Following the discovery of leptin in 1994, the scientific and clinical communities have held great hope that manipulation of the leptin axis may lead to the successful treatment of obesity. This hope is not yet dashed; however the role of the leptin axis is now being shown to be ever more complex than was first envisaged. It is now well established that leptin interacts with pathways in the central nervous system and through direct peripheral mechanisms. In this review, we consider the tissues in which leptin is synthesized and the mechanisms which mediate leptin synthesis, the structure of leptin and the knowledge gained from cloning leptin genes in aiding our understanding of the role of leptin in the periphery. The discoveries of expression of leptin receptor isotypes in a wide range of tissues in the body have encouraged investigation of leptin interactions in the periphery. Many of these interactions appear to be direct, however many are also centrally mediated. Discovery of the relative importance of the centrally mediated and peripheral interactions of leptin under different physiological states and the variations between species is beginning to show the complexity of the leptin axis. Leptin appears to have a range of roles as a growth factor in a range of cell types: as be a mediator of energy expenditure; as a permissive factor for puberty; as a signal of metabolic status and modulation between the foetus and the maternal metabolism; and perhaps importantly in all of these interactions, to also interact with other hormonal mediators and regulators of energy status and metabolism such as insulin, glucagon, the insulin-like growth factors, growth hormone and glucocorticoids. Surely, more interactions are yet to be discovered. Leptin appears to act as an endocrine and a paracrine factor and perhaps also as an autocrine factor. Although the complexity of the leptin axis indicates that it is unlikely that effective treatments for obesity will be simply derived, our improving knowledge and understanding of these complex interactions may point the way to the underlying physiology which predisposes some individuals to apparently unregulated weight gain.
Obesity is a major health issue in much of the human population. In the UK over 16% of men and 17.5% of women are obese and a further 45% of men and 35% of women are overweight,1 representing a doubling since 1980. In the USA, it is estimated that over one-third of the population are overweight by at least 20%, and this proportion is increasing.2,3 Recently, a survey by Brown et al4 showed that 48% of Australian women aged 45–49 were overweight. Increasingly, obesity has become a serious medical problem in developing nations, with the incidence correlating with urbanization and a more plentiful food supply. The morbidities associated with obesity, including type 2 diabetes, cardiovascular disease, osteoarthritis and several forms of cancer, represent a major health risk to the obese population.
Physiologically, obesity is a disorder of energy balance; excess energy is stored as fat whenever energy intake exceeds energy expenditure. Thus, the optimal treatment for obesity would be one that both suppresses food intake and increases energy expenditure. Energy intake and energy expenditure are closely regulated processes, as reflected in the relative stability of body weight in the presence of large daily fluctuations in energy intake.5 Complex interactions between hormone axes in the periphery are integral in maintaining homeostasis of a diverse range of functions. Feedback from these complex interactions is ultimately integrated at the level of the central nervous system (CNS) by a similarly complex array of neurotransmitter signals.
An important regulator in both the CNS and the periphery is leptin. It is now over 8 y since the mouse and human ob genes were cloned,6 leading to the discovery that their expression encoded a 16 kDa protein, which was termed ob protein or leptin (from the Greek leptos, meaning thin). Although it was demonstrated that correction of leptin deficiency in the ob/ob mouse caused a marked reduction in food intake and a normalization of its obesity syndrome,7,8,9 the role of leptin in mediating short-term food intake is still a matter for debate. Both leptin deficiency (in ob/ob mice) and leptin resistance (in db/db mice, having a defective leptin receptor10,11) are characterized by hyperphagia and decreased energy expenditure, leading to an obese phenotype that is notable in animals less than 1 month old.12 The weight of the animals stabilizes at 60–70 g, compared with 25–35 g in control littermates (Figure 1). Both ob/ob and db/db mice manifest numerous other abnormalities, such as non-insulin-dependent diabetes mellitus with severe insulin resistance, hypothermia and cold intolerance, infertility and decrease in lean body mass.12 Studies of both mouse models have provided valuable understanding of many aspects of obesity and the role of the leptin axis. However, these pathologies are not always instructive in the understanding of normal leptin biology and the data presented throughout this review should be considered with this in mind.
This review will focus upon the peripheral interactions of leptin, and will address the aspects of leptin biology which are germane to our understanding of these roles, including the sites and regulation of both leptin and leptin receptor (Ob-R) synthesis, the mechanisms of leptin signalling, and its interaction with other regulatory signalling pathways from the level of the whole body, to the intracellular and molecular levels. These complex interactions will be illustrated using examples from in vitro models, and studies conducted in vivo in animals and in humans. Where there is species variation, we have attempted to clarify the known differences, especially between animal and human studies. However, there are many examples illustrated here in which leptin interactions are highly consistent across the species studied, and add great interpretive value to leptin biology and the factors contributing to obesity in humans.
Transcription of the leptin gene in mice yields a mRNA of ∼3.5 kb that is expressed primarily in adipose tissues, but recent studies have confirmed that some other tissues also express leptin, including placenta, ovaries, skeletal muscle and stomach.13,14,15,16,17 In humans, leptin is encoded by a gene located in human chromosome 7q31.3, and is similar to that in rodents.18
Leptin is translated as a 167 amino acid protein with an amino-terminal secretory signal sequence of 21 amino acids. The signal sequence is functional, and results in the trans-location of leptin into microsomes with the subsequent removal of the signal peptide.6 Therefore, leptin circulates in the blood as protein of 146 amino acid residues.
Tissue sites of leptin synthesis
White adipose tissue (WAT) is the main site of leptin synthesis, but it is now evident that the hormone is also produced in other tissues as well. There is some evidence that brown adipose tissue (BAT) is also a site of leptin production.19,20,21,22,23,24,25,26 A major issue is the physiological role of leptin produced by brown fat. One possibility is that it simply adds to the pool of circulating hormone, thus its contribution would be small relative to that of white fat in mature animals and man.27 Other reports suggest that the expression that is observed may be a reflection of contamination or infiltration of brown fat with white adipocytes.28 However, Moinat et al19 reported that the level of mRNA in interscapular brown fat was about 40% of that in the epididymal white fat depot, a result that is unlikely to be explained by infiltration or contamination. If the ob gene is expressed by brown adipocytes, then there are important implications for the physiological role of leptin. The interaction of the leptin axis and thermogenesis is discussed below.
Comparison of the levels of ob mRNA in different adipose tissue depots suggests that there are site-specific variations in the expression of the ob gene in both rodents,28,29,30 and humans.31,32 In humans leptin expression appears to be greater in subcutaneous than in omental adipose tissue.31,33,34 The differences in mRNA level between various adipose tissue sites may reflect differences in fat cell size; the larger the adipocytes, the greater the expression of the ob gene.28,29,30,34 However, Lonnqvist et al35 using in situ hybridization, found no differences in leptin expression between subcutaneous and omental fat in a small group of four obese humans.
