recent insights into how fetal growth restriction affects phenotype in humans, and animal studies used to investigate possible mechanisms. The discussion is not meant to be all inclusive but focuses on some of the most recent and thought-provoking observations. The second part discusses recent insights into how growth restriction in the premature infant potentially leads to later morbidities and how dietary interventions may either contribute to or moderate these effects.
Fetus/Intrauterine Growth Restriction
Multiple studies from different regions of the world containing racially distinct cohorts have associated fetal growth restriction with adult morbidities, as previously discussed. One of the more recent trends is the realization that the lasting effects of fetal growth restriction are evident in both early life and adulthood. Furthermore, weight gain and nutrition modify the effect of IUGR on these early results. These findings pertain to issues involving glucose homeostasis, lipid biology, and hypertension; however, an important central theme to this literature, as well as the literature focusing on adult phenotype, is that cohorts differ, whether as a result of fetal growth restriction or postnatal consequences. Consequently, the findings of these studies differ slightly, which has led to the conundrum that multiple mechanisms may be involved. For example, Mericq et al. followed up SGA and appropriate for gestational age (AGA) infants through the first 3 years of life. This group measured infants’ weights and lengths at 1, 2, and 3 years. At 48 hours of life, glucose and insulin levels were measured in these infants, and at 1 and 3 years of life an intravenous glucose tolerance test was administered after an overnight fast. A calculation of insulin resistance was determined by using the homeostasis model. At 48 hours of life, SGA infants had lower insulin levels compared with AGA infants. At 1 and 3 years of age, SGA children exhibited higher fasting insulin levels and lower insulin sensitivity compared with AGA children.
Growth restriction also appears to affect cellular energy homeostasis. Chessex et al. compared 6 SGA and 13 AGA newborns, in terms of energy expenditure. Energy expenditure in SGA infants was 4.8 kcal/kg/day greater than in AGA infants, primarily via increased fat oxidation. In contrast, when infant oral glucose disposal was studied in prepubertal and early pubertal children (ages 8-14 years) with a history of IUGR, and compared with that of healthy age- and weight-matched control children, lower glucose oxidation characterized the IUGR subjects, although no decrease in overall energy expenditure was noted. Interestingly, lipid oxidation was increased nearly twofold, although not significantly based on variation.
The study by Arends et al. also noted other differences between SGA and AGA counterparts, similar to the large epidemiologic studies discussed above. First, systolic blood pressure was significantly increased in the SGA children. Second, although fasting serum free fatty acids, triglycerides, total cholesterol, HDL, and LDL levels were not significantly different between groups, 6 of the 28 children in the SGA group had serum free fatty acids above the normal range. In terms of lipid biology, much of the focus has been on leptin and adiponectin. For example, in a population of 1-year-old SGA infants from Santiago, Chile, SGA infants were characterized by decreased leptin levels (0.29 ± 0.19 nM versus 0.40 versus 0.07 nM; P < 0.05), as well as a trend toward increased triglycerides ( P = 0.053).
The next sections will delve into specific mechanisms through which IUGR induces many of these phenotypic changes involving glucose metabolism, lipid homeostasis, and other cardiovascular risk factors. The list is neither exhaustive nor exclusive, and, in fact, evidence continues to accumulate that IUGR adult phenotype is a result of many moderate adjustments that are complementary and probably interdependent ( Table 11.2 ). One theme that becomes evident when looking at the whole body of IUGR physiologic studies is that many of the adjustments that are implemented early not only provide short-term advantages in terms of survival but also lead to adult morbidities, such as diabetes, dyslipidemia, and cardiovascular disease, with the passage of time ( Fig. 11.1 ).
| Adiponectin | Leptin |
| Androgens | Glucocorticoids |
| Angiotensin | Insulin and its binding proteins |
Metabolism
Leptin
Leptin is a 167–amino-acid protein that is produced by adipocytes, which performs multiple functions. Among the most important functions of leptin is the regulation of hypothalamic centers that determine, at least in part, whole-body energy expenditure and fat mass. In general, leptin increases energy expenditure and reduces food intake. As a result, humans and rodents lacking either leptin or the leptin receptor develop severe obesity and hyperphagia. In practical terms, increased serum levels of leptin characterize most obese conditions, and in fact, leptin levels typically correlate directly with body fat mass and BMI. The failure of increased levels of leptin in regulating weight loss suggests a potential state of leptin resistance in many cases of obesity.
