Lifetime consequences of abnormal fetal pancreatic development

Abstract

There is ample evidence that an adverse intrauterine environment has harmful consequences for health in later life. Maternal diabetes and experimentally induced hyperglycaemia result in asymmetric overgrowth, which is associated with an increased insulin secretion and hyperplasia of the insulin-producing B-cells in the fetuses. In adult life, a reduced insulin secretion is found. In contrast, intrauterine growth restriction is associated with low insulin secretion and a delayed development of the insulin-producing B-cells. These perinatal alterations may induce a deficient adaptation of the endocrine pancreas and insulin resistance in later life. Intrauterine growth restriction in human pregnancy is mainly due to a reduced uteroplacental blood flow or to maternal undernutrition or malnutrition. However, intrauterine growth restriction can be present in severe diabetes complicated by vasculopathy and nephropathy. In animal models, intrauterine growth retardation can be obtained through pharmacological (streptozotocin), dietary (semi-starvation, low protein diet) or surgical (intrauterine artery ligation) manipulation of the maternal animal. The endocrine pancreas and more specifically the insulin-producing B-cells play an important role in the adaptation to an adverse intrauterine milieu and the consequences in later life. The long-term consequences of an unfavourable intrauterine environment are of major importance worldwide. Concerted efforts are needed to explore how these long-term effects can be prevented. This review will consist of two parts. In the first part, we discuss the long-term consequences in relation to the development of the fetal endocrine pancreas and fetal growth in the human; in the second part, we focus on animal models with disturbed fetal and pancreatic development and the consequences for later life.

Introduction

Fetal growth and development are determined primarily by the genetic potential of the fetus. However, the genetically predetermined growth and development can be influenced by environmental factors, which can exert stimulatory or inhibitory effects. For the supply of nutrients, the fetus depends on the nutritional status of the mother and on the capacity of the placenta to transport these nutrients to the fetus. The fetus also has its own growth factors, which influence growth and differentiation. Normal fetal growth is the result of an equilibrated interplay between these different factors. Any imbalance between these factors can result in fetal growth restriction (microsomia) or fetal overgrowth (macrosomia). Both abnormalities in fetal growth are related to an abnormal intrauterine environment.

Gestational diabetes in humans results in fetal macrosomia due to an increased supply of glucose and other nutrients (Pedersen, 1977; Freinkel, 1980). However, severe maternal diabetes complicated by vasculopathy and nephropathy results in intrauterine growth restriction (Van Assche et al. 1998). Yet, intrauterine growth retardation is mainly due to a reduced uteroplacental circulation, maternal undernutrition or malnutrition.

In animal models, ‘moderate’ diabetes or mild hyperglycaemia induces fetal hyperinsulinaemia whereas severe diabetes, semi-starvation, a low protein diet and a reduced uterine blood flow induce fetal hypoinsulinaemia (Van Assche et al. 1998).

The fetal endocrine pancreas, fetal growth and consequences for later life in the human

Gestational diabetes is characterised by an increased placental transport of glucose and other nutrients due to an increased availability at the maternal site, resulting in fetal and neonatal macrosomia (Pedersen, 1977; Freinkel, 1980). Islet hypertrophy and hyperplasia of the insulin-producing B-cells are typical features of fetuses and newborns of diabetic mothers (Dubreuil & Anderodias; 1920; Miller, 1946; Cardell, 1953; Jackson & Woolf, 1958; D'Agostino & Bahn, 1963; Naeye, 1965; Van Assche, 1970). In normal human pregnancy, insulin production and secretion by fetal B-cells are observed as early as 19 weeks of gestation (Van Assche et al. 1976). An increase in insulin secretion coincides with an increased number of insulin-producing B-cells. Levels of insulin-like growth factors (IGFs) also increase at this time. Both cord blood insulin and IGF concentrations are related to birth weight (Verhaeghe et al. 1993; Wiznitzer et al. 1998).