In rodents the opposite is apparent as leptin expression may be lower in subcutaneous fat than in the internal depots, and the highest level of expression is generally evident in the epididymal (males) and perirenal adipose tissues.29,31 However, site-specific leptin expression also varies with maturation as in adult rodents; the ob mRNA level is much higher in the gonadal and perirenal sites than in subcutaneous depots,28 but in suckling rats during the first 1–2 weeks after birth, the subcutaneous fat is the main site of ob gene expression.36
Sexual dimorphism in humans is evident in both ob mRNA expression,33,37 and in the correlation between leptin concentrations and fat mass.38,39 Kennedy and co-workers39 proposed that the observed gender differences in leptin synthesis are because of the stimulating roles of oestrogens and/or a suppressive effect of circulating androgens, while other authors have been unable to find a correlation between sexual dimorphism and sex hormones.40
An important new dimension to leptin biology has emerged with the recognition that the placenta and ovary express the leptin gene and that they are sites of production of the hormone. This has been demonstrated in mice, rats and humans.13,15,41,42,43 Two hypotheses about the role of placental synthesis of leptin are that it may either be a new growth factor, or act as a signal of energy status between the mother and the foetus. The placenta also expresses the leptin receptor gene, implying that the organ is a target for the action of leptin as well as being a source of the hormone.13 These observations suggest that leptin may act in an autocrine manner. The expression of leptin by syncytiotrophoblasts13,44,45 has added support to the hypothesis of the importance of leptin in nutrient transfer. In situ hybridization and immunohistochemical studies on pregnant mice have demonstrated leptin synthesis in several regions of the foetus, including the heart, bone and cartilage, choroid plexus of the foetal brain, lung, kidney, heart and liver, and cells of the hair follicle.46
Leptin has also been detected in rat stomach.16 Leptin mRNA was identified after screening of total RNA extracted from fundic epithelium scrapings using the polymerase chain reaction after reverse transcription of RNA. Leptin mRNA was not detected in any other gastrointestinal tissue, including the liver and pancreas. Subsequent studies demonstrated that leptin immunoreactive cells were localized in the lower half of the fundic glands, a site similar to that of the pepsinogen secreting chief cells.16,47 Recent studies have also shown leptin synthesis by other tissues, including rat skeletal muscle,17 rat and mouse pituitary gland,48,49 human mammary epithelial cells,50 human and newborn mouse bone marrow51,52 and mouse liver.53
Taouis et al,54 reported that leptin gene expression in chicken was not only localized in adipose tissue, as in mammals, but also was present in liver. The expression of leptin in liver may be associated with a key role of this organ in avian species in controlling lipogenesis. The diverse picture of where leptin is produced in the body indicates that the functions of the hormone may extend beyond the basic lipostatic model originally envisaged.
Regulation of leptin synthesis
The level of ob mRNA in white adipose tissue and the circulating leptin concentration are increased markedly in obesity, as shown in both human studies and in studies of several types of obese animal.30,55,56,57 Indeed, in human subjects there is a high correlation between body mass index (BMI) and circulating leptin.30 Thus, the greater the amount of adipose tissue, the higher the level of the hormone. In addition, as previously mentioned, adipocyte size appears to be another major determinant of leptin mRNA expression.58 Hormonal factors have also been studied in relation to leptin and are discussed below.
Exercise training induces an increase in metabolic rate and overall energy expenditure.59 The mechanisms which regulate leptin mRNA expression and circulation levels during exercise are still unknown. Controversy exists over whether exercise has any effect on leptin levels. Landt et al60 demonstrated that prolonged endurance exercise, such as marathon running, decreased leptin concentrations. Hypoleptinemia was also detected in female and male elite gymnastics of pubertal age.61 Leptin gene expression was also decreased in Sprague–Dawley rats after 4 weeks of exercise training.62 However, other studies63,64,65,66 have shown that moderate-intensity aerobic exercise and acute and chronic exercise in men and women do not affect leptin levels. In contrast, Hickey et al67 reported that moderate intensity aerobic exercise in women reduced leptin levels by 17.5% after 12 weeks. Pasman and co-workers68 investigated the effect of 4 month moderate-intensity training on leptin levels in obese males and found that serum leptin levels were reduced by 23%. Hickey and co-workers67 have suggested that the gender-specific responses to training reflects differences in insulin-resistance between males and females; males, being the more insulin-resistant, may need more time and a greater stimulus to respond with lower leptin levels. There are a number of potential mechanisms by which exercise might regulate leptin production, such as via changes in insulin sensitivity to fatty acid concentrations or by alterations in sympathoadrenal activity.69 As leptin concentrations are determined by energy balance, it is conceivable that energy expenditure may also be a regulatory signal for leptin production.
It has been shown that cold exposure reduces circulating leptin levels and leptin expression,19,28,70,71,72 suggesting that leptin may participate in the adaptive mechanism triggered by variations in external temperature. Trayhurn and co-workers28,73 hypothesized that the inhibitory effect of low temperature on leptin results from an increase in the adrenergic tone induced by exposure to cold temperatures, which would in turn act through the β3-adrenoceptor (β3-AR) present in adipose tissue, inducing a reduction in the expression of leptin mRNA. It has been reported that this was the case for mice19 and rats.28 Hardie et al70 also found that acute cold exposure decreased the level of circulating leptin in lean Zucker rats. However, it appears that this is not so for other rodents, such as the Djungarian hamster.21 For animals living in areas with large seasonal temperature variations, a decrease in serum leptin may represent an adaptive mechanism for maximising the size of fat deposits when environment temperature is low. Trayhurn et al28 showed that the leptin mRNA observed in brown adipose tissue was further reduced on cold exposure, paralleling the response observed in mouse white adipose tissue. These authors postulated that these findings may suggest the existence of a feedback loop between the hypothalamus and BAT, that would result in an inhibition of leptin gene expression when the sympathetic system is activated. Interestingly, Trayhurn et al28 reported contrasting results to those reported by Moinat et al,19 who showed no change in leptin expression in the white adipose tissue of rats which were acutely cold-exposed (6°C for 24 h).