Within the context of IUGR, maternal serum leptin levels are generally noted to be increased, whereas fetal IUGR serum leptin levels are generally noted to be decreased. For example, Pighetti et al. measured maternal and fetal serum leptin levels in 43 “normal” term pregnancies and 27 pregnancies complicated by asymmetric IUGR. Women from both groups had normal pregravid BMIs (20-27 kg/m 2 ); pregnancies complicated by diabetes or hypertension were excluded. In utero ultrasonography identified fetal growth restriction by an abdominal circumference below the 10th percentile, which was confirmed at birth if the birth weight was similarly below the 10th percentile. IUGR correlated with increased serum leptin levels in the mothers (45 ng/mL versus 29 ng/mL; P < 0.01) and decreased umbilical cord serum levels in the infants (8.4 ng/mL versus 13.1 ng/mL; P < 0.01). No significant differences were noted between male and female infants in either group. As expected, umbilical cord serum leptin levels correlated with neonatal birth weight.
In general, maternal undernutrition or malnutrition results in growth restriction and rapid catch-up growth such that by early adulthood the body weight of the progeny with IUGR exceeds that of the control. In most animal studies, the animals develop components of the morbidities afflicting growth-restricted humans, including insulin resistance, dyslipidemia, and hypertension. In a remarkable methodical study, Fernandez-Twinn et al. reduced maternal protein intake to 50% of controls and measured the circulating levels of several hormones, including leptin, throughout the pregnancies. The low-protein diet significantly reduced both placental and fetal body weights, and at term, the low-protein diet also significantly decreased the placental/body weight ratios. In adulthood, the pups that suffered intrauterine low-protein undernutrition developed diabetes, hyperinsulinemia, and tissue insulin resistance in adulthood. The low-protein diet increased maternal serum leptin levels at day 17 of gestation (term 21 days) and decreased maternal serum leptin levels at term. No significant differences in perinatal serum leptin levels were noted between the pups suffering the low-protein maternal diet and the control pups.
To begin the process of defining leptin homeostasis in the postnatal growth-restricted rat, Krechowec et al. used a similar model of maternal food restriction, in which dams received 30% of ad libitum food intake versus the controls. This study demonstrated that prenatal malnutrition may lead to the developmental of leptin resistance, particularly in adult female rats. Mechanisms responsible for this phenomenon may include cross-talk between the insulin and leptin receptors, as well as hepatic insulin resistance leading to hypertriglyceridemia and subsequent impairment of leptin transport across the blood–brain barrier. Regardless, such studies suggest that leptin biology plays a significant role in the effects of early nutrition on the adult phenotype. Future studies will delve further into “chicken and egg” issues to more clearly differentiate the relative importance of leptin and other molecules produced by adipocytes.
Adiponectin
Adiponectin is also produced by adipocytes, and its receptors adiponectin receptor-1 and adiponectin receptor-2 are found in skeletal muscle and liver, respectively. A 244–amino-acid protein is translated from the most abundant messenger RNA (mRNA; apM1) found in human adipocytes. In mice, administration of adiponectin reduces insulin resistance and serum glucose levels. In humans, serum levels of adiponectin correlate directly with insulin sensitivity, but not with serum lipid profiles or obesity.