Two forms of fetal overgrowth, symmetric and asymmetric macrosomia, can be distinguished. Symmetric fetal overgrowth is genetically determined; asymmetric fetal overgrowth is induced in a disturbed intrauterine environment as in gestational diabetes, and is characterised by an enlarged thoracic and abdominal circumference in relation to the head circumference. B-cell hyperplasia and hyperinsulinaemia are characteristic features of asymmetric overgrowth (Van Assche, 1997a).

Pregnancy complicated with severe diabetes is associated with vasculopathy and reduced renal function. The elevated glucose concentrations in the mother, accompanying hyperglyacaemia in the fetus, lead to degranulation of the fetal B-cells, resulting in fetal hypoinsulinaemia (Van Assche et al. 1983). Indeed, the pancreas of a newborn of a badly controlled diabetic mother (blood glucose > 16.7 mmol l⁻¹) shows degranulation of the majority of B-cells with swollen mitochondria, extended rough endoplasmatic reticulum and very few granules (Van Assche et al. 1983).

The study of anencephalics provides additional interesting data. Anencephalics display normal development of the fetal endocrine pancreas. However, anencephalics born to diabetic mothers display the characteristic adaptations of the fetal pancreas to the diabetic intrauterine milieu only when a functional hypothalamic-hypophyseal (HH) axis is present. These anencephalics are macrosomic (Van Assche, 1970). A summary of the morphometric data is presented in Table 1.

Table 1.

Morphometric data of the human fetal endocrine pancreas

	Volume density endocrine tissue	Percentage of B cells
Normal controls (n = 40)	5.1 ± 1.6	40 ± 7.5
Maternal diabetes (n = 10)	12.9 ± 4.2**	63.8 ± 8.9**
Anencephalics without a functional HH system of a diabetic mother (n = 8)	5.0 ± 1.6	38 ± 9.7
Anencephalics with a functional HH system of a diabetic mother (n = 7)	11.6 ± 4.9**	59.2 ± 6.9**

Values are means ± S.D.

^**P < 0.001versus controls.

The consequences of diabetes during pregnancy are not confined to fetal and neonatal life. Several epidemiological data show that consequences extend to adult life and even to the next generation through the maternal line. Knowler and coworkers (1985) showed that the risk for diabetes is significantly higher in the offspring of mothers who have non-insulin-dependent diabetes. Additionally, 35 % of patients with gestational diabetes are offspring of diabetic mothers compared with only 5 % of normoglycaemic mothers, and gestational diabetes occurs more frequently in the offspring of diabetic mothers (35 %) than in offspring of diabetic fathers (7 %) (Martin et al. 1985). Most convincing are the studies on Pima Indians that have shown that, besides a genetic transmission of diabetes, the diabetic intrauterine milieu can also induce a diabetogenic tendency in the offspring. Non-insulin-dependent diabetes mellitus (NIDDM) is more frequent in children of mothers who had diabetes during pregnancy than in children of mothers who developed diabetes after pregnancy (45 % versus 1.4 % at age 20–24 years) (Pettitt et al. 1988).

An extensive study over several generations demonstrated a predominance of type II diabetes in great-grandmothers of infantile onset diabetes on the maternal side compared with the paternal side. In addition, a predominance of familial diabetes aggregation in first- and second-degree relatives was found on the maternal side compared with the paternal side. Moreover, a systematic prevention of hyperglycaemia and impaired glucose tolerance in pregnant women has significantly decreased the prevalence of diabetes mellitus in their children (Dörner et al. 1984, 1987).

Besides the transmission of diabetes to the next generation through the maternal line, women who were macrosomic at birth are at increased risk of developing a breast carcinoma in later life. The increased fetal insulin and IGF levels may explain this, as both fetal ‘growth factors’ have mitogenic effects on fetal breast tissue (Van Assche, 1997b).

Asymmetric fetal growth retardation is most commonly caused by a reduced uteroplacental circulation or by maternal malnutrition (Van Assche et al. 1977). It has been well documented that insulin action and insulin-like growth factor (IGF) are reduced in fetal growth retardation (Verhaeghe et al. 1993). Furthermore, we have demonstrated that the number of insulin-producing B-cells and the total amount of endocrine tissue are reduced in the pancreas of human growth retarded fetuses (Table 2; Van Assche et al. 1977). In contrast, one study found no developmental pancreatic abnormalities in human IUGR fetuses: the percentage of B-cells within the islets was normal in the IUGR group (Beringue et al. 2002).

