Introduction

Adiposity has become a worldwide concern particularly in the developed world. Studies have shown that adiposity tends to run in families so that children of obese parents are at a greater risk of developing obesity throughout their life . This familiar relationship of adiposity suggests a heritable element in the aetiology of the condition, which may reflect a synergistic interrelationship between the genetic make-up of the individual and environmental factors, namely the role of nature and nurture in the development of the adipose state . A detailed understanding of the interrelationship between genetic and environmental factors is essential to enable the design of effective preventive interventions .

Genetic factors

In a proportionately few number of individuals, syndromic adiposity develops directly or indirectly as a secondary consequence of particular genetic defects. Such defects include trisomy 21 acting through the altered production of obesity-related hormones , and conditions associated with errors in genomic imprinting such as Prader–Willi syndrome ( chromosome 15q partial deletion ), Albright’s hereditary osteodystrophy ( Gs alpha subunit deficiency ), Cohen syndrome ( VPS13B gene mutation ), Borjeson–Forssman–Lehmann syndrome ( PHF6 gene mutation ), Beckwith–Wiedemann syndrome ( chromosome 11p15 rearrangements ) and the range of Bradet–Biedl syndromes. These are relatively rare genetic conditions that play a minimal role in the aetiology and increasing prevalence of adiposity in the general population. A further limited role is played by the specific monogenic disorders identified by focused candidate-gene linkage studies as rare causes of hyperphagia and adiposity beginning in early childhood. These include mutations in the leptin-encoding gene , the leptin receptor gene and proopiomelanocortin. Association genetic testing has further identified an association with a number of low-frequency coding variants of the melanocortin-4 receptor (MC4R) . However, these gene linkage study techniques have not been as useful as one would have wished in determining common forms of obesity and/or diabetes .

Better success at identifying associations between DNA sequence variants and adiposity has been achieved by the use of systematic large-scale population-based genome-wide association studies examining BMI value ranges. These studies have identified at least 30 loci that appear to affect the risk of developing adiposity with the strongest association being with variants within the fat-mass and the obesity-related ( FTO ) gene localised on chromosome 16 . Other adiposity gene associations, including TMEM18 , KCTD15 , GNPDA2 , SH2B1 , MTCH2 and NEGR1 , have suggested an interrelationship between adiposity and hypothalamic function disorders ( Table 1 ). In addition, a number of genetic determinants, such as the B3-adrenergic receptor gene , have been identified as being contributory towards the development of weight gain, insulin resistance and the earlier susceptibility to develop type 2 diabetes mellitus (T2DM) . This gene polymorphism seems to predominantly express itself in adipose tissue to regulate lipid metabolism and thermogenesis , while a further interrelationship with inflammation has also been described . The reported data, however, remain generally conflicting .

Genetic factors

In a proportionately few number of individuals, syndromic adiposity develops directly or indirectly as a secondary consequence of particular genetic defects. Such defects include trisomy 21 acting through the altered production of obesity-related hormones , and conditions associated with errors in genomic imprinting such as Prader–Willi syndrome ( chromosome 15q partial deletion ), Albright’s hereditary osteodystrophy ( Gs alpha subunit deficiency ), Cohen syndrome ( VPS13B gene mutation ), Borjeson–Forssman–Lehmann syndrome ( PHF6 gene mutation ), Beckwith–Wiedemann syndrome ( chromosome 11p15 rearrangements ) and the range of Bradet–Biedl syndromes. These are relatively rare genetic conditions that play a minimal role in the aetiology and increasing prevalence of adiposity in the general population. A further limited role is played by the specific monogenic disorders identified by focused candidate-gene linkage studies as rare causes of hyperphagia and adiposity beginning in early childhood. These include mutations in the leptin-encoding gene , the leptin receptor gene and proopiomelanocortin. Association genetic testing has further identified an association with a number of low-frequency coding variants of the melanocortin-4 receptor (MC4R) . However, these gene linkage study techniques have not been as useful as one would have wished in determining common forms of obesity and/or diabetes .