In growth-restricted humans, the impact of the early malnutrition appears to vary with age. Iniguez et al. measured leptin levels in 1- and 2-year-old SGA and AGA infants. The SGA infants experienced catch-up growth such that at 2 years of age, only moderate differences existed between the two groups. Although no significant differences existed between SGA infants and control infants in absolute serum adiponectin levels, differences in adiponectin levels were inversely related to weight gain between 1 and 2 years of age. In contrast to leptin, adiponectin levels were unrelated to insulin levels, and multiple regression analysis found that adiponectin related only to postnatal age. If postnatal age was excluded from the analysis, then determinants of adiponectin levels included lower postnatal body weight ( P < 0.001) and male gender ( P < 0.03). Although the findings on adiponectin levels are not as immediately satisfying as the leptin data in this study, those findings are thought provoking because of the many gender-specific effects of early growth restriction that have been noted.
In contrast to the study by Iniguez et al., in the study by Lopez-Bermejo et al., adiponectin levels were measured, and both insulin resistance and insulin secretion were assessed by using the homeostasis model of assessment (HOMA) in 32 prepubertal SGA children (mean age 5.4 ± 2.9 years) and AGA children (mean age 5.9 ± 3.0 years). Gestational age–adjusted birth weight <10th percentile and >25th percentile defined the SGA and AGA children, respectively. As expected, SGA infants were lighter and shorter but had BMIs similar to those of controls. HOMA analysis of SGA and AGA children >3 years of age demonstrated a tendency toward insulin resistance in the SGA children ( P = 0.046) after adjustments for gender, age, and BMI SD score. Surprisingly, serum adiponectin levels were significantly higher in the SGA children ( P < 0.0001), although when the data from the SGA children were broken into BMI quartiles, the findings became more complicated. Compared with lean SGA children, the higher-quartile SGA children had lower serum adiponectin levels ( P = 0.02). The higher-quartile SGA children also had markedly higher fasting insulin levels ( P = 0.03), borderline higher HOMA insulin resistance ( P = 0.05), and higher HOMA beta-cell insulin secretion ( P = 0.01). Finally, in a multiple regression analysis, HOMA insulin resistance explained 35% of adiponectin variance, and SGA or birth weight status explained an additional 10% or 15% of adiponectin variance.
In general, early malnutrition and subsequent growth restriction appear to affect serum adiponectin levels. Little has been done at this point using specific rodent models to analyze adiponectin biology within the context of growth restriction. As in the case of leptin, we are still at the point of trying to figure out “chicken and egg” type issues. Whether this is a marker for, or cause of, impending morbidity remains to be seen.
Insulin-Like Growth Factor-1
Insulin-like growth factor-1 (IGF-1) is a polypeptide, whose homology resembles proinsulin. The bioavailability and subsequent actions of IGF-1 are regulated by binding proteins, which generally moderate the actions of IGF-1 by competing with IGF-1 receptors. Most tissues synthesize IGF-1, and IGF-1 homeostasis appears to have both systemic and paracrine implications, of which the latter are probably unappreciated. In general, IGF-1 levels in human fetuses increase from 18 to 40 weeks of gestation. IGF-1 and IGF-binding protein-3 (IGFBP-3) serum levels slowly increase before adolescence, steeply increase during puberty, and decrease thereafter. In contrast, IGFBP-1 serum levels gradually decline before adolescence such that the lowest levels are found in puberty.
IGF-1 and its associated binding proteins are intriguing players in the mechanisms relating early malnutrition to adult disease for several reasons. First, IGF-1 plays a key role in fetoplacental growth throughout gestation. Null mutations of IGF-1 in mice reduce fetal size by approximately 40%. In humans, IGF-1 gene deletion results in severe prenatal growth failure. Second, IGF-1 and IGFBP-3 mediate many of the anabolic and mitogenic actions of growth hormone in postnatal life. Short children have lower IGF-1 and IGFBP-3 levels compared with tall children. Third, a recent nested case-control study found that low IGF-1 and high IGFBP-3 levels in adulthood predicted increased risk for developing ischemic heart disease. Finally, IGF-1 regulates or moderates insulin sensitivity in adulthood. Hepatic IGF-1 is of vital importance for normal carbohydrate metabolism in both mice and humans. In mice, elimination of hepatic IGF-1 production increases serum levels of insulin without significantly affecting glucose elimination. In humans, recombinant IGF-1 is approximately 6% as potent as insulin in the production of hypoglycemia. Severe IGF-1 deficiency leads to insulin resistance, which can be reversed with recombinant IGF-1.