Table 2.

Morphometric data of the human fetal endocrine pancreas in IUGR

	Gestational age (weeks)	Birth weight (g)	Volume density endocrine tissue	Percentage B cells
Normal controls (n = 20)	34.3 ± 1.6	2013 ± 793	4.8 ± 1.3	42 ± 6.3
IUGR (n = 20)	34.6 ± 1.9	1073 ± 394*	2.6 ± 0.6**	22.5 ± 3.1**

Values are means ± s.d.

^*P < 0.01

^**P < 0.001versus controls.

An expansion of epidemiological data shows that intrauterine growth restriction is related to diseases in later life. Cardiovascular diseases and insulin resistance are the most frequent alterations documented (Barker et al. 1993; Phillips et al. 1994; Barker, 1995; Leon et al. 1996). These studies introduced the term ‘fetal origin of adult diseases’. Intra-uterine growth restriction is also related to the occurrence of pre-eclampsia in later life (Hanssens et al. 1996). This is an important finding, since pre-eclampsia is characterised not only by renal and vascular alteration but also by increased insulin resistance.

Animal models

Maternal diabetes

Hyperglycaemia during rat pregnancy, induced by streptozotocin (Kervran et al. 1978; Mulay et al. 1983; Oh et al. 1988; Aerts et al. 1989, 1990a,b, 1997; Holemans et al. 1991a,b, 1993, 1999a; Ryan et al. 1995) or during a glucose infusion (Ktorza et al. 1990), has profound influences on fetal development and metabolism. Disturbances in glucose handling persist into adult life and into the second and third generation offspring.

In fetuses of diabetic rats, the development of the endocrine pancreas is enhanced by increased blood glucose concentrations, which results in an increased percentage of endocrine tissue because of hyperplasia and hypertrophy of the islets of Langerhans.

A moderate elevation (±20 %) of maternal glucose concentrations (Aerts & Van Assche, 1977; Aerts et al. 1989, 1990a, 1997) and consequently fetal glucose concentrations induces an increase in the number and biosynthetic activity of the fetal B-cells (Kervran et al. 1978; Aerts et al. 1990a). The insulin response to glucose stimulation, both in vivo and in vitro, is increased in the fetuses of mildly hyperglycaemic rats (Kervran et al. 1978). Circulating amino acid levels in the fetuses of mildly diabetic rats are severely decreased, as a result of increased amino acid uptake stimulated by fetal hyperinsulinaemia (Aerts et al. 1989).

After withdrawal of the hyperglycaemic stimulus at birth, the lactation period is a poor stimulus for further development of the endocrine pancreas, which ends up hypoplastic by the time of weaning. In the adult offspring of mildly diabetic rats, plasma amino acid levels remain low, including the neurotransmitters taurine, GABA and carnosine (Aerts et al. 1989, 2001).

A severe hyperglycaemia in the maternal rat results in hyperglycaemia and hypoinsulinaemia of the fetuses and fetal growth retardation (Aerts & Van Assche, 1977; Kervran et al. 1978; Devaskar et al. 1990). Fetal pancreatic weight is decreased but the percentage of endocrine tissue is increased (Aerts & Van Assche 1977). Fetal B-cells are overstimulated and they show degranulation on electron microscopy: a ‘B-cell exhaustion phenomenon’ is probably the cause of decreased pancreatic and plasma insulin concentrations (Aerts & Van Assche, 1977; Kervran et al. 1978). Similar structural and metabolic changes have been reported in fetuses of spontaneously diabetic BB rats (Verhaeghe et al. 1989) and in fetuses of the Goto-Kakisaki (GK) rat, a genetic model for NIDDM (Serradas et al. 1998; Miralles & Portha, 2001). The B-cells of severely hyperglycaemic fetuses are incapable of insulin secretion in vivo and in vitro; indeed the pancreas was insensitive to glucose (Kervran et al. 1978) and other secretagogues (Bihoreau et al. 1986a,b). Only arginine induced a sustained monophasic insulin secretory response, suggesting that the defect may concern stimulus-secretion coupling rather than the insulin-releasing mechanism (Bihoreau et al. 1986b).

