The Fetal Origins of Type 2 Diabetes Mellitus

Like other living creatures, human beings are plastic in early life and are molded by the environment. Although the growth of a fetus is influenced by its genes, studies in humans and animals suggest that this growth is usually limited by the environment, particularly by the nutrients and oxygen the fetus receives (1). There are many possible evolutionary advantages in the tendency of the body to remain plastic during development rather than have its development driven only by genetic instructions acquired at conception (2). Studies in animals show that a fetus may adapt to malnutrition by altering hormone production or the sensitivity of tissues to these hormones (3, 4). Among the hormones that regulate fetal growth, and hence the requirements for nutrients, insulin has a central role (5). The fetus can alter its metabolism (for example, by switching from glucose to amino acid oxidation) (4). It can redistribute its cardiac output to protect key organs, especially the brain. Slowing of growth is also an adaptive function because it reduces the requirements for substrate. Unlike physiologic adaptations in adulthood, adaptations during development tend to lead to permanent changes in the structure and function of the body. Experiments show that even minor modifications to the diet of pregnant animals may be followed by lifelong changes in the offspring in ways that can be related to human disease (for example, elevated blood pressure and altered glucose-insulin metabolism) (6). At a molecular level, such programmed changes may reflect the alteration of gene expression in utero by nutrient availability, acting directly on the cell or through hormonal signals. In the early 1990s, a study in Hertfordshire, England (7), was the first to show that persons with low birthweights had higher rates of type 2 diabetes later in life. The study was part of a research program investigating the fetal origins hypothesis, which states that coronary heart disease, stroke, type 2 diabetes, and hypertension originate from adaptations that the fetus makes to malnutrition; these adaptations cause permanent changes in the body's structure and physiology (3, 8). The Hertfordshire study entailed examination of men and women born from 1911 through 1930 whose size at birth and in infancy was recorded. Later studies in Europe and the United States (9, 10) also examined detailed birth records and confirmed the association between low birthweight and the development of type 2 diabetes or reduced glucose tolerance. Although birthweight serves as a marker of fetal growth and nutrition, it is crude. The same birthweight may be the result of many different paths of growth (11). More detailed measurements of body size at birth, when available, may give insights into the adaptations made by the fetus. For example, low-birthweight babies who are thin tend to be insulin resistant as children and adults and are therefore likely to develop type 2 diabetes. This suggests that thin babies responded to malnutrition in utero through endocrine and metabolic changes (3, 9, 12). In this issue, Rich-Edwards and colleagues (13) report the results of the largest study in this area to date. Women in the Nurses' Health Study were able to obtain their birthweights, and the incidence of type 2 diabetes among these women was ascertained by a mailed questionnaire. When data were adjusted for current body mass index and age, the risk for type 2 diabetes across the range of birthweights decreased twofold; this finding was similar to those in previous studies (7). A useful feature of the Nurses' Health Study is that it has collected information on childhood socioeconomic status and adult lifestyle factors. The associations with birthweight changed little after adjustment for these factors, providing further evidence that they reflect processes linked to intrauterine growth rather than the confounding effects of influences after birth. The Nurses' Health Study also throws further light on a question that is often raised. Because mothers who have gestational diabetes tend to deliver babies with high birthweight, who are themselves at increased risk for type 2 diabetes, why is not high birthweight associated with type 2 diabetes? The answer is that it is, but only in the offspring of mothers with gestational diabetes. In the Pima Indians, among whom diabetes in pregnancy is unusually common, the association between birthweight and type 2 diabetes is U-shaped. The association with high birthweight is abolished when the offspring of mothers with gestational diabetes are excluded. Similarly, in the Nurses' Health Study, exclusion of mothers who had a history of diabetes at any age strengthened the association of type 2 diabetes with low birthweight and led to a fourfold decrease in the risk for type 2 diabetes across the range of birthweights. In a study in Mysore, South India (14), men and women who had low birthweights tended to be insulin resistant; however, short, fat babies of heavier mothers tended to develop type 2 diabetes and were both insulin resistant and insulin deficient, with a low 30-minute insulin increment. These findings have led to a novel explanation for the current epidemic of type 2 diabetes in urban and migrant Indian populations. Widespread fetal malnutrition in the Indian population, whose average birthweight is only 2.9 kg, predisposes them to insulin resistance. After moving to an urban area, a person's level of physical activity tends to decrease. Young women who are no longer required to do agricultural work or walk long distances to fetch water and firewood may gain weight and therefore become more insulin resistant. These women would be less able to maintain glucose homeostasis during pregnancy, even at relatively low levels of obesity, and could become hyperglycemic (although not necessarily diabetic). Their children, who would be exposed to high circulating glucose concentrations in utero, may have impaired pancreatic -cell development and become insulin deficient. These ideas need to be confirmed by further studies in India. It is known, however, that maternal hyperglycemia, even within the normal range, is associated with increased fetal birthweight (macrosomia) without increased fetal length (15). In addition, women with gestational diabetes have babies who tend to be macrosomic and are at increased risk for type 2 diabetes as adults. During the past 10 years, some persons have argued that the associations between low birthweight and cardiovascular disease and type 2 diabetes in later life are the results of confounding variables (16, 17). Their voices are now seldom heard. Furthermore, as Rich-Edwards and colleagues (13) remark, results of experiments on animals directly support the fetal origins hypothesis. So how should we proceed? It seems that the strategy described by the National Institutes of Health (18) is to ignore the issue. Others, however, will recognize the possible importance of these findings in preventing type 2 diabetes. If the disease is a deferred consequence of successful adaptation in utero, then its primary prevention lies in protecting fetal development. Continuing epidemiologic studies will need to use better markers of fetal development than birthweight. Furthermore, we now know that the fetus can be programmed by nutritional influences that do not influence size at birth. Persons who were in utero during the Dutch famine in 1944 and 1945 now have reduced glucose tolerance and evidence of insulin resistance, but these changes are independent of the modest effect of famine on birthweight (19). The mechanisms by which undernutrition and retarded growth in utero lead to lifelong changes in glucose-insulin metabolism need to be explored. Studies of children in Europe, Jamaica, and India suggest that these mechanisms still operate today (20, 21). Further clinical and basic science research may therefore be a matter of urgency and cannot be deferred until we learn whether genetics offers solutions to the worldwide epidemic of type 2 diabetes.