Although not as widely recognized, IGF-1 biology also affects glucose homeostasis before adulthood. Moran et al. hypothesized that the normal increase in insulin resistance that occurs concomitantly with puberty will be associated with changes in IGF-1, IGFBP-1, and IGFBP-3 levels. Studies were performed on 357 adolescents (mean age 13 ± 1.2 years). IGF-1 levels significantly correlated with insulin sensitivity in both boys ( P = 0.0006) and girls ( P = 0.02), although IGF-1 correlated significantly with fasting insulin levels only in girls. IGFBP-1 was negatively associated with insulin resistance in both genders, whereas IGFBP-3 was positively associated with insulin resistance only in boys. The findings of this study are provocative in that they can be interpreted to suggest that the IGF-1 axis either contributes or responds to the insulin resistance of puberty. Considering the data on adults, the latter appears to be more likely.
Multiple investigators have found that infants with IUGR or SGA status have lower fetal or cord IGF-1 concentrations compared with AGA infants. In the fetus, both genetic and environmental factors regulate fetal IGF-1 levels. The importance of the latter factor is supported by the observation that serum IGF-1 levels are similar in discordant monochorionic twins but significantly different in discordant dichorionic twins. This suggests that placental function and subsequent in utero substrate delivery may override genetic determinants of IGF-1 production. Furthermore, placental IGF-1 signaling appears altered in association with poor in utero fetal growth. In a recent study of 14 control and IUGR pregnancies, IUGR was found to reduce both IGF-1 receptor protein levels and IGF-1 signal transduction. In contrast, no differences were found in insulin receptor protein levels. The simple conclusion is that IGF-1 homeostasis is altered in both the fetus and the placenta in pregnancies complicated by poor fetal growth. The key question will be which comes first, the poor growth or the altered homeostasis, and the likely answer is that both are possible along the spectrum of the human environmental continuum.
The trend in altered IGF-1 homeostasis continues to be evident in the postnatal period. When comparing IUGR and AGA infants from birth through age 6 to 9 months, Ozkan et al. found that IUGR decreased IGF-1 serum levels and increased serum IGFBP-1, relative to the control infants. When IUGR babies were subdivided into those with catch-up growth and those without, the infants without catch-up growth had the lowest IGF-1 values ( P < 0.05). Finally, birth weight, postnatal weight, and postnatal height correlated directly with IGF-1 and IGFBP-3 levels, but not IGFBP-1 levels. Similarly, when Fattal-Valavski et al. determined IGF-1 levels in preadolescent IUGR children (mean age 6.5 ± 2.1 years; n = 57) versus control children (7.6 ± 2.8 years; n = 30), IGF-1 serum levels were found to be significantly decreased only in the non–catch-up IUGR group. Again, significant correlations between IGF-1 serum levels and both height and weight percentiles were observed in this study.
The interaction between birth weight and IGF homeostasis continues into early adolescence. Tenhola et al. investigated the relationship between serum IGF-1 and insulin sensitivity in 55 SGA and AGA age-matched children. SGA was defined in this study as birth weight, length, or ponderal index >2 SD below the respective mean for gestational age. Insulin sensitivity was determined by using HOMA. After adjusting for BMI, gender, and puberty, SGA increased serum IGF-1 concentrations ( P = 0.006). In multiple logistic regression analyses, HOMA insulin resistance predicted high serum IGF-1 levels in SGA children, but not in the AGA control group. Furthermore, SGA children in the highest IGF-1 quartile had higher BMIs ( P = 0.021), weight ( P = 0.038), and weight for height ( P = 0.040), and well as lower birth weights ( P = 0.077) versus SGA children in the lower IGF-1 quartiles.