Despite the abundance of nutrients, fetuses of severely diabetic rats are growth retarded (Aerts & Van Assche, 1977; Kervran et al. 1978; Devaskar et al. 1990). Hypoinsulinaemia and a reduced number of insulin receptors on target cells (Mulay et al. 1983) in fetuses of severely diabetic rats may lead to a reduction in fetal glucose uptake; a reduced fetal glucose uptake has been demonstrated in hypoinsulinaemic streptozotocin-injected fetal lambs (Phillips et al. 1991). In addition, levels of the glucose transporter GLUT1 are decreased in skeletal muscles of fetuses of diabetic rats (Schroeder et al. 1997). The growth of the fetal protein mass is suppressed in severely diabetic rat pregnancies: fetal protein synthesis is consistently lower than in controls whereas protein degradation increases sharply towards the end of gestation (Canavan & Goldspink, 1988). Circulating amino acid levels in the fetuses of severely diabetic mothers are decreased but in proportion to the maternal amino acid levels (Aerts et al. 1989; Copeland et al. 1990).

During the suckling period, the percentage of islet tissue decreases to values lower than normal. The B-cells can partly restore their secretory capacity by normalisation of B-cell granulation. Malnutrition and neglect of the newborn rats by their mothers is the cause of an increased postnatal death, and subnormal glucose and insulin levels in the survivors until the time of weaning. Body weight in these offspring remains significantly lower during their entire postnatal life, both during the suckling period and after weaning.

Offspring of diabetic rats have apparently recovered from the consequences of the abnormal intrauterine milieu and neonatal development by the time they reach adulthood The offspring of mildly diabetic rats have a normal mass of endocrine pancreatic tissue, with a normal distribution of the different islet-cell types (Aerts et al. 1990a). In vivo and in vitro insulin secretion, in response to glucose stimulation, is deficient. Glucose tolerance in the adult animal that suffered from perinatal hyperinsulinaemia is impaired (Kervran et al. 1978; Oh et al. 1988; Aerts et al. 1990a; Ktorza et al. 1990; Plagemann et al. 1992).

In adult offspring of severely diabetic rats the endocrine pancreatic mass exceeds control values, and this excess of islet mass is due to a high number of very small islets of Langerhans (Aerts et al. 1997), suggesting an increased contribution of B-cell neogenesis, rather than cell replication (Aerts et al. 1997). In vivo and in vitro, stimulation of B-cells results in increased insulin secretion (Aerts et al. 1990a), suggesting the existence of an insulin resistance in the offspring of severely diabetic rats. As revealed by euglycaemic hyperinsulinaemic clamp studies, these offspring are indeed markedly resistant to the action of insulin (Holemans et al. 1991a, 1993; Ryan et al. 1995). The decreased sensitivity to insulin is observed in the liver as well as the extrahepatic tissues (Holemans et al. 1991a) and peripheral glucose uptake is specifically reduced in skeletal muscles (Holemans et al. 1993). Exposure to severe maternal diabetes during fetal and neonatal life has profound consequences for cardiovascular function in the offspring. Despite normal systolic and diastolic blood pressure, there is evidence of pronounced bradycardia. The offspring of severely diabetic rats also show abnormalities of vascular function in vitro. Mesenteric arteries isolated from adult offspring of diabetic rats show a reduced relaxation in response to the endothelium-dependent dilators acetylcholine and bradykinin, suggestive of an impaired synthesis of endothelium-derived vasodilators. The defect in sensitivity and maximal relaxation to ACh is not observed in the presence of cyclooxygenase, NO synthase and guanylate synthase blockade, suggesting that PGI₂/NO-induced relaxation is responsible for the defect in endothelium-dependent relaxation in offspring of severely diabetic rats (Holemans et al. 1999a). The normal sensitivity to sodium nitroprusside also suggests that the defect does not arise from reduced sensitivity of the smooth muscle to NO, but from reduced NO synthesis. The constrictor response to noradrenaline is enhanced. The enhanced sensitivity, but similar maximal response, to noradrenaline is indicative of abnormal receptor-mediated tension development.