Changes in physiological state induced by fasting result in reduced ob gene expression and a subsequent fall in circulating leptin.28,74,75 Rapid decreases in leptin levels in response to energy restriction or fasting appear to be greater than the decrease in fat mass in both rodents55 and humans.76 Thus, leptin may serve as a sensor of short-term changes in energy stores. Another study in mice has shown that re-feeding after fasting led to a rapid restoration of gene expression and plasma leptin.77 The inhibitory effect of fasting on leptin expression might also be mediated by the sympathetic nervous system and the β3-AR. Leptin is involved in increased energy utilization, possibly through enhanced thermogenesis mediated by tissue uncoupling proteins (UCPs). Studies on UCP subtype expression in rodents have shown that UCP-1 is expressed at higher levels during cold exposure and decreased by fasting78 and that leptin treatment enhances79 or does not alter80 BAT UCP-1 mRNA, compared with untreated ad libitum fed controls. These contrasting findings indicate that the interaction between cold exposure, thermogenesis and leptin production requires further investigation.
The role of leptin has been investigated in the pathogenesis of eating disorders. In most studies, serum leptin concentrations in patients with anorexia nervosa, bulimia, non-specific eating disorders81 and depression82 are similar to those of healthy persons with comparable BMI. In contrast, Haluzik et al83 found that serum leptin levels in patients with anorexia nervosa were significantly lower than those in the control group. However, Mantzoros et al84 found that patients with anorexia nervosa have relatively higher transport of leptin to the cerebrospinal fluid at lower serum leptin concentrations, which may explain both the symptoms of anorexia nervosa and the difficulty that these patients have in regaining weight.
The sympathetic nervous system, especially the β-AR axis, has been implicated in the regulation of leptin gene expression. For example, the β-AR agonists, noradrenaline and isoprenaline decreased leptin gene expression in WAT in mice73 and decreased serum leptin.85 It has been shown, both in vivo and in vitro, that β3-agonists induce a decrease in the expression of leptin in WAT and BAT and decrease serum leptin.27,73,86,87,88 The significance of β3-AR in the control of leptin production in human subjects is uncertain, although there is evidence that the receptor may play an important role in the control of lipolysis in human omental and subcutaneous adipose tissue.89,90 Mantzoros et al86 reported that acute treatment of mice with the β3-agonist, CL 316,243 suppressed leptin expression in WAT. Inhibition of leptin expression was observed only in normal (control) mice, while the effect of β3-AR agonist was absent in knockout mice (with targeted disruption of the β3-AR gene). Also, despite the falling leptin levels, β3-agonist administration acutely suppressed food intake. Li et al87 showed that another β3-agonist, CGP-12177, also caused reduction in food consumption in F-344 rats. In another study,19 β3-AR stimulation by Ro 16-8714 decreased leptin mRNA in WAT of lean (Fa/Fa), but not obese (fa/fa) Zucker rats. Thus it appears that the leptin response to specific β3-AR activation is analogous to classical adrenergic activation. Thus, it appears that the body's response to a stressor such as low temperature conserves triacyl-glycerol (TAG) from mobilization for use by muscle and promotes its use for thermogenesis. In addition, leptin has been shown to increase noradrenaline turnover in thermogenic interscapular BAT91 and to also increase the sympathetic outflow to other tissues,92,93 consistent with this hypothesis.
Leptin-treated mice lose more weight than pair-fed vehicle-treated animals, implying that leptin also increases energy expenditure.8 Furthermore, Campfield et al7 showed that leptin-treated animals have higher core temperatures and metabolic rates than controls. These observations are consistent with the peripheral interaction of leptin with insulin, mobilizing fuels and inhibiting energy storage mechanisms as outlined in more detail below. It is also possible that leptin increases overall sympathetic nerve activity, thereby leading to a significant increase in energy expenditure. Summaries of variables that appear to regulate leptin production are shown in Tables 1 and 2.
The leptin gene promoter is also positively regulated by the functional binding site for CCAAT/enhancer binding protein α (C/EBP-α), a transcription factor important in adipocyte differentiation.94,95 In contrast, thiazolidinediones, a class of new antidiabetic agents which activate another transcription factor involved in adipocyte differentiation, peroxisome proliferator-activated receptor γ(PPARγ), suppress leptin expression in adipocytes.96,97 Thus, the relationships between the control of leptin expression and adipocyte differentiation are not well understood.
Leptin receptors (Ob-R)
The identification of the leptin receptor (Ob-R) was realised through an expression cloning strategy.10 Tartaglia and co-workers10 constructed a cDNA library from mouse choroid plexus. A cDNA that encoded the Ob-R was found by expression cloning using an alkaline phosphatase-Ob fusion protein as a probe. Expression of the Ob-R in COS cells was achieved, and displaceable binding of mouse and human leptin was observed. Scatchard analysis revealed a high affinity receptor with a binding constant of about 0.7 nM. The human homologue of the mouse leptin receptor was cloned using a human infant total brain library.10 The murine and human leptin receptors are highly similar in amino acid sequences for both the extracellular (78% identity) and intracellular domains (71% identity).98In situ analysis with probes generated from the extracellular domain (common to all Ob-R forms), RNase protection or polymerase chain reaction (PCR) revealed expression of Ob-R in multiple tissues. Ob-R exhibit a widespread distribution including liver, heart, kidneys, lungs, small intestine, testes, ovaries, spleen, pancreas and adipose tissue.11,13,52,99,100,101,102 However, analysis has shown that the majority of transcripts detected are those encoding short intracellular domain forms.52,99 Within the hypothalamus, the long form leptin receptor has been found in several hypothalamic nuclei, including the arcuate nucleus, ventromedial, dorsomedial and lateral hypothalamic nuclei, and the paraventricular nucleus.103,104
Previous data suggest that the diabetes (db) gene encodes the receptor for the obese (ob) gene product, leptin.6,105 Leptin receptor maps to the same 300 kb interval as db, and has at least six alternatively spliced forms.11 Lee and co-workers11 isolated candidate genes for db from two bacterial artificial chromosomes, 43 and 242, and complementary DNA selection from mouse hypothalamus. Mouse cDNAs predicted proteins with differing cytoplasmic portions, designated Ob-Ra, Ob-Rb, Ob-Rc (also known as B219) and Ob-Rd. Ob-Re has a different predicted amino-acid sequence after His 796.11
All isoforms of the murine leptin receptor share an identical extracellular, ligand-binding domain of 816 amino acids, while the intracellular domain at the C-terminus is different.10,11 Ob-Ra, Ob-Rb, Ob-Rc, Ob-Rd and Ob-Rf contain transmembrane domains of 34 amino acids, whereas Ob-Re is truncated before the membrane-spanning domain and is therefore likely to be secreted.11 Secreted extracellular domains of other cytokine receptors are also known to function as specific binding proteins.106 Ob-Rb (the human form of which is also referred to as Ob-RL; L=long) has a long cytoplasmic domain of 320 amino acids. Recently, another splice variant of leptin receptor, B219, was cloned from murine foetal liver.107
Sequencing of the original murine cDNA revealed a single membrane-spanning receptor, which belongs to the gp 130 family of cytokine receptors.10 The extracellular region of these receptors is characterized by the presence of multiple domains. Each of these domains is characterized by unique consensus residues.108,109,110 For cytokine receptors, the CK-F3 domain seems to form the ligand binding site. In contrast to the growth hormone receptor, the extracellular region of the leptin receptor contains two repeating CK-F3 domains. Such repeating CK-F3 domains are not commonly found in cytokine receptors, and localization of the ligand binding site was not described until recently. Fong et al111 found that the first CK-F3 domain of Ob-R is not required for leptin binding and receptor activation, whereas the second CK-F3 domain is the most likely leptin binding site.
