Adipocyte regulation in early life

There are at least two key factors in early life that are critical in determining adipose tissue function in the newborn: the amount of fat present and its ability to generate heat through the brown adipose tissue-specific uncoupling protein (UCP; UCP1)⁽Reference Symonds, Mostyn and Pearce^5, Reference Cannon and Nedergaard⁶⁾. There are considerable differences in UCP1 both within different large-mammal species⁽Reference Symonds and Lomax⁷⁾ and between small and large mammals⁽Reference Symonds, Stephenson and Gardner⁸⁾. For example, brown fat is not present in the newborn pig, which is dependent on shivering thermogenesis in order to maintain body temperature after birth⁽Reference Lossec, Lebreton and Hulin⁹⁾. In newborn sheep, which are primarily dependent on non-shivering thermogenesis in order to prevent newborn hypothermia, brown fat is rapidly recruited after birth⁽Reference Clarke, Heasman and Firth¹⁰⁾. In contrast, both mice and rats are precocial newborns and depend on huddling with their littermates within a nest in order to maintain body temperature⁽Reference Cannon, Connoley, Obregon, Kunzel and Jesen¹¹⁾. Maturation of non-shivering thermogenesis at this time is therefore coincident with maturation of the hypothalamic–pituitary axis⁽Reference Girard, Ferre and Pegorier¹²⁾, a process that occurs over the final third of gestation in large mammals, including man, as compared with the lactational period in rodents⁽Reference Symonds and Budge¹³⁾. Ultimately, these substantial differences in maturity at birth reflect the very different environments inhabited by each of the species, in conjunction with the profound differences in brain maturation that appear to have an over-riding influence on adipose tissue development⁽Reference Symonds, Mostyn and Pearce⁵⁾.

The ontogeny of brown fat development

The most prominent example of differences in fat function and distribution resides within the ontogeny and location of brown fat between large and small mammals; primarily reflecting the very different methods of adapting to the extrauterine environment. In mice, for example, there is a precipitous rise in gene expression for UCP1 in late gestation (between 18 d and 19 d) that continues until ⩾8 d after birth⁽Reference Rim, Xue and Gawronska-Kozak¹⁴⁾. The abundance of UCP1 mRNA then declines by 1 month of age but is still retained throughout the life cycle (Fig. 1). This pattern contrasts with both human subjects and sheep in which UCP1 gradually increases in abundance through late gestation⁽Reference Clarke, Bryant and Lomax¹⁵⁾ to peak at birth⁽Reference Clarke, Heasman and Firth¹⁰⁾ before declining over the first month to undetectable levels in sheep⁽Reference Clarke, Buss and Juniper¹⁶⁾, a process that takes approximately 9 months after birth in human subjects⁽Reference Lean¹⁷⁾. These substantial differences between species must be given full consideration when considering the translational relevance of findings from small mammals to human subjects. Currently, for large mammals such as man and sheep there is no evidence that brown fat can be reactivated as, although there is some indirect evidence that brown fat is present in adult patients with cancer⁽Reference Nedergaard, Bengtsson and Cannon¹⁸⁾, its location is completely different from that seen in the newborn⁽Reference Lean¹⁷⁾. After the peak period of brown adipose tissue activation after birth not only is the UCP1 gene rapidly lost but this deficit is potentially accompanied by the in vivo, but not in vitro, loss of the key transcription factors PPARα and PPAR coactivator 1α⁽Reference Lomax, Sadiq and Karamanlidis¹⁹⁾ from ovine brown adipose tissue. At the same time, there is loss of glucose-regulated protein 78, an endoplasmic reticulum chaperone crucial for protein synthesis, that may further reflect the changes in protein content of adipocyte depots as they become primarily white in character⁽Reference Sharkey, Symonds and Budge²⁰⁾. Accompanying the loss of brown fat there is a substantial rise in the amount of white fat and white adipose tissue depots represent the fastest-growing organ⁽Reference Clarke, Buss and Juniper¹⁶⁾.

Fig. 1.

Comparison of the ontogeny of uncoupling protein (UCP) 1 between small (—, – –) and large mammals (, ). —, , UCP1 protein; – –, , UCP1 mRNA; HPA, hypothalamic–pituitary–adrenal; dGA, d of gestation.