Endothelial dysfunction, similar to that which we report in offspring of diabetic rats, is not only observed in adult diabetic subjects (Johnstone et al. 1993) and animals (Taylor et al. 1992), but also in other conditions with high cardiovascular risk, particularly hypercholesterolaemia (Girerd et al. 1990; Goode et al. 1995). It is possible, therefore, that the intrauterine diabetic milieu has conferred upon the offspring of diabetic rats a predisposition to severe cardiovascular disorders in later life.

In experimental models, it is also possible to obtain information on the transmission of the diabetogenic effect to the next generation. Pregnant offspring of mildly diabetic pregnant rats have increased glucose levels. Consequently, fetuses are macrosomic and they display islet hypertrophy and hyperinsulinaemia. When these offspring become adult, they have an impaired glucose tolerance (Aerts et al. 1990b; Ktorza et al. 1990).

In addition, the offspring of severely diabetic rats develop signs of glucose intolerance during pregnancy; they have higher glucose and lower insulin levels than normal pregnant rats. These offspring, already insulin resistant before pregnancy, do not show the normal pregnancy-induced insulin resistance in the 2nd half of pregnancy: there is no further decrease in peripheral tissue sensitivity to insulin (Holemans et al. 1991b). Furthermore, vascular dysfunction shows only a slight deterioration during pregnancy (Holemans et al. 2000). Interestingly, the plasma concentration of the lipid peroxide 8-epi prostaglandin (PG)F_2α in pregnant offspring of diabetic rats was also raised above that of the pregnant offspring of control rats (Holemans et al. 2000). Possibly, the mild hyperglycaemia during pregnancy in these offspring (Holemans et al. 1991b), could contribute directly to enhanced free radical synthesis and lipid peroxidation (Hunt et al. 1988). Pregnancy seems to confer additional ‘stress’ and so unmask an already compromised balance between free radical synthesis and antioxidant status. We suggest that oxidative stress in the diabetic pregnant rat and her pregnant offspring could potentially play a role in fetal ‘programming’ and the transmission of a diabetogenic tendency to the next generation through permanent alteration of DNA and tissue damage in the developing fetus. This can be seen in conjunction with the fact that the disadvantaged fetal pancreas has lost its full adaptive possibility needed during pregnancy.

Semi-starvation of the rat

Maternal food-restriction (50 % of normal food intake) during pregnancy in the rat induces perinatal growth retardation (Holemans et al. 1996, 1997b, 1999b; Woodall et al. 1996a,b). The decreased availability of nutrients for transplacental transport and a decreased placental blood flow result in a decreased nutrient supply to the fetus (Rosso & Kava, 1980). The poor nutrient supply fails to stimulate the development of the fetal endocrine pancreas, B-cell mass and insulin content are significantly reduced (Garofano et al. 1997) and circulating insulin reaches only half of the normal value at term gestation (Holemans et al. 1996; Woodall et al. 1996a). At weaning, pups of rats food-restricted during pregnancy and lactation are severely growth retarded and their B-cell mass reaches about 30 % of the normal value (Garofano et al. 1998).

With maternal malnutrition, the combination of fetal hypoinsulinaemia and low substrate availability decreases fetal whole body glucose utilisation rates in the rat, mainly due to a decrease in glucose uptake by the fetal skeletal muscle and heart (Leturque et al. 1989). A decreased glucose transporter activity, protein and mRNA, has been reported in the lung of small-for-gestational-age fetuses in the rat (Simmons et al. 1992).