In adulthood, the relationship becomes even more complicated, and again, genetic and environmental diversities enter the picture, as does the impact of puberty. Kajantie et al. investigated 421 subjects, who were singleton births between 1924 and 1933 at the Helsinki University Central Hospital. Detailed birth records were available for these subjects, including birth weight, length, head circumference, and gestational age, as well as measurements of height and weight between ages 7 and 15 years. The average birth weight of these subjects was 3504 ± 422 g (males) and 3342 ± 406 g (females). Fourteen of the subjects were classified as SGA and were >2 SD below the norm. When adjusted for gender, current age, and BMI, IGF-1 concentrations did not correlate with any of the measures obtained at birth. However, IGFBP-1 did positively correlate with birth weight ( P = 0.03) and ponderal index at birth ( P = 0.01). A positive correlation also existed between adult IGFBP-1 concentration and BMI at age 7 years. Furthermore, serum IGF-1 concentrations were positively associated with adult fasting glucose levels and both systolic and diastolic blood pressures.
Childhood malnutrition also appears to affect adult IGF-1 biology. Elias et al. utilized a group of 87 postmenopausal women who were exposed to the Dutch famine between ages 2 and 20 years. These women were divided into moderately exposed and severely exposed groups on the basis of weight loss, and 163 unexposed women of similar ages were used as controls. After adjusting for characteristics, such as BMI, waist/hip ratio, and cigarette-smoking habit, exposure to famine resulted in a significant increase in IGF-1 ( P = 0.038) and IGFBP-3 ( P = 0.045) serum levels.
The relationship between postnatal IUGR and IGF-1 levels is intriguing, although not completely clear, and if one looks at the trend in the studies by age, IGF-1 appears to gradually increase from the initial low levels in the SGA groups relative to the control groups. This may represent an overcompensation that permits catch-up growth, which is a good thing that helps cope with life as an adolescent, but the higher IGF-1 levels may have an as-yet-undefined pathologic or physiologic effect later in life.
Not all of these effects are necessarily bad, however. Early life events have been shown to affect hippocampal neurogenesis and increase age-related learning impairments in the rat. Interestingly, IGF-1 regulates the neurotropic response to aging through multiple mechanisms, including increasing mRNA levels of brain-derived neurotropic fats and stimulating hippocampal neurogenesis. Moreover, a recent study by Gunnell et al. on 547 white singleton boys and girls found that IGF-1 levels positively correlated with intelligence: For every 100 ng/mL increase in IGF-1, the intelligence quotient (IQ) increased by 3.18 points.
Furthermore, aging also leads to decreased brain levels of IGF-1. Although the study by Darnaudery et al. used rats, the findings of the study are worth noting. These authors exposed pregnant dams from day 14 of pregnancy until term (approximately 21.5 days) to confinement as well as exposure to bright light three times a day. After weaning at 21 days, the offspring of stressed and nonstressed dams were housed under identical conditions. At age 24 months, the progeny of the stressed dams were further divided into two groups: one group received vehicle (sodium chloride [NaCl]) into the right lateral ventricle for 21 days, the other group received IGF-1 for the same amount of time. Both groups were then exposed to a water maze task. Females whose dams had been stressed during pregnancy exhibited learning impairment in the water maze task. IGF-1 infusion restored the performance of spatial learning in the water maze such that it approximated that of control animals. One of the debates in the field is whether the adult consequences of early life events are a nonspecific response to early injury or is, in some way, teleologically protective, an evolutionary response, if you will. Such experiments suggest that the latter may be true, at least in some cases. There is a possibility that the benefits of early programming have not been appreciated, but the cost is evident, particularly as humans, as a species, have become more sedentary.
Investigators have used maternal malnutrition in the rat or bilateral uterine artery ligation of the pregnant rat to define this toll. This latter model is attractive in that it produces asymmetric IUGR through fetal hypoxia, acidosis, hypoglycemia, and decreased levels of branched-chain amino acids, characteristics shared with human infants suffering from uteroplacental insufficiency. IUGR rat pups in this model are 20% to 25% lighter than control pups, and birth weights are normally distributed within and between litters. Furthermore, litter size does not significantly differ between the control and IUGR groups. IUGR pups in this model develop early adipose dysregulation, insulin resistance, and adult-onset diabetes.