Better success at identifying associations between DNA sequence variants and adiposity has been achieved by the use of systematic large-scale population-based genome-wide association studies examining BMI value ranges. These studies have identified at least 30 loci that appear to affect the risk of developing adiposity with the strongest association being with variants within the fat-mass and the obesity-related ( FTO ) gene localised on chromosome 16 . Other adiposity gene associations, including TMEM18 , KCTD15 , GNPDA2 , SH2B1 , MTCH2 and NEGR1 , have suggested an interrelationship between adiposity and hypothalamic function disorders ( Table 1 ). In addition, a number of genetic determinants, such as the B3-adrenergic receptor gene , have been identified as being contributory towards the development of weight gain, insulin resistance and the earlier susceptibility to develop type 2 diabetes mellitus (T2DM) . This gene polymorphism seems to predominantly express itself in adipose tissue to regulate lipid metabolism and thermogenesis , while a further interrelationship with inflammation has also been described . The reported data, however, remain generally conflicting .

Metabolic programming

The familiar, apparently hereditable, predisposition towards the development of adiposity thus appears to be generally not a simple direct relationship to a genetic predisposition. The nutritional environment presented to the developing foetus in utero has been shown to influence the risks of developing later-life adiposity and other co-morbidities of the metabolic syndrome through the process of metabolic programming . Both over- and undernutrition have now been shown to be implicated in contributing towards the foetal origins of adult-onset disease.

The relationship between intrauterine overnutrition and the predisposition to neonatal adiposity has long been known. The link between macrosomia and the diabetic state during pregnancy was originally made in the 1920s but was formalized by Jørgen Pedersen in 1967 who formulated the hypothesis concerning how the foetus is affected by maternal hyperglycaemia – now known as the ‘Pedersen Hypothesis’ . Epidemiological studies have further shown that a high birth weight is related not simply to the glycaemic status of the mother during pregnancy but also to the mother’s body mass and weight gain during pregnancy irrespective of the carbohydrate metabolism status of the woman .

The relationship between overeating and excessive birth weight was first proposed by Hugo Ehrenfest in 1919 . It has been demonstrated that, in a central Mediterranean island population, the mean foetal birth weight increased proportionately with both an increase in maternal BMI and increasing weight gain during pregnancy. A macrosomia rate >10% was associated with an antenatal weight gain of ≥15 kg in women with a BMI 25–29 kg/m ² , and with a weight gain of ≥10 kg in women with a BMI >30 kg/m ² . Women with a BMI <25 kg/m ² showed a macrosomia rate of <10% irrespective of the degree of weight gain during pregnancy . A high birth weight has further been correlated with the risks of childhood adiposity. In the Pima Indian population, children born to women suffering from gestational diabetes and hence presumably overfed in utero had higher risks of childhood obesity than those born to women with a normal carbohydrate metabolic profile. This difference persisted after correction for other influencing factors . A similar interrelationship between macrosomia and the eventual increased predisposition to childhood adiposity has been demonstrated in 5- and 9-year-old central Mediterranean island children . A relationship has further been shown between macrosomia and an increased risk of the eventual development of gestational diabetes .

Animal-model transgenerational studies have indicated that the predisposition to macrosomia is mediated by a number of different mechanisms involving foetal pancreatic and hypothalamic development. Second- and third-generation pups of streptozotocin (STZ)-injected pregnant diabetic rats were shown to have developed pancreatic islet cell hyperplasia and beta cell generation in response to the hypernutritional intrauterine state increasing the volume density of endocrine tissue but decreasing the volume density of granulated beta cells . Similar observations relating islet cell hypertrophy and beta cell hyperplasia in foetuses of diabetic human mothers had long been made , the degree being positively related to maternal glycaemia and body mass index . Further maldevelopmental processes occurring in the foetuses of diabetic mothers include the malprogramming of hypothalamic orexigenic and anorexigenic neurons, possibly mediated through a raised foetal and neonatal leptin, predisposing to the later development of adiposity . These long-term consequences were similarly noted in the offspring of pregnant rats who were exposed to a hyperglycaemic episode before birth or who were overfed during the early post-partum period . An increased predisposition for childhood obesity has been seen in 5-year-old children who had been bottle-fed when compared to breastfed children .