Adult female offspring of rats food-restricted during both pregnancy and lactation have a lower body weight from fetal life onwards (Holemans et al. 1996; Woodall et al. 1996a). Offspring of rats food-restricted during pregnancy and cross-fostered at birth with a control lactating rat achieve normal body weight by the time of weaning. However, from puberty onward growth is delayed in these offspring and at the age of 100 days their body weight is significantly lower than in the control group (Holemans et al. 1997a).

Non-fasting plasma glucose levels are increased in offspring that experience malnutrition during both fetal and neonatal periods, indicating that glucose tolerance has deteriorated in this group (Holemans et al. 1996).

Adult female rats subjected to malnutrition during both fetal and neonatal life, or to malnutrition during fetal life only, are resistant to the action of insulin. This resistance to insulin is the result of a decreased responsiveness of the liver. Insulin action at the peripheral tissues, however, remained normal (Holemans et al. 1996, 1997b). The pathogenesis of the lower responsiveness of the liver to exogenous insulin is unclear. Fetuses of food-restricted rats were hypoinsulinaemic, which may have affected the normal development of insulin receptors in the liver (Mulay et al. 1983). In addition, the increased glucagon-to-insulin ratio in the plasma of fetuses of food-restricted rats (Girard et al. 1977) influences the activity of glycolytic and gluconeogenic enzymes in the liver: the glucokinase activity was found to be lower and the phosphoenolpyruvate carboxykinase activity higher in offspring of rats fed a low-protein diet during gestation (Desai et al. 1997). The hypoinsulinaemia at adult age may itself be involved: insulin resistance in the liver was reported in mildly diabetic rats subjected to long-term malnutrition (Rao & Menon, 1993).

Using an intra-arterial pressure transducer, we found a normal systolic and diastolic blood pressure and heart rate in adult offspring of food-restricted rats (Holemans et al. 1999b). This would contrast, apparently, with a study by Woodall et al. (1996b), which has shown a small but significant increase in systolic blood pressure (5–8 mmHg) in much older (30 weeks) offspring of severely dietary restricted (70 %) rats. The degree of food reduction or an age-related development of raised blood pressure could potentially explain the difference in results, and the data might suggest that dietary restriction in early pregnancy could be an important determinant of offspring blood pressure. In contrast to the small rise in blood pressure observed by Woodall et al. (1996b) and the lack of rise in blood pressure observed by us (Holemans et al. 1999), protein restriction (6, 9 and 12 % casein) throughout rat pregnancy results in a very significant increase in systolic blood pressure in 9- to 21-week-old offspring (Langley & Jackson, 1994). This may indicate, therefore, that a balanced food restriction during pregnancy has a lesser effect on blood pressure in the offspring than the restriction of specific nutrients.

Determination of in vitro vascular function showed subtle changes in the small mesenteric arteries investigated. Small, but significant, reductions were observed in the dilator responses to acetylcholine and bradykinin, indicative of a reduction in synthesis of the endothelium-dependent vasodilator nitric oxide (NO). Relaxation in response to sodium nitroprusside, which evaluates responsiveness of the vascular smooth muscle to exogenous NO, was enhanced. Theoretically, this could reflect a compensatory response to tonic NO depletion. We suggest that one abnormality offsets the other, the reduction in NO synthesis or availability being paralleled by increased sensitivity, so resulting in no net change of NO-induced vasodilatation, and thus no effect on blood pressure (Holemans et al. 1999b).

Perinatal malnutrition affects not only the first generation offspring. Indeed, food-restriction during the perinatal period impairs the adaptation to a subsequent pregnancy. A reduced B-cell mass, insulin content and islet number are features of second generation fetuses, putting them at increased risk of developing diseases in later life (Blondeau et al. 2002).

Rats on protein-deficient diet

An isocaloric low protein diet (8 % versus 20 % protein content) during rat pregnancy also induces fetal growth retardation. This dietary protocol compromises nutrient delivery to the fetus by downregulation of specific amino acid transport proteins (Malandro et al. 1996). Fetuses of protein-restricted rats display a decreased islet size and B-cell proliferation capacity, a deficient insulin content and vascularisation of the islets of Langerhans (Snoeck et al. 1990; Petrik et al. 1999), and an increased apoptotic rate (Merezak et al. 2001). Insulin secretion from these islets after in vitro stimulation with amino acids is impaired (Dahri et al. 1991). At birth B-cell number and islet size remain decreased in pups of protein-restricted rats.

