The global epidemic of obesity, a multifactorial disease and key risk factor in the development of insulin resistance, type 2 diabetes, hypertension, and cardiovascular diseases is caused by both genetic and environmental factors. Its incidence has undergone a dramatic increase in recent decades resulting in a major impact on human morbidity, mortality, and quality of life, and becoming the main nutritional disorder in children and adolescents in developed countries (Friedman, 2003; Cripps et al., 2005). Overweight children have great risk of becoming overweight in adulthood (Von Kries et al., 1999); thus, the high prevalence of obesity together with the difficult success of its treatments makes it necessary to identify strategies to prevent obesity from the beginning of life. In this sense, there is a growing body of evidence suggesting that maternal nutritional conditions during critical stages of early life result in long term consequences on the future metabolic health of offspring, besides affecting their propensity to obesity in adulthood. In fact, both epidemiological studies in humans and intervention studies in animal models agree that the metabolic programming of energy balance begins with and can be modified by nutrition in the very early stages of development (Redman and Sweney, 1976; Lucas, 1991; Langley-Evans, 2006; Levin, 2006; Taylor and Poston, 2007; Wells, 2007; McMillen et al., 2008). Therefore, pregnancy and lactation are revealed as critical periods, where food restriction may lead to permanent adaptations with lasting effects on metabolic mechanisms in the offspring, thereby changing the propensity to obesity in adult life (McMillen et al., 2005; Martin-Gronert and Ozanne, 2006). However, evidence is increasing that the different outcomes of these adaptations on later health depend on the type, duration, period, and severity of the restriction and that they are, at least in part, gender specific.

Barker and collaborators in 1991 proposed the foetal origin of adult disease hypothesis (Barker and Osmond, 1986; Godfrey and Barker, 2001; Hales and Barker, 2001), This suggests that poor gestational nutrition leads to metabolic adaptations increasing not only the survival of the foetus but also the propensity to obesity and related alterations in adulthood, particularly under an obesogenic environment (Hales and Barker, 2001; Wells, 2007). Since then, considerable epidemiological evidence has been published showing an association between poor foetal growth and susceptibility to obesity and other diseases such as cardiovascular diseases, type 2 diabetes, and osteoporosis (Martorell et al., 2001; Langley-Evans, 2004). A number of animal studies have also related undernutrition during foetal life—due to maternal dietary restriction by calorie, protein or specific nutrient limitation—with lasting detrimental consequences on the homeostatic control of energy balance and obesity (Desai et al., 2007a; McMillen et al., 2008; Plagemann, 2008; Palou et al., 2010a).

On the other hand, unlike the number of studies found on foetal undernourishment, the evaluation of the long-term outcomes of early postnatal undernutrition during the suckling period, has received less attention. Studies carried out in humans have focused primarily on the consequences of energy restriction on the volume of milk production, but not on the potential effects on children's health (Dusdieker et al., 1994; Motil et al., 1995). In animal models, the positive or negative lasting effects of undernutrition during lactation are not clearly elucidated, as the conclusions are masked by the specific conditions tested: while severe food restriction during suckling—generally by means of increased litter size—has been related with growth retardation (Remmers et al., 2008a,b), moderate calorie or protein restrictions in dams during lactation have been associated with an improvement of insulin and leptin sensitivity in offspring later in life (Zambrano et al., 2006; Palou et al., 2010b).

All in all, there is increasing evidence of a close link between perinatal environment and adult metabolic health, affecting susceptibility to obesity and related metabolic alterations. Here we review the effects of maternal calorie restriction during gestation and lactation (see Figure 1) as well as the underlying mechanisms involved in the developmental programming of the processes controlling energy balance, by which a different propensity to metabolic disorders is early programmed, in order to identify prevention strategies from early stages of development.

The importance of maternal wellbeing on the future metabolic health of offspring is not new. There is strong evidence linking malnutrition during foetal life and a major propensity to acquire overweight later in life (Desai and Hales, 1997; Martin-Gronert and Ozanne, 2006). The Thrifty Phenotype hypothesis proposes that poor nutrition during early life leads to changes in tissue and organ functions, whereby the subject remains programmed to accumulate energy under a positive nutritional environment (Inoue et al., 2007; Wells, 2007; McMillen et al., 2008). Much evidence supports this idea, such as the emblematic example of the Dutch famine (Ravelli et al., 1976), where it was found that 19-year-old men conceived during the acute famine that devastated the western part of Holland during the last 6 months of World War II—whose mothers experienced poor nutrition in the first and second trimesters of their pregnancies—were more likely to be obese than their peers—whose mothers did not experience poor nutrition. Outcomes were different depending on the time of exposure. During the last trimester of pregnancy and the first months of life, exposure to energy restriction produced significantly lower obesity rates. Other epidemiological studies in humans have also related a low-birth weight to a greater risk of developing cardiovascular disease, hypertension, type 2 diabetes, and central obesity in adulthood (Barker et al., 1993; Martyn and Barker, 1994; Achard et al., 2006). In fact, human studies performed in the last decades suggest that 25–63% of adult diabetes, hypertension, and coronary heart disease could be attributed to low birth weight and/or to accelerated newborn-to-adolescent weight gain (Barker et al., 2002).

