Fetal programming occurs during embryonic and fetal development, a critical period in which tissues and organs are created. Insufficient nutrition during this time results in permanent alterations to certain structural and physiological metabolic functions of the fetus. British epidemiologist Barker first established the hypothesis, known as the "Barker hypothesis,” which states such programmed changes during this critical period predispose the fetus to certain postnatal diseases [1]. The critical period coincides with the timing of rapid cell differentiation [2]. Essentially, programming refers to the process of sustaining or affecting a stimulus or impairment that occurs at a crucial point in its development [34].

Rickets has long demonstrated that undernutrition in the critical early stages of life brings about a continuing change in structure [1]. A recent new doctrine suggests that fetal programming can affect diseases in adulthood. That is, the body's “memory” of undernutrition during the early stages of development translates into a pathology that determines future diseases [5]. This idea is based on animal studies that demonstrate how undernutrition in utero can alter blood pressure, cholesterol metabolism, insulin response to glucose, and other metabolic, endocrine, and immune functions important to human diseases [36]. This paper reviews evidence of the correlation between fetal undernutrition and diseases found in previous studies and considers the mechanism of fetal programming and the role of the placenta.

The nutritional status of the mother, which is an important factor that affects the programming of the body, involves factors such as maternal body composition, maternal dietary intake, blood flow to the uterus and placenta, and fetal genes. The fetus adapts to maternal malnutrition through changes in the production of fetal and placental hormones that regulate metabolism, redistribute blood flow and control growth (Fig. 1) [7]. The immediate metabolic response of the fetus to malnutrition is to consume its substrate to produce energy through catabolism [8]. Undernourishment of the fetus causes metabolic dependence on glucose to both decrease and increase the oxidation of other substrates, such as amino acids and lactic acid. Extended malnutrition results in delayed growth, reducing substrate use and lowering the metabolic rate, to improve fetal viability. Therefore, it can be hypothesized that the metabolic process of storing glucose continues into adulthood and that insulin resistance is the result of a similar process in which a decrease in the rate of oxidation in peripheral tissues that are insensitive to insulin creates increased insulin resistance. In late pregnancy, when tissues and organs are rapidly developing, any growth delays significantly affect organs and bring about disproportions in organ size. During the deceleration of growth, the fetus attempts to protect tissues, especially brain tissue, that are critical for immediate survival through redistribution of blood flow, thus resulting in a greater loss of liver and other abdominal visceral tissue [910]. In terms of endocrine changes, fetal insulin and insulin-like growth factor (IGF), which play a key role in growth control, are believed to respond rapidly to changes in fetal nutrition [11]. A decline in maternal food intake and the resulting drop in maternal IGF are likely to trigger decreases in fetal insulin, IGF and glucose levels. This drop reduces the transfer of amino acids and glucose from the mother to the fetus and ultimately slows the rate of fetal growth [12]. When the fetus has limited nutrient availability, anabolic growth control hormones, such as IGF-1 and insulin, decrease while the concentration levels of catabolic hormones like glucocorticoids, a major contributor to cell differentiation, increase [713].

A study traced 16,000 men and women born in Hertfordshire, England, from 1911 to 1930 from birth and showed that the mortality rate from coronary heart disease was twice as high in those with low birth weights than it was in the high birth weight group (Table 1) [1415]. Birth weight and adult blood pressure also showed an inverse correlation. Prominent alterations associated with fetal programming included insulin resistance, hypertension and increased serum low-density lipoprotein (LDL) cholesterol and fibrinogen concentrations, which are all characteristics of metabolic syndrome. The odds ratio of metabolic syndrome in relation to current body mass index was 18 times higher in the lowest birth weight group than in the highest birth weight group [16].

An association between low birth weight and coronary heart disease was confirmed in a cohort study of approximately 15,000 people born in Uppsala, Sweden [17]. The association of thin body types and type 2 diabetes was also verified with the observation that the prevalence rate of diabetes mellitus was 3 times higher in those with a low Ponderal Index (birth weight/length³) at birth. The Ponderal Index had stronger associations with diabetes than birth weight when compared to the prevalence rate of diabetes mellitus in people who had a low birth weight (Table 2) [18].

One recent retrospective cohort study followed 13,517 patients born in Helsinki, Finland, between 1924 and 1944. This study found that not only blood pressure but also the risk of hospitalization and death from coronary heart disease were influenced by birth weights. The mortality rate from coronary heart disease was 5 times greater for those with a low Ponderal Index at birth who were overweight by age 11 when compared to those with a high Ponderal Index at birth who were thin during childhood. Their study revealed another aspect of fetal programming associated with the pace of childhood growth. Children who showed the greatest growth changes and ultimately become obese in adulthood were at the highest risk for hypertension, cardiovascular disease, and type 2 diabetes [19]. That is, those who were born small and quickly became overweight while playing “catch-up” had the greatest risk.

The results of a recent systematic analysis of 34 studies, including 66,000 people from different populations, also supported the association between low birth weight and increased blood pressure in adulthood [20]. A 1-kg change in birth weight carried a systolic blood pressure difference of 3.5 mmHg. While it does not appear significant at the individual level, this is an important change in terms of a population's mean blood pressure.