The evidence strongly confirms that, through Pederson’s mechanism hypothesis, hypernutrition during intrauterine and neonatal life predisposes towards pancreatic and hypothalamic changes, which predispose the individual to develop childhood and adulthood adiposity and related co-morbidities. At the other end of the spectrum, undernutrition has also been shown to be correlated to childhood and adult adiposity. In 1990, DJ Barker proposed an interrelationship between birth weight and the development of cardiovascular disease in individuals aged 50 years. This led to the introduction of the concept that “the womb may be more important than the home” . In animal models, maternal malnutrition has been shown to be related to foetal growth restriction as a consequence of nutrient deficiency and alteration of the physiological adjustments that occur during pregnancy . The Dutch famine of 1944–1945 provided epidemiological evidence for an interrelationship between nutritional intrauterine starvation and growth restriction of the foetus . These individuals were shown to have a greater predisposition to develop chronic adult-onset disease. The predisposition to adiposity seems to be more likely if the developing foetus is nutritionally deprived during the first half of pregnancy rather than later . A later follow-up study of the same cohort of individuals now aged 50 years showed that those females exposed to intrauterine starvation had higher BMI levels that non-exposed woman. There were no significant differences between famine-exposed or non-exposed men . A relationship between low-birth-weight infants and a lean maternal BMI and poor weight gain during pregnancy has been confirmed during normal metabolic conditions in a central Mediterranean island population. A low-birth-weight (<2.5 kg) term infant rate >5% was noted in women with a BMI of 20–24 kg/m ² and a pregnancy weight gain of <5 kg, and in women with a BMI <20 kg/m ² and a pregnancy weight gain of <10 kg . Low-birth-weight infants in this population have been shown to be more likely to predispose to the eventual development of childhood adiposity at 9 years of age and to gestational diabetes .

A number of epidemiological and animal studies have now confirmed the association between antenatal foetal and early postnatal nutrition with insulin resistance, the development of adiposity and the various co-morbidities of the metabolic syndrome . The mechanisms behind ‘metabolic programming’ or ‘developmental plasticity’ have been linked to direct developmental permanent loss in the structural and functional capacity within the β-cells of the pancreas. Animal studies have confirmed that antenatal undernutrition predisposes the foetus to developmental reduction in pancreatic cell number and β-cell mass leading to a decrease in pancreatic insulin concentration and circulating foetal insulin levels . In addition, changes in the hepatocytes, skeletal muscle cells and adipocytes contribute towards the development of peripheral insulin resistance increasing gluconeogenesis in the liver and decreased insulin inhibition of lipolysis in the adipocyte. There is further an increasing body of literature to support the link between antenatal and early postnatal nutritional deprivation with the methylation of cytosine residues, thus altering genomically imprinted genes, such as IGF2, H19 and IGF2R, through covalent modifications of DNA without altering the nucleotide sequence of DNA . These developmental and epigenetic changes serve to allow the nutritionally deprived foetus and young infant to better deal with future starvation – the thrifty phenotype hypothesis as coined by Hales and Barker in 1992. However, they also predispose the individual towards developing obesity and type 2 diabetes in later life if the environmental and life circumstances alter from a state of starvation to a state of plenty .

Intrauterine and postnatal starvation from whatever cause will set up a vicious cycle whereby the intrauterine nutritionally deprived individuals will predispose to a thrifty phenotype physiology and in times of adequate or excessive food availability predispose towards the subsequent development of metabolic abnormalities. These will in turn affect their subsequent offspring during intrauterine life predisposing to the biological changes related to the Pedersen cycle. This in turn will again predispose this generation to higher rates of abnormal metabolic physiology . This cycle of metabolic modulation has been demonstrated in Maltese women born during the siege condition of the Second World War when for 3 years the population was severely deprived of food. It was demonstrated that women born during the siege period delivered infants with statistically higher mean birth weights than their counterparts born before and after the siege years . In this central Mediterranean island population, the immediate post-Second World War period was characterised by a generally low socio-economic environment, which predisposed the population to malnutrition. During 1949, a survey of 15,468 children aged 5–14 years showed that while at 5 years of age the mean body weight was just below the 50th percentile when compared to the British percentile weight standards; children aged 14 years had a mean body weight at about the 10th centile . The socio-economic situation of the Maltese population changed dramatically in the subsequent decades. This has led to significant nutritional and lifestyle changes that have contributed towards the previously thrifty phenotype population tending towards adiposity and thus influencing the subsequent generation with an adverse metabolic physiology through the Pedersen cycle. Contrasting with the situation prevalent in 1949, the mean BMI value for 9-year-old Maltese children born in 2001 was above the World Health Organization (WHO) adiposity cut-off levels so that the 47.9% of boys and 39.5% of Maltese girls aged 9 years were defined as adipose .