Ob-R-induced signal transduction
Class I cytokine receptors are known to act through cytoplasmic tyrosine kinases of the Janus kinase family (JAKs) and signal transducers and activators of transcription (STAT). The intracellular domain of the truncated leptin receptor variant, in the db/db mouse, contains the Box 1 motif, but lacks the Box 2 and 3 motifs, required for activation of JAK.11,112 In the db/db mouse, the short form of the leptin receptor, appears to be unable to activate the JAK/STAT pathway.99,113 Zucker fatty (fa/fa) rats have a full-length, but dysfunctional leptin receptor. The study of Yamashita et al114 showed that transformed Chinese hamster ovary cells (CHO) which expressed fa-type Ob-Rb exhibited slightly reduced leptin binding activities and reduced signal transduction, in comparison to CHO cells which expressed wild-type Ob-Rb. The affinity values obtained by Yamashita et al114 were 0.4 nM for normal rat Ob-R and 1.2 nM for fa rat Ob-R. These data illustrate that only a relatively minor decrease in the affinity of the leptin-Ob-R interaction results in a loss of receptor function. It is unknown whether this single amino acid substitution in the extracellular domain has any direct effect on second messenger signalling.
Upon leptin binding to its receptor, receptor dimerisation occurs, which seems to be necessary for signalling activity.115,116,117 After leptin binding and receptor homodimerization, the long form of the receptor (Ob-Rb) activates JAK/STAT pathways. The activated JAKs phosphorylate tyrosine sites on the intracellular domain of the receptor, which serve as docking sites for the Src-homology domains (SH2, SH3) that occur in all of the STATs.108,110 The phosphorylated intracellular domain then provides a binding site for a STAT protein. The activated STAT proteins dimerise and translocate to the nucleus, where they bind DNA and activate transcription. Ob-Rb activation may also phosphorylate JAK leading to the activation of insulin receptor substrate (IRS-1) and mitogen-activated protein kinase (MAPK). Upon leptin binding, the short form of the leptin receptor (OB-Ra) phosphorylates IRS-1 and consequently activates MAPK. The activation of MAPK leads to the activation of pp90 S6-K.118,119 Thus, as described above, leptin receptors share, at least in part, some of the signalling pathways characteristic for the class I cytokine receptor family.
The intracellular domains of Ob-Ra–Ob-Rd contain a Box 1 motif, which is highly conserved among most members of the cytokine family.110 A less conserved Box 2 motif is also found in a number of receptors of this family. Mutational analyses of the conserved regions in several receptors suggested that these two domains are required for activation of the JAK/STAT pathway. Since short isoforms of leptin receptor do not contain Box 2, it has been predicted that they are signalling inactive.10,11,120 However, in some cases, cytokine receptors of the GH receptor subfamily with a mutation in Box 2 motif and/or with only short intracellular domains have been shown to have signalling capability.113,121,122 Bjorbaek and co-workers118 have found that Ob-Ra was able to activate JAK and IRS-1 tyrosine phosphorylation in a leptin-dependant manner in transfected COS cells. Their results showed that Ob-Ra has signalling capability, and that these required Box 1, but not Box 2, as was predicted before.
The activation of STAT proteins in response to leptin was assayed in a variety of mouse tissues or cells known to express Ob-R. In vivo, in mice, administration of recombinant leptin caused activation of STAT3 in the hypothalamus of wild-type and ob/ob mice.123,124,125 However, in vitro, binding of leptin to leptin receptors activates STAT1, STAT3 and STAT5.99,113,120,124,126,127 Recently, a new family of cytokine-inducible inhibitors of signalling has been identified, including cytokine-inducible sequence (CIS) and suppressor-of-cytokine-signalling-3 (SOCS3).125,128,129,130,131,132,133 CIS and SOCS3 may function as part of an intracellular negative-feedback loop, inhibiting JAK activity, and thereby switching off or dampening cytokine signal transduction.131,132
Leptin receptor function
The reasons why several distinct short intracellular forms of leptin receptors are produced and their respective roles are still not clear. It has been suggested that the short intracellular domain form in the choroid plexus plays a role in transporting leptin from the blood into the cerebrospinal fluid,11,98 where it can move by diffusion to the brain centres that regulate body weight. It has been shown that leptin enters the brain by a specific and saturable mechanism,133,134 but the role of Ob-Ra in this process has not yet been fully demonstrated. However, the recent study by Kastin et al135 showed that the transport of leptin across the blood–brain barrier in rats lacking the Ob-Ra was decreased, indicating that Ob-Ra may have a functional role in leptin transport. As mentioned before, it has been shown that Ob-Ra may perform signal transduction, although different from that induced by the long form.118,136 Murakami and co-workers136 have found that expression of mRNA for the immediate-early genes c-fos, c-jun and jun-B, which are induced by addition of leptin, is observed in both CHO cells expressing Ob-Rb and also in cells expressing Ob-Ra. Thus, it appears that short and long isoforms of the leptin receptors have the potential to mediate signal transduction. Although the short forms have an attenuated signal, their relative abundance in peripheral tissues may mediate specific effects of leptin on those tissues.
It has been implied that Ob-Rc and Ob-Rd isoforms may play a role in clearance of leptin from the circulation.11 Two of the leptin receptor isoforms have been described in only one species: Ob-Rd in the mouse11 and Ob-Rf in the rat.137 It has been proposed that the soluble isoform Ob-Re functions as a binding protein. Fei et al,104 using RT-PCR, found that Ob-Re was not expressed significantly in mouse tissues, while the study of Chua et al138 showed that Ob-Re was not expressed at all in human tissues (brain, skeletal muscle, heart, liver and lung). The exception to this is during pregnancy,139 which is covered in greater detail below. There are, however, reports of an Ob-R mutation in humans.140,141 A consanguineous family with both homozygous and heterozygous members had higher plasma leptin concentrations than expected given their BMIs: 526, 600 and 670 ng/ml and 145, 212, 240, 294 and 362 ng/ml, respectively. The mutated Ob-R had no transmembrane or intracellular domains and >80% of circulating leptin was in a high molecular weight associated form, compared with 7.5% bound in a non-mutated sibling. However, Lollmann et al,142 using RNase protection assay, found that in mice Ob-Re expression was ubiquitous and occurred in rather large amounts. These authors142 concluded that Ob-Re is produced at a level that is sufficiently high to act as a buffering system for free circulating leptin.
