Baron-Cohen et al. [7] discussed a number of amniocentesis-based correlational studies, which indicate a positive correlation between the mental rotation rate and fetal testosterone and an inverse relation between fetal testosterone levels and language comprehension. The authors used stored samples of amniotic fluid and performed longitudinal studies in these ‘amniocentized children’ over 48 months after delivery. At 12 months of life they correlated fetal testosterone (fT) and the ability to make eye contact as a marker for social development in 71 children. Girls made significantly more eye contact than did boys. Eye contact decreased with increasing levels of fT. The authors further examined language development between 18 and 24 months of age in 87 toddlers and found that girls had a significantly higher vocabulary size than did boys and that there was a significantly inverse relationship between fT and vocabulary size for the children as a group but not within either sex group.
Maternal and Fetal Nutrition and Programming of Adult Disease
The concept of a relationship between fetal undernutrition and subsequent development of adult disease was introduced two decades ago by Barker et al. [10]. The ‘Barker theory’ or the ‘fetal origin hypothesis’ states that low birth weight reflects intrauterine malnutrition which programs the fetus for later cardiovascular disease, diabetes, and hypertension. Today it is generally accepted that the maternal diet affects fetal health and that maternal nutrition and specifically undernutrition with a low protein diet may be associated with subsequent development of obesity, cardiovascular disease, and hypertension. Leeson et al. [11] showed that low birth weight is associated with impaired endothelial function in childhood, which is an important risk factor for subsequent cardiovascular disease. The discovery that endothelial dysfunction in adults may be due to prenatal life events is of great importance since it indicates that atherosclerosis may also have its roots in prenatal events which are not genetically controlled. The authors calculated the relation between birth weight and endothelium-dependent flow-mediated dilatation (FMD) and concluded that a 1-kg difference in birth weight is equivalent to 4.5 cigarette pack-years as far as risk for subsequent vascular disease is concerned. This means that the adverse impact of reduced birth weight on vascular function across the birth weight range is as great as the effect of smoking. The correlation between reduced caloric intake during pregnancy and adverse fetal outcomes has been studied in populations in times of war and natural catastrophes. From November 1944 to May 1945, the German blockade of Western Netherlands led to what has been termed ‘The Dutch Famine or Hungerwinter’. Rations went as low as 400 calories a day, equivalent to 25% of the minimal daily requirement. At the Academic Medical Center in Amsterdam, the Dutch Famine Cohort Study was initiated; it investigates the outcomes of men and women who were born in the Wilhelmina Gasthuis in Amsterdam between November 1943 and February 1947 (www.hongerwinter.nl). Thanks to well-preserved documentation, the researchers were able to strongly corroborate the ‘Barker theory’. They found that individuals who were exposed in utero to malnutrition weighed 200-300 g less than controls and suffered later in life from a variety of diseases, including diabetes, cardiovascular disease, higher LDL/HDL ratios, obesity, coagulation disorders, hypertension, pulmonary disease, and renal disease. Ijzerman et al. [12] demonstrated in dizygotic, and hence genetically unlike twins that low birth weight was associated with insulin resistance and lower HDL levels, indicating a pathology independent of genetic factors. Intrauterine malnutrition seems to affect male and fetal fetuses differently. Sardina et al. [13] examined the central anorexigenic effects of insulin in rats. Female but not male adult rats who were exposed to intrauterine undernutrition developed in adulthood adiposity, impairment of hypothalamic insulin signaling, and loss of insulin-induced hypophagia.
It is becoming increasingly apparent that many forms of adult disease may have their roots in fetal programming and in early childhood. Obesity is one example: obesity has become a major public health problem affecting the majority of adults in the USA. Extremes of birth weight, i.e. low birth weight and increased birth weight, may cause future obesity and metabolic syndrome. It has been shown that prenatal androgen exposure of female mice causes impaired insulin secretion both as an organizational effect and as an activational effect which is based on defective islet cell function. Chronic diseases like osteoporosis, mood disorders, and psychiatric syndromes as well as polycystic ovary syndrome have been related to fetal programming. Other adult diseases such as cardiovascular disease, a variety of metabolic and endocrine pathologies, diabetes, and various other diseases may also be due to inadequate determination of set points during fetal programming. There is now preliminary evidence that eating disorders, which are more common in females that in males, are organized by low levels of prenatal testosterone and then activated by increasing estrogen levels during puberty in girls.
The Impact of Maternal Prenatal Stress on the Developing Embryo and Fetus
Four decades ago, Ward [14] reported that male rats which were exposed prenatally to stress showed low levels of male copulatory behavior but increased female lordotic sexual responses. There is a growing body of prospective data which links antenatal maternal stress with impaired, gender-specific emotional and cognitive development of offspring in childhood. Prenatal stress in pregnant animals can cause permanent impairment of neurodevelopment in offspring, including shorter attention spans, anxiety, and impaired cognitive function. Weinstock [15] showed in rats that maternal stress increases corticosteroid levels in the fetal brain and decreases testosterone levels and aromatase activity in male fetuses. Moreover, noradrenalin activity increased and dopamine activity decreased in male fetuses and levels became similar to those of female fetuses as a result of maternal stress. Male offspring had more learning deficits while female rats exhibited anxiety, depression, and an increased response of the pituitary-adrenal axis to stress. Based on rodent studies, Weinstock [15] further pointed out that maternal stress during critical periods of fetal brain development may exert long-term effects on the hypothalamus-pituitary-adrenal axis, including an impaired feedback mechanism and impairment of the circadian release of corticosteroids. This effect seems to be time sensitive and impinges on the fetal brain only during specific time windows. It has been shown that the increased incidence of schizophrenia in men whose mothers were subject to severe stress, as during the May invasion of the Netherlands in 1940, was restricted to the second trimester of pregnancy and did not occur during the first trimester of pregnancy [16]. Kashan et al. [17] studied a cohort of 1.3 million births in Denmark, which occurred between 1973 and 1995. They reported that death of a relative in the first but not in the second and third trimester of pregnancy was associated with an increased gender-specific risk for schizophrenia in offspring (adjusted RR 1.67; 95% CI 1.02-2.73). The incidence per 100,000 person-years was higher in males (55.4%) than in females (41.3%). The death of a relative occurring up to 6 months before conception had no such effect. Malaspina et al. [18] reported on a cohort of 88,829 births which occurred in Jerusalem between 1964 and 1976. They linked the data to Israels’ Psychiatric Registry and found an increased risk (RR = 2.3; 95% CI 1.1-4.7) for schizophrenia in individuals whose mothers were subject to substantial stress during the second month of pregnancy. In contrast to the report of Kashan et al. [17], they reported that the risk was higher in females (RR = 4.3; 95% CI 1.7-10.7) than in males (RR = 1.2; 95% CI 0.4-3.8). This sex preference for females is more in accord with other reports.