Nevertheless, most of the evidence associating foetal malnutrition and adult obesity are from animal models, particularly rats (Rasmussen, 1998). In this sense, severe maternal calorie restriction during pregnancy—up to 50%—which usually results in a reduced birth weight, has been associated with a greater incidence of obesity in adulthood, particularly when animals were exposed to an obesogenic environment (Thompson et al., 2007; McMillen et al., 2008). However, different outcomes have been described regarding subsequent body weight depending on the type, severity and timing of the restriction, and on the gender of the progeny (Symonds et al., 2004). Thus, male offspring of 50% calorie-restricted rats during the first 2 weeks of gestation displayed higher body weight and food intake than controls from the age of 5 weeks, whereas no significant differences were found in females (Jones and Friedman, 1982; Jones et al., 1984). On the other hand, Anguita and collaborators (Anguita et al., 1993)—using a comparable energy restriction treatment on pregnant rats—observed an impairment of normal weight in males and females despite no changes in food intake. However, the differences did not appear at the same age in both genders: whereas in males intrauterine malnutrition led to in impairment of normal weight gain and fat deposition from day 1 to day 53 of age, in females this resulted in pronounced fat accretion by the age of 53 days. In a different model of energy restriction, consisting of rat dams submitted to 50% food restriction from day 10 of pregnancy until delivery, pups showed intrauterine-growth restriction, but experienced very rapid catch-up growth after birth. Both males and females became obese at the age of 9 months (Desai et al., 2005). A more moderate gestational calorie restriction (30%) throughout gestation or during the first 2 weeks of pregnancy (Vickers et al., 2000, 2005) has also been related to hyperphagia, together with greater fat accumulation in offspring, particularly under a hyper-caloric diet, but without significant differences in body weight compared to controls. In addition, we have also described that lower calorie restriction (20%) during the first 12 days of gestation also has lasting effects in male and female offspring, programming both genders for increased food intake, but only resulting in greater body weight and body fat content in males (Palou et al., 2010a, 2012).

Thus, experimental studies in animals have contributed to providing a link between reduced birth weight in humans and subsequent health outcomes. However, the causes appear to be more complex than originally envisaged. Changes in maternal nutrition do not always result in altered foetal growth and can affect offspring health independently of size at birth (Harding et al., 2011). Moreover, maternal nutrition is important not only after conception. There is evidence showing that maternal undernutrition at distinct periods around conception has different effects on foetal growth. Periconception undernutrition in sheep has been shown to alter foetal metabolic and endocrine function and growth regulation and trajectory without intrauterine growth restriction (IUGR) (Rumball et al., 2009). Moreover, preconception undernutrition alone has been described to alter foetal growth responses to late gestation stressors (Rumball et al., 2009). These observations suggest that maternal nutrition is also important before conception.

In addition to maternal nutrition, other factors may also be of crucial importance concerning foetal development and subsequent health-related consequences. These factors include uterine and umbilical blood flows as well as placental transport and metabolism, since they may affect oxygen and nutrient supply to the foetus. Any interference in these steps may affect foetal nutrition and commonly results in growth restriction (Bloomfield and Harding, 1998). Different animal models of IUGR have been developed to investigate the mechanisms whereby this condition predisposes to subsequent development of metabolic diseases. These models are generally based upon uterine artery ligation, prenatal food or protein restriction, and exposure to specific hormones or to hypoxia during gestation. However, it must be pointed out that these conditions may differently influence later catch-up growth, which is believed to be a central component in the link between IUGR and the subsequent development of metabolic diseases in humans (Haugaard and Bauer, 2001; Shahkhalili et al., 2010).

