Metabolic Imprinting and Programming
METABOLIC IMPRINTING AND PROGRAMMING
METABOLIC IMPRINTING AND PROGRAMMING. It has long been recognized that environmental influences during early development have profound and long-lasting effects on humans and other animals. Metabolic imprinting describes a subset of such effects, comprising subtle but persistent responses to prenatal and early postnatal nutrition. This article will describe how the term "metabolic imprinting" was conceptualized and will provide some examples of putative metabolic imprinting phenomena to illustrate the important roles they may play in human health.
A growing body of human epidemiologic data suggests that the quality and quantity of nutrients available during prenatal and early postnatal development can affect susceptibility to various adult-onset chronic diseases, including cardiovascular disease, type-II diabetes, and hypertension. Some of the earliest indications of such relationships were gleaned from ecological data showing a regional correlation between infant mortality and mortality from cardiovascular disease several decades later. These data led researchers to postulate that a poor environment that causes high infant mortality could also impair the development of surviving infants, increasing their susceptibility to cardiovascular disease in adulthood.
David Barker's group in the United Kingdom bolstered the ecological data by collecting retrospective data on individuals. By linking obstetric records from the early 1900s with mortality records from several decades later, his group found that individual birth weight was related inversely to adult cardiovascular disease mortality. In numerous populations in industrialized countries, similar relations were later found between birth weight and morbidity and mortality associated with cardiovascular disease, coronary heart disease, and type-II diabetes. Viewing birth weight as a proxy for the quality of the prenatal environment, these relations were interpreted to indicate that prenatal factors (such as maternal nutrition) could "program" the development of adult chronic disease. Extensive epidemiologic research investigating this so-called "fetal origins" hypothesis is underway. However, the inherent weaknesses of human epidemiologic research in this field, including the long period of follow-up from exposure to outcome and inability to accurately adjust for various potential confounding variables, limit our ability to draw causal inferences from epidemiologic relations alone.
The Concept of Metabolic Imprinting
It remains unknown whether developmental responses to fetal and early postnatal nutrition are major determinants of chronic disease susceptibility in humans. Understanding the biological mechanisms underlying such phenomena in appropriate animal models should help to gauge their potential importance to human health. The term "metabolic imprinting" was proposed to provide a framework for the investigation of such biological mechanisms. Metabolic imprinting encompasses adaptive responses to specific nutritional conditions early in life that are characterized by 1) susceptibility limited to a critical ontogenic period early in development (the critical window), 2) a persistent effect lasting into adulthood, 3) a specific and measurable outcome, and 4) a dose-response relationship between exposure and outcome (Waterland and Garza, 1999). "Programming" has been used as a more general term to describe effects that occur when "an early stimulus or insult, operating at a critical or sensitive period, results in a permanent or long-term change in the structure or function of the organism" (Lucas, 1991; p. 39). However, imprinting, first used to describe the setting of animal attachment behavior based on early experience, more effectively conveys the important characteristics of the phenomena under consideration. As Konrad Lorenz noted, ethological imprinting occurs only during "a quite definite period" in the animal's life and the imprinted behavior "cannot be 'forgotten' " (Lorenz, pp. 126 and 127). The remainder of this section will discuss the most salient features of metabolic imprinting.
Adaptive responses are those that contribute to survivability of the organism. Limiting the scope of metabolic imprinting to adaptive responses excludes persistent effects of severe nutritional deficiencies or exposure to toxic levels of specific nutrients during critical stages of development. For example, it is well established that neural tube defects can be caused by maternal folate deficiency during early embryonic development. Also, fetal exposure to pharmacological levels of vitamin A can cause teratogenesis. Clearly, both of these examples represent persistent effects of early nutrition, but neither is adaptive. It may seem illogical to propose that adaptive responses putatively characterized as metabolic imprinting could contribute to adult chronic disease susceptibility. However, adaptive responses in early development are not necessarily beneficial throughout an individual's life. A metabolic response that increases prenatal or early postnatal survivability in one environment may prove detrimental to the individual in a different environment or at a later ontogenic stage.
The "critical window" criterion in the definition of metabolic imprinting limits consideration to effects resulting from nutritional perturbation of developmental pathways. Mammalian development occurs in stages. At each stage, specific processes must be completed in a limited time frame to enable progression to the next stage. A given nutritional stimulus may have diverse effects, depending on the developmental processes underway during the stage at which the stimulus is applied. Therefore, each specific metabolic imprinting phenomenon should have a finite critical window.
To illustrate, the fertilized mammalian egg proceeds through a series of reductive cell divisions (cleavage) to the formation of the three germ layers (gastrulation), to the establishment of organ rudiments (organogenesis), to the stage of histogenesis, when cellular differentiation results in the formation of specialized cell types and, lastly, to metabolic differentiation, during which the specialized functions of the different cell types proceed toward functional maturation. Organogenesis requires inductive interactions between adjacent germ layers. Hence, during organogenesis, localized concentrations of diverse nutrients or their metabolites could alter organ structural development by interacting with these signaling systems. As another example, the wave of rapid cell proliferation that follows cellular differentiation in various organs puts a high demand on the nutrients necessary for the synthesis of cellular components. Transient deficiencies (or excesses) of these food-derived precursors during limited periods of rapid cell proliferation could result in permanent alterations in cell number and, hence, metabolic activity.
Examples of Putative Metabolic Imprinting Phenomena
It should be emphasized that metabolic imprinting remains a theoretical construct intended to provide a framework for investigations into the biological mechanisms linking early nutrition and adult chronic disease susceptibility. Research is underway to identify candidate phenomena that satisfy the criteria of metabolic imprinting. This discussion therefore focuses on candidate phenomena that appear most consistent with metabolic imprinting.