Another splice variant of the leptin receptor was cloned from murine foetal liver, B219.107 Results from the Northern analysis and PCR analysis of ObR/B219 mRNA showed that the B219 isoform was associated with haemopoietic stem cells, thus these authors proposed a possible role of leptin in haemopoiesis.
Expression of Ob-R in peripheral tissues
There are numerous reports of the expression of leptin receptors in peripheral tissues, including liver, heart, kidneys, lungs, small intestine, pituitary cells, testes, ovaries, spleen, pancreas, adrenal gland and adipose tissue.11,13,14,49,52,99,100,101,102,107,143 Furthermore, a number of studies have now reported partial characterisation of Ob-R using radioligand binding, confirming the presence of high- affinity leptin binding sites in rodent kidney144 (and Margetic et al, submitted) as well as in the CNS.105,145,146,147 These studies are supported by binding analysis in a range of transformed cells expressing particular isoforms of Ob-R.10,113,114,126,148,149 These findings suggest that leptin may have more complex roles than initially thought, raising the question of additional physiological functions of leptin in different tissues (Figure 2).
Hoggard et al14 examined expression of leptin receptor splice variants in murine peripheral tissues. Reverse transcriptase-polymerase chain reaction indicated that the leptin receptor, Ob-R, in particular the short splice variant of the leptin receptor, Ob-Ra, was expressed in a wide range of tissues. Expression of Ob-R was localized by in situ hybridization to specific sites in the spleen, testes, intestine, heart, lung, liver skeletal muscle and adrenal gland. However, the long form of the leptin receptor, Ob-Rb, was only detected at significant levels in the medulla of the adrenal gland and in the inner zone of the medulla of the kidney.
Chen et al52 investigated tissue localization of Ob-R mRNA in brain and peripheral tissues of adult and newborn mice using in situ hybridization. This study confirmed previously published reports99,103,150,151 that the long isoform of leptin receptor was predominantly expressed in hypothalamic nuclei, while low levels were observed in choroid plexus. Chen and co-workers52 further described the expression of Ob-Rb in peripheral tissues of adult mice as either very low or undetectable. In newborn mice, the expression of Ob-Rb was similar to that of adult mice, although bone marrow was the site of highest Ob-R expression.
Cao et al152 demonstrated the presence of Ob-Ra on adrenaline-secreting chromaffin cells in rat adrenal medulla, suggesting that leptin may directly affect the adrenal medulla. In addition, Hoggard et al14 and Takekoshi et al143 showed that Ob-Rb was highly expressed in the mouse and porcine adrenal medulla, respectively.
Recently, Tsuchiya et al153 found that the leptin receptor expressed in human lungs is the long isoform, Ob-Rb, suggesting that leptin may be involved in the peripheral regulation of respiratory function in humans. However, several studies10,14,104 have shown the presence of only Ob-Ra in lung, or both Ob-Ra and Ob-Rb.101,142 Thus, it appears that there may be a further unidentified role for leptin in the lungs.
De Matteis et al101 studied localization of leptin receptor splice variants in mouse peripheral tissues by immunohistochemistry. Their results indicate that most of the organs expressing Ob-Ra (adrenal gland, adipose tissue, heart, liver, lung, ovary, endocrine pancreas, skeletal muscle and testis) also express Ob-Rb in the same cell type, although in some (adrenal gland, liver, endocrine pancreas and testis) staining for Ob-Rb was weaker. Thus, many tissues may contain a heterologous mix of Ob-R subtypes.
Characterization of Ob-R in peripheral tissues by radioligand binding analysis
Excluding studies of binding analysis of Ob-R in the CNS, binding has also been characterized in mouse lung, intestine, kidney, liver, skin, stomach, heart and spleen.154 Additionally, we have characterized leptin binding in the bovine kidney (Margetic et al, submitted) and leptin binding in rat kidney has been characterized by other workers.144
The study of Serradiel-Le Gal et al144 found the affinity of the leptin binding site to be ∼0.5–0.7 nM, which appears to be low in comparison to other studies in which the affinity of leptin for the Ob-R was accurately determined.114,145,149,154
The study of Dal Farra et al154 appears to be thorough, the binding affinity being characterized in a range of tissues, Kd=0.1–0.3 nM. However, of concern is that the highest concentration of radioligand used was not stated, but appears to be only ∼0.6 nM. Given the experimentally determined affinity (∼0.2 nM), saturation analysis would require a total radioligand concentration of at least 4 nM. Thus, this study has not fulfilled the underlying criteria for saturation analysis, and their results must be regarded with caution. Despite this, the affinity is in close agreement with that found in studies in our own laboratory (Margetic et al, submitted), in which leptin binding affinity to bovine kidney membranes was 0.1 nM. However, the number of binding sites characterized in our study, 46 fmol/mg protein, and in the study of Serradiel-Le Gal et al144 in rat kidney, 45 fmol/mg protein, may have been underestimated in the study of Dal Farra et al,154 which found a Bmax=6 fmol/mg protein in mouse kidney membranes.
Tissue distribution studies and leptin pharmacokinetics
The authors155 and others156 have demonstrated in pharmacokinetic studies in normal rats that the loss of leptin from the circulation is best described by a two-pool model, consistent with the occurrence of leptin in two forms: free and bound to other proteins. The biological forms which constitute the bound pool of leptin include the hormone bound to tissue receptors, to non-specific sites in the tissues, and to a carrier molecule(s) in plasma155 (and Margetic et al, submitted). The rapid removal of the free form of leptin from plasma (half-life, 3.4 min) is consistent with data for other peptide hormones which occur in the free form such as insulin (half-life, ∼4 min157) or insulin-like growth factor-1 (IGF-1) in which the free form is cleared rapidly (half-life, 12–15 min158,159). Furthermore, clearance of the bound form of leptin which is retained in the plasma for a much longer period (half-life, 71 min155) is consistent with the values obtained for IGF-1, which is known to also occur bound to one of several specific binding proteins (half-life, 148–264 min158,159), suggesting that a specific binding molecule for leptin may also be important in the rat.