A variation of this model is unilateral uterine artery ligation, with pups from the unligated side acting as controls. Vileisis and D’Ercole used this model and found that fetal weight correlated with serum glucose ( P < 0.001), liver IGF-1 protein ( P < 0.001), and serum IGF-1 protein ( P < 0.001) levels. No correlation was evident for either serum insulin or lung IGF-1 protein. Interestingly, serum fetal glucose concentrations correlated positively with liver ( P < 0.001) and serum ( P < 0.002) IGF-1 protein levels, implicating fetal glucose delivery in the regulation of IGF-1 hepatic synthesis.
Using bilateral uterine artery ligation at day 17 of gestation, Houdijk et al. found that neither male nor female animals exhibited catch-up growth in terms of body weight as they reached adulthood, although the female IUGR rats did catch up to controls in terms of nose-anus length. As one might expect from the data on human children who do not exhibit significant catch-up growth, no differences were noted between control and IUGR IGF-1 levels at 100 days, regardless of gender. Interestingly, baseline growth hormone levels were significantly decreased ( P < 0.05), suggesting that IUGR in this model may impact tissue, particularly liver, responsiveness to growth hormones.
When maternal malnutrition is utilized, IGF-1 biology is also affected. In 1991 Bernstein et al. noted that 72 hours of maternal fasting decreased serum IGF-1 levels in term rat pups. Similarly, Woodall et al. investigated the effects of both protein and caloric malnutrition by restricting pregnant rat intake to 30% of an ad libitum control group. As expected, mean body weights were significantly decreased in the term IUGR fetuses versus controls, and this trend continued through the first 90 days of life. Furthermore, plasma IGF-1 levels were significantly decreased in the IUGR group from the end of gestation ( P < 0.01) through the first 9 days of life ( P < 0.05). The effects of IUGR on IGF-1 biology have also been well described in the bilateral uterine artery ligation model of IUGR. Growth-restricted pups have decreased serum IGF-1 at birth and in adolescence, at 21 days of life. Fu et al. also measured levels of the various IGF-1 mRNA transcripts in the liver. IGF-1 mRNA processing involves two different start sites involving exon 1 and exon 2, respectively, as well as different 3’ processing resulting in Ea and Eb variants. At birth, all mRNA transcripts were significantly reduced in both male and female IUGR pups relative to their gender-matched controls. By day 21 of life, after some catch-up growth, hepatic mRNA levels of transcripts produced from promoter 2, as well as the Eb variant, were still reduced in male and female rat pups compared with their gender-matched controls. An important component of the studies mentioned previously involves analysis of the epigenetic code of the IGF-1 gene and the effects of in utero nutritional stress on that epigenetic code. Fu et al. have begun to elucidate the role of epigenetics as a mechanism contributing to the complex relationship between in utero nutrition and IGF-1 biology.
Another potential mechanism through which IUGR may alter IGF-1 levels is zinc deficiency. Zinc is one of the most abundant divalent ions in living organisms and performs multiple functions secondary to its unique physiochemical properties. Two of zinc’s most important properties include (1) the ability to assume multiple coordination numbers and geometries, which make this ion stereochemically adaptable; and (2) its resistance to oxidation and reduction under physiologic conditions.
Zinc deficiency is a worldwide problem, and it is the second most important deficiency in infants. In Western societies, zinc deficiency is often associated with a low socioeconomic status. The effect of prenatal zinc deficiency is IUGR. The specificity of this effect is suggested by a double-blind study, which found that zinc supplementation significantly reduced the incidence of IUGR and improved most measured indices of fetal health. Zinc also plays a significant role in determining postnatal growth.