Retrospective studies linking birth weight to adult chronic disease outcomes represent by far the most numerous class of epidemiologic data in support of metabolic imprinting. Studies of many different populations have found that individual birth weight is related to adult risk of coronary heart disease, hypertension, type-II diabetes, stroke, overweight, and other disorders. In most cases, there is a simple inverse correlation, with adult disease risk increasing as birth weight decreases. Some adult outcomes, such as type-II diabetes, appear to follow a "U-shaped" relation with birth weight, with adult risk increasing at very low and very high birth weights.
Proponents of the fetal origins hypothesis conclude that these examples demonstrate that nutritional status during critical periods of fetal life influences the development of diverse organ systems, leading to effects on adult chronic disease risk. As discussed above, however, there are many limitations to the interpretation of such long-term retrospective epidemiologic studies. Moreover, there are several reasons that birth weight is not an ideal proxy for fetal nutritional status. For example, maternal nutritional status during pregnancy is just one of many factors that determine infant birth weight. Also, an individual's genetic makeup could influence both fetal growth and adult susceptibility to a specific disease, leading to an association between birth weight and adult chronic disease that is not mediated by fetal nutritional environment. Conversely, it is likely that fetal nutrition can affect metabolic development without affecting birth weight. Hence, a misplaced focus on birth weight as the major early-life predictor of adult chronic disease susceptibility could result in a gross underestimation of the importance of metabolic imprinting to human health.
For these reasons, studies of human populations with documented exposure to extreme nutritional conditions in early life are an important source of support for the metabolic imprinting hypothesis. The best characterized of such populations comprises individuals exposed perinatally to starvation during the Dutch famine of 1944–1945. In 1976 Gian-Paolo Ravelli and co-workers published an analysis of Dutch military draft induction records of 300,000 young men and found that, compared to those born in non–famine-stricken control areas, men who were exposed to famine conditions at some time during the first six months of fetal development experienced an 80 percent higher prevalence of overweight. In contrast, those exposed to famine during the last trimester of gestation and/or the first five postnatal months experienced a 40 percent lower prevalence of overweight. Perinatal famine exposure was later related to adult glucose tolerance in several hundred individuals from the Dutch famine cohort. During standardized glucose tolerance tests, plasma glucose concentrations were higher in adults who had been exposed prenatally to famine than in individuals born before the famine. Because the timing, severity, and geographical extent of nutritional deprivation caused by the Dutch famine were well documented, these studies are not weakened by the methodological issues inherent to observational studies predicated on birth weight.
In the only large-scale experimental trial designed to investigate metabolic imprinting–like phenomena in humans, Alan Lucas and co-workers randomized preterm infants to receive standard infant formula, a special preterm formula, or banked breast milk during the first month of postnatal life. At the start of the trial, none of the feeding alternatives were clearly superior for preterm infants, allowing ethical randomization of early diet in this cohort of several hundred individuals. In this ongoing study, long-term follow-up has already shown that subtle differences in early postnatal diet affect cognitive development and bone mineralization in childhood. Conversely, childhood growth, body composition, and blood pressure were not associated with the early dietary exposure.
Controlled experimental investigations in animal models have demonstrated the biological plausibility of metabolic imprinting. Such studies have confirmed that subtle variation in prenatal and also early postnatal nutrition can affect adult outcomes including glucose-stimulated insulin secretion, blood pressure, body weight, organ structure, and lipid metabolism. Guided by the framework of metabolic imprinting, future animal studies of candidate phenomena should identify the tissues responsible for their effect persistence and characterize their critical windows to generate testable hypotheses about the effects of early nutrition on specific developmental processes.
Subtle nutritional perturbation of developmental pathways may have an important impact on human health. The immediate challenge for researchers in this field is to elucidate the specific mechanisms by which nutrition influences biological development in appropriate animal models. Doing so will suggest specific areas of focus for future research into the biological links between early nutrition and adult metabolism in humans. Clearly, the potential importance of gaining this information is great, especially in populations that continue to be at high risk for marginal nutritional status in early life.
Barker, David J. P. Mothers, Babies and Disease in Later Life. London: BMJ Publishing Group, 1994. Summarizes much of the data leading to development of the fetal origins hypothesis.
Kalthoff, Klaus. Analysis of Biological Development. New York: McGraw-Hill, Inc., 1996.
Lucas, Alan. "Programming by Early Nutrition: An Experimental Approach." The Journal of Nutrition 128 (1998): 401S–406S. Describes the ongoing study of preterm infants randomized to receive normal formula, preterm formula or human milk during the first postnatal month.
Lucas, Alan. "Programming by Early Nutrition in Man." In The Childhood Environment and Adult Disease. CIBA Foundation Symposium 156. Chichester, U.K. and New York: Wiley, 1991.
Lorenz, Konrad. Studies in Human and Animal Behaviour, volume I, translated by Robert Martin. London: Methuen, 1970.
Ravelli, Gian-Paolo, Zena A. Stein, and Mervyn W. Susser. "Obesity in Young Men after Famine Exposure in Utero and Early Infancy." The New England Journal of Medicine 295 (1976): 349–353. Historic study describing the Dutch famine of 1944–1945 and demonstrating relationships between early famine exposure and adult overweight.
Waterland, Robert A., and Cutberto Garza. "Potential Mechanisms of Metabolic Imprinting That Lead to Chronic Disease." The American Journal of Clinical Nutrition 69 (1999): 179–197. Reviews data in support of metabolic imprinting and discusses potential underlying biological mechanisms.
Robert A. Waterland