Apart from the effects of overall maternal undernutrition during pregnancy, other studies have focused on the effects of foetal undernutrition by maternal protein restriction during gestation—generally performed by exposing dams to an isocaloric low protein diet (containing around 8% protein), compared with controls, fed 20% protein. Offspring born to protein restricted dams (8% protein) but suckled by normally fed dams were smaller at birth, but underwent rapid catch-up growth when suckled by normally fed mothers (Jennings et al., 1999). In adulthood these animals showed body weights similar to controls from 21 days of age, but had reduced longevity when fed ad libitum on standard chow diet. In another study, this very rapid catch-up postnatal growth after foetal protein restriction (8% protein) was also associated with increased body weight with respect to their controls at the age of 16 days (Coupe et al., 2010). In contrast, similar protein restriction (9% protein, compared with controls fed 18% protein) during pregnancy did not affect body weight of offspring of either males or females at the age of 1, 9, or 18 months; however, in senescence, these animals developed a metabolic syndrome-like phenotype (Erhuma et al., 2007). Neither were significant differences in body weight found in female offspring of moderate protein-restricted rats (30%) during gestation at the age of 24 weeks, although they did exhibit significantly greater food intake than their controls (Ikenasio-Thorpe et al., 2007).

Central and peripheral resistance to insulin and/or leptin signaling have been proposed as important mechanisms responsible for the deregulation of energy homeostasis, which may lead to obesity (Levin and Dunn-Meynell, 2002; Lustig et al., 2004; Esteghamati et al., 2009; Palou et al., 2010a) (see below). On the other hand, adiponectin, which has been described as an important adipokine related with obesity and insulin sensitivity, may also play a role in the early programming mechanisms involved in the effects of gestational undernutrition, as it exerts important effects on carbohydrate metabolism, improving glucose metabolism by increasing insulin sensitivity (Berg et al., 2001; Yamauchi et al., 2001). Furthermore, adiponectin has also antiatherogenic and anti-inflammatory properties (Fantuzzi, 2005). Unlike leptin, its levels are negatively correlated with body weight and fat content (Arita et al., 1999). Reduced adiponectin levels have been described in children born small for gestational age, in relation with their predisposition to develop insulin resistance and atherosclerosis (Cianfarani et al., 2004). Moreover, adiponectin levels in IUGR children were particularly low in those who showed postnatal catch-up growth, compared with the levels in IUGR children who remained small during childhood (Sancakli et al., 2008). These data suggest that adiponectin insufficiency may also play a role in the metabolic abnormalities observed in IUGR children and adults. However, there is still limited information available, and controversial data showing normal or higher adiponectin concentrations in IUGR insulin-resistant children have also been described (Evagelidou et al., 2007), suggesting that other factors and conditions may be influencing this association (Briana and Malamitsi-Puchner, 2009).

Another important aspect in the control of body weight and feeding behavior is the preference for certain foods. This is particularly significant in humans, where the obesity state is generally associated with greater food intake and appetite together with a preference for highly caloric food (Rissanen et al., 2002). A descriptive human study comparing the dietary preference of monozygotic twins who were discordant for obesity showed that the obese twin had a greater preference for fatty foods than the lean co-twin (Rissanen et al., 2002). Hence, programming food preferences in favor of high or low calorie food would be very important in the prevention of obesity development, even more so when high-fat foods are widespread in developed countries. In this sense, nutritional changes during perinatal development have been described to lead to differential programming of eating behavior, affecting not only the appetite but also food preferences, and being related with a propensity to obesity later in life. For instance, 20% maternal calorie restriction during the first half of gestation changed the dietary preferences of male offspring in adulthood; these animals showed a greater preference for fat-rich food than their controls. This could explain, at least in part, their increased food intake, under high-fat diet conditions, and their greater body weight. In contrast, females did not show differences in their dietary preferences with respect to their controls, and no significant differences in body weight were observed (Palou et al., 2010a). Thus, changes in food preferences could be one of the mechanisms that might explain the greater impact of moderate calorie restriction during gestation on subsequent body weight in male offspring (see Table 1). On the other hand, both male and female offspring of 50% protein-restricted-fed pregnant rats were programmed for a specific preference for energy-dense foodstuffs and hyperphagia, which was also related with greater body weight and body fat content in adulthood (Bellinger et al., 2004).

Thus, the different studies mentioned above mostly show a greater propensity to develop altered energy metabolism in adult life, in particular overweight and/or hyperphagia, after undernourishment during foetal development. This enhanced propensity to develop obesity is particularly clear in offspring of overall calorie-restricted dams, particularly when they undergo rapid catch-up postnatal growth, and is also exacerbated when animals are exposed to an obesogenic diet in adulthood. Mechanisms underlying the deregulation of food intake and energy balance, due to perinatal nutrition, could be related with hypothalamic alterations (Garcia et al., 2010) and a lower capacity to respond to insulin and leptin signaling (Palou et al., 2012), which are further commented below.