Different origins of brown and white adipocytes

Recent studies have highlighted the importance of early-life events in determining aspects of adipocyte function and distribution. It has been shown that not only are white adipocyte progenitor cells committed to the adipose lineage during the late fetal–early postnatal period but there is also a marked expansion of this cellular pool as a result of proliferation during postnatal life⁽Reference Tang, Zeve and Suh²¹⁾. It must be noted that in species adopted for the latter study (i.e. mice), this is the time at which changes in fat distribution are greatest⁽Reference Rim, Xue and Gawronska-Kozak¹⁴⁾. In view of the strong possibility that the origin and fate of brown fat is very different between small and large mammals, it is important to reconsider the recent proposal that brown adipocytes have the same lineage as skeletal myoblasts, a process that may be regulated by bone morphogenetic protein 7 acting through PR-domain-containing (PRDM) 16⁽Reference Seale, Bjork and Yang^22, Reference Tseng, Kokkotou and Schulz²³⁾. The possibility of a common mechanism relating brown adipocyte and muscle development, and that it differs substantially from white adipocytes, is in accord with similarities between brown adipocytes and skeletal muscle and the distinct differences in myogenic gene expression found between brown and white cells⁽Reference Timmons, Wennmalm and Larsson²⁴⁾. Additionally, it has been suggested that functional brown fat may be present within muscle and other fat depots in some strains of mice⁽Reference Almind, Manieri and Sivitz²⁵⁾ and the thymus in rats⁽Reference Carroll, Haines and Pearson²⁶⁾. It is most unlikely that such depots have major biological importance⁽Reference Brennan, Breen and Porter²⁷⁾ in overall energy balance as UCP1 protein is barely detectable⁽Reference Almind, Manieri and Sivitz²⁵⁾. This finding is not surprising, given that mRNA abundance for UCP1 in both skeletal muscle and fat sampled from the subcutaneous and epididymal regions is <1% of that found in native brown fat (located within the intrascapular region) even when maximally stimulated by a β₃-adrenergic agonist⁽Reference Almind, Manieri and Sivitz²⁵⁾.

The UCP1 gene has also been identified in fetal, but not adult, human muscle-cell cultures⁽Reference Crisan, Casteilla and Lehr²⁸⁾. Consequently, UCP1 gene expression in vitro in fetal muscle CD34+ cells exhibits a much greater capacity to respond to the PPARγ agonist rosiglitazone than skeletal muscle cells isolated from adult human subjects⁽Reference Crisan, Casteilla and Lehr²⁸⁾. However, the pronounced increase in UCP1 mRNA transcription seen in fetal, as compared with adult, muscle that has been allowed to differentiate in vitro (i.e. fetal 5·5-fold v. adult 0·5-fold) may simply reflect an exaggerated adaptation as a result of an inability to translate the greater abundance of UCP1 mRNA to protein. Indirect support for this proposal comes from the observation that maximal stimulation of PRDM-16-expressing myotubes still results in substantially lower protein expression of UCP1 compared with native brown adipocytes, whereas PRDM-16 is only present at very low levels in native brown fat⁽Reference Seale, Bjork and Yang²²⁾. Thus, it is likely that although PRDM-16 has a role in adipocyte cell lineage, this action only occurs very early in development. Consequently, in adults overexpression of PRDM-16 for 4 weeks, although promoting energy expenditure, has a much smaller effect on UCP1 gene expression⁽Reference Tseng, Kokkotou and Schulz²³⁾. One likely mediator of PRDM-16 action is through the sympathetic nervous system⁽Reference Champigny, Ricquier and Blondel²⁹⁾, which could also explain the 0·4°C rise in body temperature of these animals following transfection with PRDM-16⁽Reference Tseng, Kokkotou and Schulz²³⁾. Ultimately, it could be predicted that it is the increased physical activity and therefore heat production in lean mice⁽Reference Almind, Manieri and Sivitz²⁵⁾ that acts to prevent excess fat accumulation. This process appears to act in combination with the persistently greater recruitment of non-exercise-activity thermogenesis, which is closely associated with a lower fat mass in adult human subjects⁽Reference Levine, Lanningham-Foster and McCrady³⁰⁾.

Early determinants of fat cell number and fat growth

Studies in human subjects have demonstrated that adipocyte number is a major determinant of fat mass in adulthood, with the number of fat cells being set during childhood and adolescence and remaining largely unchanged through adulthood in both lean or obese individuals⁽Reference Spalding, Arner and Westermark³¹⁾. Importantly, the developmental up-regulation of adipocyte number with obesity appears to increase during infancy. Consequently, this process occurs 4 years earlier in obese (2·1 years) children compared with those that remain lean (5·7 years)⁽Reference Spalding, Arner and Westermark³¹⁾. The extent to which such a process maybe be reset by the early nutritional environment is unknown, but in species with a long gestation fat cells first appear at about mid gestation⁽Reference Vernon, Buttery, Haynes and Lindsay^32, Reference Brennan, Gopalakrishnan and Kurlak³³⁾ before total fat mass increases up to term, as the fetus lays down sufficient energy reserves to enable it to meet the cold challenge of the extrauterine environment⁽Reference Clarke, Bryant and Lomax¹⁵⁾. Importantly, the net effect is to promote the abundance of brown fat in the offspring⁽Reference Symonds, Bryant and Clarke^34, Reference Budge, Bispham and Dandrea³⁵⁾ under thermal or nutritional environments that may act to enhance fetal development. In addition, term infants are born with substantial amounts of subcutaneous fat⁽Reference Widdowson, Dickerson, Comar and Bronner³⁶⁾ that increase during lactation, before being mobilised during infancy with its increased energy expenditure in motor development. The postnatal period therefore encompasses a period in which there is a dramatic increase in total fat mass but modulation of fat depot location within the body⁽Reference Clarke, Buss and Juniper¹⁶⁾.