Van Heek et al100 determined brain and whole body localization and distribution of 125I-leptin after intraperitoneal (i.p.) administration to ob/ob and db/db mice. After a single injection of 125I-leptin, radioactivity was detected at early time points in all tissues sampled in both ob/ob and db/db mice. Analysis of the tissue distribution of 125I-leptin revealed that leptin rapidly appears in the serum and accumulates in the fat, intestines, liver and kidneys in both ob/ob and db/db mice. These similarities may be unexpected, as these two models are at the ‘extremities’ in terms of leptin biology (ob/ob, non-functional leptin with up-regulated receptors; db/db, high circulating leptin with non-functional receptors). Thus, leptin distribution may be affected by its hydrophobic character, as non-specific interactions with abundant lower affinity binding sites, independent of binding to specific Ob-R, may have a large effect on its tissue distribution. Alternatively, it must be remembered that these results were obtained in animals which received very high doses (0.3 mg/kg) of exogenous leptin. Thus, it is likely that receptors were saturated and that non-specific binding sites were disproportionately represented in this study, possibly obscuring differences between ob/ob and db/db mice.
Klein et al160 used an arteriovenous balance method in humans and found that the rate of clearance of plasma leptin was 1.50 ml/kg/min. This lower rate of clearance may indicate that leptin in humans remains biologically active for a longer period than in rodents. These results were recently supported by a study in another primate, Rhesus monkeys,161 which found a similar clearance rate (1.82 ml/min/kg). It is likely that leptin association with carrier molecules which appear to be different across the species investigated to date162,163 is an important factor affecting leptin clearance.
Pharmacokinetics studies also provide evidence on the relative tissue distribution of leptin.155 The high abundance of leptin distributed throughout the peripheral tissues compared with brain is consistent with data from studies of Ob-R mRNA expression in both rodents and humans.11,14,99,107,150 Using Northern blot analysis Cioffi et al107 showed that the mouse tissues with the highest abundance of Ob-RNA were lung and kidney, with less in heart, liver, spleen, muscle and brain, and none in testis. The distribution of tissue expression may be quite species-specific, as in human tissues the highest expression was found in heart, liver, small intestine, prostate and ovary, whereas levels were relatively lower in lung and kidney. Ghilardi et al99 quantified the results of RNase protection assays using densitometry. In normal mouse tissues, mRNA from both the long and the short isoforms of the receptor were detected, and the relative abundance of the long isoform expressed as a percentage of the total Ob-R mRNA. Interestingly, stomach (11%), small intestine (7%), pancreas (11%), skeletal muscle (8%), and heart (10%) had comparable abundance to that in the brain (18%). However, the level in hypothalamus was higher (36%).
In addition to providing corroborating evidence of Ob-R mRNA expression in defining the tissue distribution of Ob-R, pharmacokinetics studies provide evidence for the relative importance of rate of synthesis of leptin, and may indicate that short-term changes in rate of leptin synthesis and thus, minor changes in circulating plasma leptin concentrations, may be of little biological significance owing to the prolonged half-life of leptin in plasma and the presence of a substantial pool of non-specific tissue and plasma binding sites, providing a large buffering capacity for circulating plasma leptin concentrations. If this is so, the role of leptin, particularly in satiety, requires further clarification, and suggests a more complex interaction between leptin, its transport across the blood–brain barrier and CNS-specific hormones such as neuropeptide Y.
Major peripheral interactions
The role of leptin in the circulatory system
Leptin and haemopoiesis
It has recently been shown that Ob-R is expressed in haemopoietic cells and that leptin may be linked to the proliferation and differentiation of haemopoietic precursors.107,164 This topic was recently reviewed by Gainsford and Alexander165 and will only be briefly reviewed here. It is perhaps not surprising that leptin appears to have a role in haemopoiesis. Leptin structure is similar to the cytokines, containing four alpha helices. Furthermore, the Ob-R is a member of the type I cytokine receptor superfamily. In cell culture studies, it appears that leptin has a mediation role, at least in murine myelocytic cells from bone marrow, in cell proliferation.166 However, at least one study using human bone marrow cells was unable to demonstrate clonal proliferation.164 Rather, in this study differentiation of transformed myelocytic cells expressing Ob-R was evoked by leptin.
Of particular interest, the study of Umemoto et al166 provided data describing the colony formation of granulocyte macrophage cells from bone marrow derived from both normal and db/db mice. The dose–response relationship for leptin was rightward shifted in db/db cells compared to normal cells. Thus, there was a significant response in normal cells at a leptin concentration of 10 ng/ml (physiological concentration), while in db/db cells (which have a dysfunctional Ob-R), a leptin concentration of 100 ng/ml evoked a similar response. This evidence suggests that intracellular events (at least in these cell types) may be evoked by supraphysiological concentrations of leptin, which are independent of the Ob-R.
Furthermore, this concentration of leptin is typical of that observed in obese individuals and poses an intriguing question: if leptin at this concentration has intracellular effects independent of the Ob-R, is obesity a ‘pathological state’ in which some of the peripheral actions of leptin are dysfunctional, and independent of Ob-R binding?
Nakata et al167 reported that relatively high concentrations (50–100 ng/ml) of human leptin corresponding to plasma leptin levels in obese individuals, promoted platelet aggregation. At lower concentrations (<10 ng/ml), leptin failed to potentiate agonist-induced platelet aggregation. This suggests that the leptin effect of potentiating platelet aggregation operates specifically in obese individuals, and it may be a key coupling factor between obesity and the cardiovascular disease associated with syndrome X (which includes glucose intolerance, insulin insensitivity and abnormal fat distribution) and diabetes. Alternatively, these data may indicate that leptin action in platelets functions via an Ob-R-independent mechanism, as suggested above.
Leptin, other roles in the circulatory system—angiogenesis
Further roles for leptin have been proposed in the circulatory system, including stimulation of endothelial cell growth and angiogenesis.168,169 Several studies have demonstrated that vascular endothelium expresses the long form of the receptor146,167,168,169,170,171 (in rodents and humans170) and short131,171,172 forms of the leptin receptor in rodents.