Interestingly, dietary intake of zinc significantly contributes to the regulation of IGF-1 levels in both humans and animals, and the liver is considered to be the major source of circulating IGF-1. Devine et al. found that zinc was the major determinant of IGF-1 concentrations ( P < 0.033) in postmenopausal women after 2 years of nutritional supplementation. Because of this study’s careful design, the authors were able to suggest that zinc intake influences IGF-1 concentrations, even in the face of adequate energy and protein intake. Similarly, although less well controlled because of the ages of the study subjects, several groups have found that zinc supplementation increases serum IGF-1 levels in children. Unfortunately, in the literature, human studies correlating zinc deprivation to depressed IGF-1 levels are lacking because of the multiple confounding factors, such as low protein and caloric intake, which complicate these studies. As a result, the animal literature is necessary to provide further insight.
A controversy exists in the zinc–IGF-1 literature on whether the effect of zinc deficiency upon IGF-1 hepatic expression is caused by dysregulation of growth hormone receptor pathway or the direct action of zinc on IGF-1. Two studies have investigated this controversy. The first study supplemented mice with zinc and found that although serum and hepatic IGF-1 mRNA levels increased in these animals, serum and hepatic growth hormone receptor mRNA levels remained unchanged. The second study demonstrated that hepatic IGF-1 synthesis requires zinc and that lack of the growth-promoting action of growth hormone in zinc-deficient animals resulted from a defect beyond growth hormone binding to its liver receptors. In other words, although growth hormone biology certainly plays a role in the regulation of hepatic IGF-1 expression, nutritional zinc also contributes in a fashion independent of the GH pathway. This is particularly true in the developing animal for the following reasons: (1) Neither growth hormone receptor mRNA nor specific GH binding is detectable in rat liver until 14 to 20 days of age ; (2) hepatic IGF-1 mRNA levels are not elevated until age 3 weeks in transgenic mice that are characterized by elevated levels of circulating GH ; and (3) liver-specific deletion of IGF-1 in mice reveals that IGF-1 regulates pituitary expression of growth hormone releasing factor, receptor, and secretagogue.
Although the above studies demonstrated causal links between zinc nutrition, IGF-1 biology, and growth, conflicting reports exist. Doherty et al. performed a double-blind, randomized intervention study of 141 children, ages 6 months to 3 years, from the Dhaka Shishu Children’s hospital in Bangladesh. Their weight for age was <60% of the National Centre for Health Statistics value. On the basis of the type of malnutrition (marasmus versus kwashiorkor), the presence of diarrhea, and the numbers of days after recruitment, each child was placed in a standardized feeding regimen, which included vitamin supplementation. Randomization further placed the children into one of three regimens: (1) 1.5 mg/kg elemental zinc for 15 days; (2) 6 mg/kg elemental zinc for 15 days; and (3) 6 mg/kg elemental zinc for 30 days. The diet increased ponderal catch-up growth rapidly, although linear growth was only moderately improved. At baseline, IGF-1 and IGFBP-3 were significantly lower than standard values for healthy, well-nourished European children. During the feeding protocol, IGF-1 and IGFBP-3 increased, reaching a plateau by 15 days. The only difference between the regimens was that IGF-1 was higher in the zinc regimen 1 at day 15 ( P = 0.04). As markers, IGF-1 and IGFBP-3 correlated best with ponderal growth at day 15 and with linear growth at day 90. The authors of this study noted that their inability to demonstrate a clearer effect of zinc in this particular study may be secondary to (1) the lower supplementation providing sufficient zinc or (2) the full nutritional supplementation masking the effects of the higher zinc dosing. The bottom line is that both prenatal and postnatal nutrition affects IGF-1 biology and alters postnatal phenotypic characteristics. On all levels—molecular, endocrinologic, and physiologic—there is a lot of information that has yet to be elicited ( Table 11.3 ). One of the more intriguing factors is the relationship between IGF-1 and early growth, particularly in how that relates to adult phenotype. Issues that need to be addressed include the effects of different macro- and micronutrients on both IGF-1 levels and sensitivity.