Postnatal development of white adipose tissue

The loss of brown fat after birth, and subsequent replacement in fat mass depots by white adipocytes that have the potential for nearly ‘unlimited’ growth⁽Reference Clarke, Buss and Juniper¹⁶⁾, is determined in part by changes in glucocorticoids⁽Reference Gnanalingham, Mostyn and Symonds³⁷⁾. Between 1 week and 6 months after birth, as fat mass increases in a linear fashion in animals raised in natural ‘free-living’ environments, there is an accompanying rise in glucocorticoid action within the adipocyte⁽Reference Gnanalingham, Mostyn and Symonds³⁷⁾. At the same time, the abundance of UCP2 peaks at approximately 30 d of age⁽Reference Gnanalingham, Mostyn and Symonds³⁷⁾, coincident with the loss of UCP1⁽Reference Mostyn, Wilson and Dandrea³⁸⁾. As the switch from brown to white adipose tissue involves the proliferation and differentiation of preadipocytes and cell loss by apoptosis⁽Reference Prins and O'Rahilly³⁹⁾, UCP2 may be involved in the regulation of this transition⁽Reference Gnanalingham, Mostyn and Symonds³⁷⁾.

Ultimately, with increasing fat mass and accompanying inflammation adipose tissue becomes insulin resistant and, if exposure to an obesogenic environment occurs, the metabolic syndrome may result⁽Reference Symonds, Sebert and Hyatt⁴⁰⁾. As this process is a developmental one (see Fig. 2), it is essential that the normal changes within the adipocyte with increased age and fat mass are understood, as only then can appropriate intervention strategies be introduced to avoid the adverse metabolic consequences. Macrophage infiltration of adipose tissue contributes to the inflammatory state observed with obesity⁽Reference Rasouli and Kern⁴¹⁾. The mechanisms that drive this infiltration and alter insulin signalling remain unknown, although toll-like receptor 4 (TLR4) appears to have a critical role⁽Reference Nguyen, Favelyukis and Nguyen⁴²⁾. TLR4 is a major component of the innate immune system and is activated by its main ligand lipopolysaccharide. In addition, NEFA, plasma concentrations of which are normally raised with obesity⁽Reference Despres and Lemieux⁴³⁾, form a natural ligand for TLR4, inducing a local paracrine loop between macrophages and adipocytes⁽Reference Nguyen, Favelyukis and Nguyen^42, Reference Suganami, Tanimoto-Koyama and Nishida⁴⁴⁾.

Fig. 2.

Summary of the main changes in the characteristics of white adipose tissue during postnatal life. – –, Glucose-regulated protein 78 (GRP 78) mRNA; , toll-like receptor 4 (TLR4) mRNA; , CD68 mRNA; , IL-18 mRNA; dGA, d of gestation.

It has recently been shown that gene expression for both TLR4 and the macrophage marker CD68 follow a similar developmental pattern in the sheep. Their abundance peaks at 30 d of age, coincident with the transition of brown to white adipose tissue and accompanying lipid accumulation⁽Reference Clarke, Buss and Juniper¹⁶⁾, before declining to low concentrations in the adult⁽Reference Nedergaard, Bengtsson and Cannon¹⁸⁾. The strong correlation between TLR4, CD68 and relative fat mass⁽Reference Lomax, Sadiq and Karamanlidis¹⁹⁾ may be indicative of elevated NEFA delivery to the adipose tissue as its abundance increases. This process would be predicted to promote TLR4 action, thereby enhancing the differentiation of preadipocytes and their subsequent pro-inflammatory properties⁽Reference Poulain-Godefroy and Froguel⁴⁵⁾. In light of the close relationship between fat mass after birth and markers of both glucocorticoid action and inflammation, it has been possible to establish how these interactions may be reset, both after birth and in adolescence and early adulthood⁽Reference Lomax, Sadiq and Karamanlidis^19, Reference Gnanalingham, Mostyn and Symonds³⁷⁾. Importantly, that such adaptations can occur in the absence of any change in absolute fat mass implies that the latter is not required for all adverse metabolic outcomes⁽Reference Sharkey, Gardner and Fainberg⁴⁾.