Offspring of protein-restricted rats weaned on a standard chow remain smaller then control offspring. Weaning the pups on the same low protein diet further impairs the growth of the offspring (Dahri et al. 1991). In adult female offspring of protein-restricted rats, the size of the pancreatic islets is increased whereas pancreatic insulin content is decreased. The vascularisation of the endocrine pancreas and its blood flow are impaired in these animals (Iglésias-Barreira et al. 1996). Only female offspring displayed decreased insulin levels at weaning and in adult life (Dahri et al. 1995). Glucose tolerance in the offspring was improved in both sexes, despite a reduced insulin release. This is compensated for by an improved insulin sensitivity (Shepherd et al. 1997; Holness et al. 1999). In addition, basal and insulin-stimulated glucose uptake rate are increased in isolated adipocytes of male offspring of protein-restricted rats (Ozanne et al. 1997) in association with a two-fold increase in insulin receptor number (Shepherd et al. 1997). However, in these offspring an accelerated deterioration in glucose tolerance was observed with advancing age (Desai et al. 1997). Gluconeogenesis from the liver is increased, with increased hepatic phosphenolpyruvate carboxykinase activity and decreased glucokinase activity (Desai et al. 1997). In contrast, in a study by Dahri and coworkers (1991) female offspring of rats on a similar low protein diet exhibited a significant impairment of glucose tolerance at younger ages.

Protein restriction throughout rat pregnancy results in a significantly increased systolic blood pressure in 9- to 21-week-old offspring as revealed by tail-cuff measurements (Langley & Jackson, 1994). However, low protein diets of differing composition used in different laboratories have yielded inconsistent data on the relationship between low birth weight and adult blood pressure (Langley-Evans, 2000). Using radiotelemetry, Tonkiss et al. (1998) found only a small baseline increase in diastolic blood pressure in offspring of protein-restricted rats, but an augmented elevation of both systolic and diastolic pressures after exposure to stress, suggesting that blood pressure measurements using the tail-cuff method might be biased by stress.

Pregnant offspring of protein-deprived rats develop gestational diabetes. Fasting plasma glucose levels are increased on day 18.5 of pregnancy and glucose tolerance is clearly impaired because of an inadequate adaptation of the endocrine pancreas to pregnancy (Dahri et al. 1995). Glucose utilisation during physiological hyperinsulinaemia is dampened in fast-twitch muscles (Holness & Sugden, 1996) and adipose tissue (Holness et al. 1999).

The penalty for the offspring must have been induced during intrauterine life, since recuperation on normal feeding after birth restores adult body weight, but does not improve glucose tolerance (Dahri et al. 1995).

Reduced uterine blood supply in the pregnant rat

Intra-uterine growth retardation (IUGR) can also be induced by reduction of the blood supply to the uterine horn using the Wigglesworth model of uterine artery ligation. IUGR fetuses displayed significantly lower glucose and insulin levels. In addition, C-peptide levels were reduced in the amniotic sac of the growth-retarded fetuses (De Prins et al. 1983).

At the level of the endocrine pancreas, islet cell number was normal but the animals with IUGR ended up with a reduced percentage of B-cells (De Prins & Van Assche, 1982). Uteroplacental insufficiency also decreases plasma levels of branched-chain amino acids in the fetuses (Ogata et al. 1986). An increased fetal oxidation of these amino acids may contribute to their decline in IUGR fetuses. Indeed, Hepatic branched-chain alpha-keto acid dehydrogenase, a mitochondrial enzyme performing the rate-limiting step in branched-chain amino acid oxidation, is significantly increased in IUGR fetuses (Kloesz et al. 2001). In 21-day-old IUGR fetuses, an alteration in the mitochondrial gene expression and function is found (Lane et al. 1998).