Sierra-Honigmann et al168 demonstrated that the leptin receptors are expressed in human vasculature and in primary cultures of human umbilical vein endothelial cells (HUVEC). Ob-Rb was detected in HUVEC using immunofluorescence microscopy and rabbit polyclonal antibodies specific for the intracellular domain of the Ob-Rb isoform, as a 170 kDa protein. The proliferative effects of human leptin on HUVEC have also been reported by Bouloumie et al.169 RT-PCR analysis revealed the expression of both short and long isoforms of Ob-R in endothelial cells. Moreover, human leptin evoked a time-dependent tyrosine phosphorylation of a number of endothelial proteins, the most prominent of which were the mitogen-activated protein kinases, Erk 1/2. Treatment of HUVECs with leptin led to a concentration-dependent increase in cell number that was maximal at a leptin concentration of 10 ng/ml. This effect was associated with enhanced formation of capillary-like tubes in an in vitro angiogenesis assay and neovascularization in an in vivo chick allantoic membrane (CAM) model of angiogenesis. We have found similar effects of leptin on angiogenesis using the CAM model.173 Of particular interest in this study was the dose–response relationship of leptin with angiogenic activity. At 0.1, 0.5 and 1 µg/disc applied to CAM, leptin had increasing angiogenic activity. However, at higher concentrations (3 and 10 µg/disc), not only was angiogenic activity increased, but there were also pronounced toxic effects including capillary leakage, and marked avascular zones. These data appear to be consistent with findings in other systems such as the colony formation induced in granulocyte macrophage cells from db/db mice166 (cited above), and may suggest that some actions of leptin, when present in high concentrations, are associated with pathophysiological effects independent of the Ob-R.
In a recent study, Ring and co-workers171 studied leptin's effect on wound repair in ob/ob and db/db mice. Systemic administration of leptin at doses ranging from 0.1 to 10 mg/kg/day induced a highly significant acceleration in wound repair in ob/ob mice, but not in db/db mice, indicating that leptin effects on wound repair were at least partially due to a direct interaction of leptin with its receptors at the wound site. Topical application of leptin also induced dose-dependent acceleration in wound repair. However, these authors also found in ob/ob mice that neither systemic nor topical leptin application changed wound haemoglobin or wound vessel density. These data may suggest that leptin is acting via a mechanism independent of angiogenesis, however further studies are required to clarify this point.
Leptin and reproduction
Recently, it was found that leptin plays an unexpected role in reproduction.41,174,175 Human ovary and prostate and murine ovary and embryo express mRNA for the leptin receptor.10,107,176 Although much of the evidence suggests a central site for the effects of leptin on the reproductive axis, a direct effect on the ovary has also been demonstrated as leptin at physiological concentrations inhibits insulin-induced oestradiol production by granulosa cells from both small and large bovine follicles.15,43,177 This is an example of non-competitive antagonism between leptin and insulin, demonstrated in a range of tissues (examined in greater detail below).
Expression of Ob-R has also been demonstrated in human granulosa cells.178 Furthermore, Lindheim et al179 showed that a significant increase in circulating leptin levels occurred during controlled ovarian hyperstimulation, suggesting that leptin plays a role in follicular growth and maturation.
To date, human and animal studies suggest that placental leptin is likely to affect maternal, placental and foetal function through both autocrine and paracrine mechanisms. It is possible that placental leptin may have physiological effects on the placenta including angiogenesis, growth and immunomodulation.175 Leptin may also be involved in regulation of foetal and uterine metabolism.
Leptin in fertility
The sterility of male and female ob/ob mice is a recognized feature of this mutation.12 Hoggard and co-workers46 found that reproductive hormone levels are reduced in ob/ob females, suggesting a functional defect in the hypothalamic–pituitary axis. Several independent reports have demonstrated that correction of leptin deficiency in ob/ob mice by peripheral injections of recombinant leptin activates the reproductive axis and restores fertility in both sexes. Barash and co-workers180 dosed ob/ob mice with human leptin (50 µg, i.p, twice daily) for 14 days. Leptin-treated females had significantly elevated serum levels of LH, increased ovarian and uterine weight, and stimulated aspects of ovarian and uterine histology compared to controls. Leptin-treated males had significantly elevated serum levels of FSH, increased testicular and seminal vesicle weights, greater seminal vesicle epithelial width, and elevated sperm counts compared with controls. The study of Chehab et al176 also showed that repeated administration of human leptin to female ob/ob mice corrected their infertility, thus restoring ovulation, pregnancy, parturition and lactation. The stimulatory effect of leptin on reproductive organs appears to have functional significance, as leptin-treated male and female ob/ob mice were able to mate successfully with their wild-type counterparts.176,181 Thus, leptin has an essential permissive role in the endocrinological processes which control fertility.
Furthermore, in vitro studies have revealed that oestrogens increase leptin mRNA expression.37,182 Testosterone, in contrast, inhibits leptin gene expression in vivo and in vitro,37,183 possibly through a direct suppressive effect. These findings suggest that androgens and oestrogens modulate leptin expression at the mRNA level through sex steroid receptor-dependent transcriptional mechanisms.37 Thus, there appear to be some counter-regulatory mechanisms between leptin and, particularly, the oestrogens. The study of Jockenhovel et al184 indicated that testosterone substitution normalized elevated serum leptin levels in hypogonadal men. These investigators have concluded that interaction of testosterone and leptin might be part of a hypothalamic–pituitary–gonadal–adipose tissue axis that is involved in body weight maintenance and reproductive function.
Role of leptin in pregnancy
This topic was recently reviewed by Henson and Castracane.185 Readers will find that the treatment of this topic in the present review has a different emphasis.
In females of most mammalian species, high leptin levels may signal the attainment of the sufficient long-term energy stores that are crucial for successful reproduction.174 During human pregnancy, leptin levels were elevated throughout compared with the non-pregnant state.42,186,187 Furthermore, in both humans42 and mice,174 serum leptin was particularly elevated during the second and third trimesters.
Both Ob-Ra and Ob-Rb were found in mouse, rat and human placenta using RT-PCR, in situ hybridization, Western blotting or immunohistochemistry.13,42,44,45,188 However, expression of leptin mRNA in rat placenta was much lower than that in the human placenta.13,189 Both the short and the long forms of leptin receptor were detected in the human placenta in early and in the full-term stages of gestation, using quantitative RT-PCR.190 Cioffi et al107 also detected transcription of the leptin receptor isoform termed Ob-R/B219 in human placenta using Northern blot and RT-PCR. In a recent study by Bodner et al45 Ob-R/B219 was not detected in human placenta cells, but was detected in blood cells of the intervillous space and foetal vessels. Using in situ hybridization and immunohistochemistry, they also found that Ob-Rb was expressed by the apical membrane of the syncytiotrophoblasts.