The timing of maternal nutrient restriction and longer-term outcomes for adipose inflammatory and endocrine responses

Maternal nutrient restriction targeted to the period of very early adipocyte appearance in the fetus increases fat mass at term in conjunction with raised glucocorticoid action⁽Reference Gnanalingham, Mostyn and Symonds³⁷⁾. This response is accompanied by raised gene expression for UCP2, a characteristic of patients with visceral obesity⁽Reference Cassell, Neverova and Janmohamed⁴⁶⁾, but occurs in the absence of similar effects on inflammatory markers⁽Reference Lomax, Sadiq and Karamanlidis¹⁹⁾. However, even when the offspring are raised under an obesogenic environment they do not show increased fat mass compared with controls⁽Reference Sharkey, Gardner and Fainberg⁴⁾, although glucocorticoid action remains raised, at least in lean animals⁽Reference Gnanalingham, Mostyn and Symonds³⁷⁾. The same animals also show reduced gene expression of IL-18, an important mediator of the innate immune system⁽Reference Gracie, Robertson and McInnes⁴⁷⁾. Although, paradoxically, IL-18-knock-out mice develop obesity and insulin resistance⁽Reference Netea, Joosten and Lewis⁴⁸⁾, plasma IL-18 concentration is actually raised in obesity⁽Reference Skurk, Kolb and Muller-Scholze⁴⁹⁾. Decreased IL-18 gene expression in the adolescent offspring born to nutrient-restricted mothers⁽Reference Lomax, Sadiq and Karamanlidis¹⁹⁾ could ultimately reduce appetite suppression and so increase food intake, promoting positive energy balance and consequent obesity. An adaptation of this type would be in accord with epidemiological studies of human populations exposed to early gestational nutrient restriction that demonstrate an increased risk of obesity in the offspring⁽Reference Painter, Roseboom and Bleker⁵⁰⁾.

At birth, offspring exposed to mothers who have been nutrient restricted in late gestation have a reduced fat mass⁽Reference Budge, Edwards and McMillen⁵¹⁾. They do, however, show catch-up growth during the early postnatal period and thus restore their fat content to the same amount as offspring born to normally-fed mothers by 1 month of age. By 1 year of age they have an increased fat mass and are more likely to become glucose intolerant⁽Reference Gardner, Tingey and Van Bon⁵²⁾, which is similar to the epidemiological findings from historical cohorts⁽Reference Painter, Roseboom and Bleker⁵⁰⁾. A discrepancy between reduced fetal fat mass and increased juvenile adiposity in these offspring exposed to late gestational nutrient restriction suggests that a switch in fat growth during the early postnatal period is an important window in which later fat mass is set⁽Reference Tang, Zeve and Suh²¹⁾. Over this period there is a marked increase in TLR4 gene expression in the adipose tissue of offspring born to nutrient-restricted mothers that is not evident in controls⁽Reference Lomax, Sadiq and Karamanlidis¹⁹⁾, in conjunction with a decrease in the abundance of CD68 on day 1 of life, a pattern that is subsequently reversed. This type of adaptation could determine later fat mass if white adipocyte precursors are committed during prenatal or early postnatal life⁽Reference Tang, Zeve and Suh²¹⁾.

At the same time, preadipocytes can be inhibited by macrophage-secreted factors, therefore suppressing conversion into mature adipocytes⁽Reference Lacasa, Taleb and Keophiphath⁵³⁾. A lower macrophage content soon after birth is predicted to have important implications for early adipogenesis as adipose tissue undergoes substantial remodelling⁽Reference Clarke, Buss and Juniper⁵⁴⁾. This stage of development therefore represents a highly-sensitive period during which the prenatal-programmed reduction in macrophage content, and hence macrophage-secreted factors, may allow more preadipocytes to mature. Ultimately, this process would enhance the adipocyte pool and could contribute to a greater fat mass in later life. Subsequently, with increasing adiposity, the marked rise in TLR4 could trigger pro-inflammatory mechanisms, in particular the NF-κB and c-Jun NH₂-terminal kinase pathways⁽Reference Nguyen, Favelyukis and Nguyen^42, Reference Jiao, Chen and Shah⁵⁵⁾, further increasing the macrophage content⁽Reference Sharkey, Gardner and Fainberg⁴⁾. This response in turn would be predicted to result in adipocyte and macrophage cross talk leading to a down-regulation of glucose–insulin signalling pathways, particularly of the glucose transporter, GLUT4⁽Reference Lumeng, Deyoung and Saltiel⁵⁶⁾, as seen at 1 year of age⁽Reference Gardner, Tingey and Van Bon⁵²⁾. To date, such adaptations have only been described in offspring reared under free-living conditions. Clearly, future investigations need to examine the extent to which exposure to an obesogenic environment may accelerate or amplify these outcomes.