In the newborns, glucose levels remained decreased but insulin levels returned to normal. Yet the percentage of B-cells remained reduced in the IUGR pups (De Prins & Van Assche, 1982). Simmons and coworkers (2001), however, found normal B-cell mass, islet size and pancreatic weight in 1-week-old IUGR rats. At birth and 2 weeks after birth, growth restricted offspring have a reduced nephron number and impaired renal function (Merlet-Bernichou et al. 1994).

Postnatally, the growth of IUGR rats accelerated and by 26 weeks of age, IUGR rats were markedly obese (Simmons et al. 2001). In contrast, others reported no catch-up in body weight at any time in male and female offspring (Jansson & Lambert, 1999; Houdijk et al. 2000; Lane et al. 2001). Both sexes display a normal growth hormone (GH)/IGF-I secretion, suggesting the intrauterine modulation of tissue responsiveness to GH and IGF-I, leading to a decreased body weight in adulthood (Houdijk et al. 2000). Furthermore, an altered hepatic fatty acid metabolism, suggested by decreased mRNA expression of carnitine palmitoyltransferase and beta-oxidation trifunctional protein, may also contribute to the adult dyslipidaemia associated with IUGR (Lane et al. 2001).

At 7 weeks of age, IUGR rats still have normal B-cell mass, islet size and pancreatic weight. However, in 15-week-old offspring, the relative mass was 50 % that of controls (Simmons et al. 2001). The effects on glucose tolerance in adulthood were sex specific: only female offspring displayed fasting hyperglycaemia and impaired glucose tolerance at an adult age (Jansson & Lambert, 1999). In contrast, Simmons and coworkers (2001) described impaired glucose tolerance, insulin resistance and even overt diabetes in adult male IUGR rats.

In this model of IUGR there is no evidence that early growth restriction is associated with hypertension in later life: Jansson & Lambert (1999) found normal blood pressures in conscious freely moving IUGR rats using telemetry. Obviously, a reduction of blood supply to the uterus also affects fetal growth and alters the development of the endocrine pancreas.

Conclusions

There is ample evidence that an abnormal intrauterine environment can induce alterations in fetal metabolism with persisting consequences in late life. The occurrence of 43 cycles of cell division between fertilisation and birth, compared with only five after birth may explain the crucial role of fetal development.

The consequences are mostly seen at older ages, since the vitality of the organism is reduced and can no longer compensate for these alterations. It is known that pregnancy induces a stress on the endocrine pancreas and the total organism; this may explain why alterations are also observed in this period.

The maternally derived changes in the fetal plasma concentrations of nutrients (glucose, amino acids, fatty acids) seen in diabetes clearly influence the development and function of the fetal endocrine pancreas. Overstimulation of the insulin-producing B-cells may result in reduced insulin secretion in later life. However, other organs and functions may also have been altered, both directly and indirectly.

High glucose concentrations are known to promote B-cell replication, but the typical B-cell hyperplasia in fetuses of diabetic mothers only occurs if the fetus has a functioning hypothalamo-hypophyseal system (Van Assche, 1970), stressing the involvement of other hormonal regulatory mechanisms. Moreover, fetal hyperglycaemia induces fetal hyperinsulinaemia, which is known to damage the ventromedial part of the hypothalamus, which controls insulin secretion by modulating the tone on the nervus vagus (Plagemann et al. 1992).

It is also important to mention that normalisation of the diabetic intrauterine milieu in the last part of pregnancy protects B-cell function and the hypothalamus (Harder et al. 2001). Fetal hypoinsulinaemia, resulting from hypoplasia of the endocrine pancreas in maternal and fetal malnutrition or resulting from B-cell exhaustion in severe diabetes, might have an opposite effect. Fetal hypoinsulinaemia presents a lack of stimulus for the development of the insulin receptor system and this may affect insulin-sensitive organs in a different way. The effect on the fetal endocrine pancreas and other fetal systems depends on the metabolic condition of the mother: maternal diabetes associated with hyperglycaemia and maternal malnutrition or reduced uteroplacental circulation associated with hypoglycaemia most probably have diverse effects on fetal metabolic processes/gene expression. This may result in divergent defects in later life.