Information on the effects of fetal malnutrition during different stages of pregnancy has been gathered by studying the Dutch famine of 1944 to 1945. During the relatively short-lived Dutch famine from September 1944 to May 1945, calories consumed per day dropped to 400 to 800 kcal from the previous 1,800 kcal average then rapidly bounced back to a normal diet of more than 2,000 kcal/day post-famine. A relationship between blood pressure and low birth weight was noted but was not significant. However, further analysis revealed that adult blood pressure was associated with a significant decrease in the protein/carbohydrate ratio in the maternal diet during specific trimesters of pregnancy. In glucose tolerance tests of adults who were gestating during the famine, a blood glucose increase in relation to birth weight was observed for 2 hours. This increase was most prominent in those who had been exposed to the famine in the last trimester of pregnancy, and the effects were also more pronounced in people who were obese in adulthood. These people showed the highest head circumference/birth weight ratio and the lowest birth weights, suggesting more severe growth limitations. The least affected people were those who were exposed to the famine in the first trimester of pregnancy, but they also showed fat metabolism disorders [21].

Other adult cohorts have demonstrated signs of impaired glucose tolerance (IGT) and insulin resistance. In the Hertfordshire group, 370 men aged 59 to 70 years were tested for glucose tolerance. The prevalence of type 2 diabetes and IGT was 40% for adults with a birth weight lower than 2,500 g. This figure dropped to 14% for adults with a birth weight greater than 4,300 g (Table 3) [22]. In these associations, birth weight was independent of social class or smoking and drinking status. However, lifestyle in adulthood also could influence the effects of the intrauterine fetal environment. For example, the prevalence of IGT was highest in the low birth weight adult obesity group. Insulin resistance testing on 103 men and women in the Preston region of the United Kingdom yielded similar results, showing more insulin resistance in those with similar body weights but a lower birth weight Ponderal Index [23].

In a study of participants living in Hertfordshire, England, Hales et al. [22] found a correlation between weight and the prevalence of diabetes. They discovered that birth weight and weight at age 1 were inversely correlated with the prevalence of diabetes in adulthood and that the risk of IGT and diabetes mellitus fell with increasing birth weight. Consequently, they proposed a thrifty phenotype hypothesis for the etiology of type 2 diabetes, which asserts that malnutrition during the fetal and neonatal period causes hypoplasia of the endocrine pancreas and triggers hyposecretion of insulin by the pancreatic beta cells [24]. If a fetus is exposed to undernutrition or malnutrition during rapid cell differentiation of the pancreatic beta cells due to maternal malnutrition or maternal and placental abnormalities, it will try to overcome these limitations through metabolic programming. This process will cause impaired development of the endocrine pancreas, which results in poor insulin secretion. This alteration cannot be reversed even with subsequent adequate nourishment; as the fetus grows into adulthood, diabetes mellitus often develops when insulin requirements increase (Fig. 2). If malnutrition persists after birth, the insufficient insulin secretion is not harmful, and these subjects will simply show increased insulin sensitivity. However, calorie intake that leads to obesity, in which nutritional intake is increased and energy consumption is decreased, produces a physiological imbalance and results in glucose intolerance. As brain development is prioritized, organs such as the liver, vessels and pancreatic tissues are subject to various permanent structural and functional disorders. Additionally, the timing and combination of the nutritional deficiency determine the profile of metabolic disorders, such as diabetes mellitus, hypertension, dyslipidemia, and insulin resistance. Many studies have confirmed that fetal growth restriction and subsequent obesity aggravate glucose intolerance and insulin resistance [25]. In the thrifty phenotype hypothesis, “catch-up” growth is a critical factor for the manifestation of potential diseases in adulthood [26].

A recent update of this theory suggests maternal hyperglycemia is a risk factor for fetal nutritional deficiency. It also holds that changes in vessel structure and function, which are fundamental to the growth and functional modification of organs, is a consequence of fetal undernutrition. Finally, it incorporates the role of the hypothalamus-pituitary gland-adrenal gland (HPA) axis and the sympathetic nervous system (Fig. 3) [25].

Changes in endometrial endocrine, nutritional and cardiovascular conditions affect the expression of the fetal genome so it can adapt during development (Fig. 4). Within this framework, the fetal genome determines the intrauterine growth potential, but actual growth is primarily determined by environmental effects, such as fetal nutrition and the hormonal environment [27]. Fetal programming can affect individual gene expression at any stage, from changes in molecular biological functions, such as receptor cell density or sensitivity, to permanent hormonal changes or even alterations in metabolism or responses to physiological stressors [28].