Lactation is one of the most important stages of development. Given that milk is the only source of food until the start of independent feeding (Babicky et al., 1972; Fiorotto et al., 1991), maternal nutrition during this period may be pivotal in programming the metabolic health of offspring. In this sense, the identification of nutritional and environmental conditions in dams that affect milk composition becomes very important in the prevention of future diseases in offspring.

However, unlike the contribution of the foetal period, the contribution of the early postnatal environment has received less attention, beyond the known beneficial effects of breastfeeding compared with infant formula against later pathologies, including obesity and related metabolic alterations (Von Kries et al., 1999). In addition, unlike the known effects of calorie restriction during foetal life increasing the risk of developing obesity, the lasting consequences of maternal undernutrition during the suckling period are uncertain. In fact, studies in humans are scarce, principally due to ethical aspects (Dewey, 1998a,b), and have been mainly focused on the amount of milk production by mothers rather than on the consequences in their infants (Coward et al., 1984; Prentice et al., 1984; Dusdieker et al., 1994). Lovelady and collaborators (Lovelady et al., 2000) described that moderate weight loss of 0.5 Kg per week—by restriction to 500 Kcal less than the control group and with moderate exercise—between 4 and 14 weeks postpartum in overweight lactating women who are exclusively breast-feeding, did not affect the normal weight gain of infants. A previous study by McCrory and collaborators (McCrory et al., 1999) also showed that an intervention for 11 days with 35% caloric restriction combined with exercise or not was not harmful for lactating women. In this study, no significant changes were detected in the amount and energy output of milk, frequency of nursing or weight gain of infants.

Regarding animal studies on intervention during lactation, most of these were performed by maternal restriction of overall calories or proteins or by direct food restriction of pups by increasing litter size. Results show different outcomes depending on the type and severity of the restriction. On the one hand, severe maternal calorie restriction (67%) throughout the lactation period (Boxwell et al., 1995) was shown to result in lower body weight of pups at weaning. Vicente and collaborators (Vicente et al., 2004) also showed that offspring of protein restricted dams (65% reduction) or energy restricted dams (to the same amount of calories eaten by the protein restricted group) during lactation had lower body weight at weaning. In adulthood the animals whose dams were fed the protein restricted diet weighed significantly less than the controls did. However, the animals whose dams were fed the energy restricted diet weighed more than the controls. Another study showed that female offspring of 50% calorie-restricted dams during the suckling period, although presenting a lower body weight than their controls, displayed increased food intake (Šefčíková and Mozeš, 2002). These conditions of undernutrition during the suckling period also affected feeding behavior in adulthood. Evaluation of feeding behavior under different experimental conditions revealed that these rats consumed significantly less food in a new (unfamiliar) environment than in the home environment (Šefčíková and Mozeš, 2002).

Comparable results to severe maternal restriction have been obtained by increasing litter size during lactation. Remmers and collaborators (Remmers et al., 2008a) showed that an increase of litter size from 10 to 20 pups per dam resulted in lower body weight and food intake of both males and female restricted pups in adulthood. However, these animals displayed reduced body dimensions at weaning and also showed incomplete catch-up growth size afterwards (Remmers et al., 2008a). Velkoska and collaborators (Velkoska et al., 2008) also demonstrated that pups raised in large litters (18 pups per litter) were lighter than their controls and remained significantly lighter throughout the study, with no evidence of catch-up growth.

On the other hand, unlike severe maternal food restriction during the suckling period, moderate maternal calorie restriction (30%) was described to result in lower body weight and fat content in both male and female offspring, but not to affect animals' body length (Palou et al., 2010b). The lower body weight was also associated to lower food intake, which was maintained from weaning to adulthood (Palou et al., 2010b). Moreover, these rats also showed a better response to high-fat diet and were protected against the increased body weight and hyperphagia exhibited by their controls under these dietary conditions (Palou et al., 2011b). Thus, these results suggest that moderate maternal calorie restriction during lactation protects against diet-induced obesity in rats, but does not impede catching up to normal body length and weight under standard diet conditions (Palou et al., 2010b, 2011b). Interestingly enough, these conditions of moderate calorie restriction during lactation also resulted in a different programming of food preferences; that is, female offspring of calorie-restricted lactating dams exhibited a lower preference for fat-rich food when exposed to high-fat diet feeding in comparison with their controls (Palou et al., 2010b).