Nutrition delivered through the placenta is an especially important determinant of fetal growth because it allows the fetus to meet the growth potential determined by the underlying genotype. Recent reports have emphasized the impact of interactions between genotypes and the uterine environment. For example, siblings who share the same mother tend to have similar birth weights and show a 0.51 correlation, while siblings with the same father show a birth weight correlation of only 0.1. Other studies have indicated that cousins on the maternal side tend to have similar birth weights, whereas those on the paternal side do not. Babies born after oocyte donation have demonstrated that birth weight is strongly associated with the weight of the recipient rather than the weight of the donor [28]. One explanation for this may be related to the differential expression of the imprinted gene. Generally, imprinted genes expressed from the paternal system (IGF-2) stimulate growth, whereas imprinted genes (IGF-2 receptor) expressed from the maternal system act as growth inhibitors. The imprint of these genes is also controlled by epigenetic modifications, including DNA methylation and histone acetylation, under the control of environmental factors and nutrition [29].

Several epidemiological studies have reported on the relationship between placental weight and the placental weight/birth weight ratio in fetal programming, and there are ongoing animal and human studies on the overall role of the placenta in fetal programming. The placenta regulates maternal-to-fetal nutritional composition and supply. It is also the source of hormonal signals that affect maternal and fetal metabolism. Proper development of the placenta is critical to normal fetal development and plays an active role in programming the in utero fetal experience, which influences diseases that may appear in adulthood. The function of the placenta develops gradually in a series of closely organized developmental stages of pregnancy. This sequential division may lead to abnormal development of placental vessels or trophoblasts [53]. The timing of any abnormalities in development will be crucial to the resulting placental function and thus to fetal programming. This damage, which alters placental development, includes hypoxia and abnormal maternal nutrition. In response, the placenta changes its transporter expression and activity to maintain fetal growth, resulting in epigenetic regulation [54] or placental gene expression [55]. Hypoxia is physiological in organogenesis, and placental tissue normally resides in a relatively hypoxic environment [56]; however, delayed intrauterine growth and preeclampsia are associated with a more severe hypoxia of trophoblasts. Metabolic activity of the placental mitochondria causes oxidative stress even in a normal pregnancy and is further intensified by intrauterine growth retardation, diabetes and preeclampsia [57]. It can also cause nitrative stress, which is known to bring about covalent modification and altered protein activity. Hypoxia, oxidation, and nitrification stress all alter placental development, and the associated changes in placental function may be the general fundamental mechanism underlying fetal programming (Fig. 5) [55].

We observed genes that induced obesity and metabolic disorders in adults who were small for their gestational age due to nutritional deficiency (nutritional status) during pregnancy. In the study, when the offspring of a pregnant mouse given only 50% of the normal amount of food had its brain dissected at 3 weeks of age, a significant difference was observed compared to the control in the expressed levels of ubiquitin carboxylase terminal hydrolase L1 and secretin (Fig. 6) [58]. Methylenetetrahydrofolate dehydrogenase 1 and S-methyl transferase 1, enzymes related to single carbon metabolism in the livers of this animal model, were also lower than in the control, which was associated with high levels of blood homocysteine (Fig. 7). As blood homocysteine levels rise, there is an increased health risk of cardiovascular diseases and Alzheimer's disease [59]. In addition, wide spaces between hepatic cells were induced in mice on a low nutrition diet compared to normal mice (Fig. 8A), and a dramatic increase in gene expression levels of the mammalian target of rapamycin and sterol regulatory element-binding protein 1 were also observed (Fig. 8B and C). Thus, these results suggest that the liver of small-for-gestational-age newborns of mice that were exposed to a low nutrition diet showed dysfunctional liver signaling and lipid metabolism and had higher incidences of insulin resistance, diabetes and non-alcoholic fatty liver disease as adults [60]. Another piece of evidence suggests that when the pregnant mice were fed a low-protein diet, the liver of the offspring of the mother showed increased expression of glucocorticoids and paroxysmal proliferator-activated receptor, which is related to metabolic function. Such gene-specific changes may influence the occurrence of hypertension and hyperlipidemia later in life [61].

Cardiovascular disease and type 2 diabetes are 2 of the leading causes of death worldwide. During the last century, an increase in the migration of rural populations to urban areas has led to significant dietary changes and an increased degree of obesity. In evolutionary terms, fetal programming is not intended to help adults survive longer but rather exists to provide a mechanism for the fetus to adapt to conditions in the intrauterine environment. Fetal programming may predispose people to certain adult diseases and is an important matter of concern in the current age of low fertility and an aging population because it suggests that the aggressive treatment of low birth weights during pregnancy can prevent many adult diseases. However, the mechanism for the fetal programming of adulthood diseases remains a hypothesis, and there is insufficient understanding of the mechanism and also no method to determine if a fetus has been predisposed to certain adult diseases. Therefore, it is necessary to look beyond the epidemiological relevance to understand and study the mechanisms at the cellular and molecular level. Nutritional studies have focused on macro-nutrients in the past. Current investigations of micro-nutrients and also consideration of programming in the fetal environment due to factors other than nutrition are therefore required.

Epidemiological studies on possible fetal origins of adult diseases have only been conducted in adults; further studies on disease susceptibility in childhood and also from birth to adulthood will be required. Mutual strategic studies that include clinical and epidemiological studies as well as laboratory experiments should additionally be carried out. Currently, the health of the fetus is considered more important than ever due to the trends in low birth rates and late childbearing; understanding the role and mechanism of fetal programming, which has a negative impact on future health as well as on fetal health, has important academic value for finding intervention strategies.