The malnutrition produced by protein restriction of dams during the suckling period also seems to be related with a lower body weight of offspring, which is maintained in later life (De Souza Caldeira Filho and Moura, 2000; De Moura et al., 2007; Fagundes et al., 2007). However, there is controversial evidence regarding programming effects on food intake. For instance, 65% maternal protein restriction during the suckling period resulted in pups with lower food intake from weaning until the age of 53 days, but from that age onwards no significant differences were found (Teixeira et al., 2002; De Moura et al., 2007). In a similar experimental design, pups of protein-restricted dams during lactation did not eat less food than their controls when referred to body weight (Fagundes et al., 2007). On the other hand, offspring of protein-free nourished dams during suckling—when fed with a normal protein diet at weaning—did not show lower food intake compared to controls. Interestingly, when they were exposed to a protein-free diet in adulthood, the reduction in food intake was more marked than controls (Moura et al., 2002). In another study, both male and female offspring of 50% protein-restricted dams during suckling showed lower body weight and also lower total food intake at 100 days of age (Zambrano et al., 2006). Notably, and contrary to what is described concerning protein restriction during gestation, it is remarkable that reduced growth during lactation due to maternal protein restriction results in a permanent reduction in size and an increase in longevity (Jennings et al., 1999). It is not known whether the effects on longevity are directly related with particular features of lactation or with growth reduction. Lim and collaborators (Lim et al., 2006) also described that 56.5% maternal protein restriction during both gestation and lactation led to lower body weight of offspring which persisted into adulthood. However, body weight of offspring of 60% protein-restricted dams during pregnancy and lactation remained lower than controls until the fifth week, but later on differences in body weight between groups disappeared (Coupe et al., 2009b). These rats also displayed hyperphagia with respect to their controls (Coupe et al., 2009b).

All in all, animal studies show that dietary restriction during suckling by maternal restriction or by increasing litter size results in a different programming of subsequent body weight and food intake, which may change propensity to obesity. Thus, although moderate restrictions may have protective effects against overweight and improve metabolic health of offspring (Palou et al., 2010b, 2011b), severe deprivation may lead to detrimental effects on the normal development of animals (Boxwell et al., 1995; Šefčíková and Mozeš, 2002; Remmers et al., 2008a) (see Table 1). However, it must be stressed that these conclusions are based on animal studies, since there are no clear data available from humans. Therefore, we should be cautious in extrapolating these results to humans due to differences between species, particularly in the different timing of birth as regards organ maturity and metabolic development.

The mechanisms underpinning maternal nutrition and programming of obesity risk in offspring are not clearly elucidated. One potential mechanism could be through permanent structural changes in key organs. In fact there are several examples in animal models showing lasting effects of the perinatal environment on brain development (Morgan and Naismith, 1982; Plagemann et al., 2000a; Garcia et al., 2010). Circulating hormones influence multiple aspects of hypothalamic development and play a role in directing formation of neural circuits. Thus, a deficiency in some of these essential factors during a critical period of development will have permanent structural consequences. Among these, leptin has been proposed as an important neurotropic factor during postnatal development (Bouret et al., 2004; Yura et al., 2005). Its deficiency during this period may have later consequences in body weight control (Palou and Pico, 2009; Pico et al., 2011). However, the brain is not the only tissue that is sensitive to perinatal conditions and responsible for the perinatal programming of body weight control. Recent data indicate that the development of other structures, such as peripheral innervations of the sympathetic nervous system (SNS), may also be altered during development and have an important impact on future adiposity and metabolism of the offspring (Garcia et al., 2011).

An alternative mechanism by which environmental factors at critical periods of development could have long-term phenotypic consequences may involve epigenetic modifications (Heijmans et al., 2009). In this section, recent progress in these fields is summarized (see Figure 1).

Given the current obesity epidemic there is an urgent need for interventional strategies, even from early stages of life. In this sense, foetal and early postnatal time periods are important for determining future risk of obesity, type 2 diabetes and other features of the metabolic syndrome, regardless of genetic or further environmental factors. Animal and epidemiological studies provide clear evidence for the relationship between maternal nutrition—including not only over-nutrition, but also under-nutrition—and the risk of obesity in offspring, although with different programming effects depending on the magnitude of the dietary alteration and the specific period. Thus, they represent invaluable tools in the investigation of mechanisms underpinning this linkage as well as in providing a potential therapeutic target to stem the growing epidemic of obesity.