At present there appears to be a number of possible explanations for the increase in leptin levels in pregnancy: increased production by maternal fat; increased expression by the placenta; and increased levels of binding protein(s) (also of placental origin in the case of the mouse191), in the maternal circulation.46 However, hyperleptinaemia during pregnancy is not associated with decreased food intake or a decline in metabolic efficiency, as might be expected given one of the roles of leptin is as a satiety factor. Explanations for this may be a possible pregnancy-induced state of leptin resistance, or a change in leptin bioavailability.192,193 It is interesting to note that some authors believe that the increase in leptin concentration during pregnancy is counter-intuitive as this is a period of increased nutritional demands, and not one in which the actions of a satiety factor are expected to increase.185 Others have hypothesized that the soluble form of the leptin receptor mediates leptin actions during pregnancy as the circulating concentrations of this protein are increased, especially in the mouse,191,194 although much less so in humans.139 Given the range of interactions between leptin and other hormone axes, and the complex pattern of central and peripheral pathways invoked, we hypothesize that, if leptin does in fact have a role as a satiety factor, this role is centrally mediated and may not be significantly modulated by the soluble leptin receptor. Interactions such as those between leptin and insulin, which modulate oxidation of FFA and lipogenesis (see below) as part of the process which mobilizes energy reserves during pregnancy, are peripherally modulated and more influenced by the balance of energy reserve status and the demands of the foetus. Studies on the interaction of leptin binding protein status with plane of nutrition and energy reserve status during pregnancy will be required to confirm or refute this.
Holness et al193 suggest the possibility that leptin may be important for optimizing pregnancy outcome. Potential roles of leptin may include facilitation of endocrine responses to pregnancy, maintenance of maternal fuel homeostasis during a period of increased nutrient intake and requirement and/or optimization of foetal growth and development.
The relative contributions of foetal-derived leptin and maternal leptin to maternal total leptin has recently been described by Linnemann et al195 using an in vitro perfusion technique. This study suggests that only 1–2% of placental leptin enters the foetal circulation, while approximately 98% enters the maternal circulation. Although not yet determined, the interaction of leptin with insulin may also be important in pregnancy. Leptin appears to act as a permissive factor with respect to its effects on insulin-mediated fuel storage in muscle and in the liver and on insulin-mediated fuel utilization by muscle (see below). Thus, one of the actions of foetal- and placental-derived leptin may be to act on maternal liver and muscle, changing the dynamics of fuel utilization. Particularly in the case where maternal nutrition may become limiting, and thus maternal leptin reduced; foetal- and placental-derived leptin would act on the maternal metabolism to ensure that fuel is preferentially utilized by the foetus, rather than rebuilding maternal stores. This hypothesis is consistent with the observation that leptin concentration is higher in arterial than in venous cord blood (see below), and suggests that both foetal- and placental-derived leptin targets maternal metabolism.
To consider this interaction further, we hypothesize that, following parturition, and the accompanying sudden decrease in leptin in the maternal circulation, insulin-mediated fuel storage undergoes a rebound phase, in which maternal fuel stores are rapidly rebuilt. In some women, it may be that this rapid phase of increasing fuel storage is not counter-regulated, and thus begins a phase in which adipocyte number increases, predisposing the subject to increased fat-to-lean ratio and, in the longer-term, in the case of repeated pregnancies, obesity. These factors are also likely to be intricately linked with lactation as serum leptin concentrations in lactating women are lower post-partum, but somewhat higher than in non-lactating women, further demonstrating a role for leptin in mobilizing fuel reserves.196
Leptin expression in the foetus
The finding that leptin levels are increased in pregnancy42,174,186,187 has raised the possibility that leptin may play an important role in nutritional signalling between mother and the foetus during pregnancy. The expression of leptin by syncytiotrophoblasts13,44,45 suggests a role for leptin in nutrient transfer. Hoggard and co-workers13 studied leptin gene expression and localization of leptin protein in a number of tissues in the 14.5-day-old murine foetus by in situ hybridization and immunohistochemistry. High levels of leptin and its receptor expressed in foetal bone and cartilage suggested a role for leptin in bone or cartilage development. Leptin receptor mRNA and leptin receptor protein were also identified in leptomeninges and the choroid plexus of the foetal brain, and in lung, kidney, heart and liver.
Schubring and co-workers197 found that maternal serum leptin concentrations do not correlate with birth weight. In contrast, leptin concentrations in umbilical cord blood correlated positively with birth weight and placental weight in humans.197,198,199 Schubring et al197 demonstrated that leptin concentrations were higher in arterial cord blood than in venous cord blood, leading to the suggestion that leptin synthesis by foetal tissue may be greater than by placental tissue. However, evidence from Linnemann et al195 now suggests that foetal and placentally derived leptin may target maternal metabolism as outlined above. The high levels of expression of both leptin and leptin receptors in the placenta and foetus suggest that leptin plays a key role in foetal development. One possible role is that of a foetal growth factor, ie a signal to the foetus of the maternal energy status. Alternatively, as the concentration of leptin in arterial cord blood is higher than in venous cord blood, leptin may signal to the placenta the energy status of the foetus, and may have further actions on maternal metabolism, particularly if maternal energy stores become limiting.
Leptin and the onset of puberty
The weight hypothesis of the development of puberty states that, when body weight reaches a certain level, puberty occurs. If rats are underfed, puberty is delayed, but with access to food rapid weight gain leads to onset of puberty.200 It has been shown that exogenous leptin accelerates the onset of puberty in normal female mice174,201 and normal prepubertal female rats.202 From this evidence it may be speculated that puberty is induced when fat stores reach a certain level, increasing the release of leptin from adipocytes into the bloodstream. Leptin may then act on hypothalamic cells to stimulate release of LHRH, thereby triggering gonadotropin release. The subsequent release of follicle-stimulating hormone (FSH) and LH stimulates gonadal steroid secretion, leading to development of the reproductive tract and induction of puberty.
However, several studies203,204,205 have demonstrated contrasting data which indicate that the timing of the onset of puberty in male Rhesus monkeys is not triggered by rising circulating leptin concentrations. Given the conflicting reports in the literature,174,205,206 it is likely that leptin is a factor permissive to the onset of puberty, and that, although some minimal threshold level of leptin is necessary to pubertal development, leptin is more likely not sufficient to initiate puberty. The interactions between leptin and oestrogens and androgens outlined above indicate that leptin's role in the onset of puberty is mediated by different mechanisms in males and females and that further work is required to elucidate these differences. The notion that leptin is a permissive factor is supported by data which show that leptin is also expressed in skeletal muscle tissue.10,17,99,107 Thus, leptin may act as an indicator of general well-being: that sufficient muscle mass has developed (and is in a phase of anabolic growth) and that a commensurate level of body fat has been deposited, such that it is favourable for puberty (and its consequences, such as pregnancy) to proceed.
Leptin and insulin
Insulin is an important regulator of energy homeostasis. It stimulates glucose, free fatty acid and amino acid uptake by tissues and tissue anabolism. It is not surprising that a link between leptin and insulin should exist in the regulation of energy homeostasis.
Insulin effects on leptin
Current knowledge suggests that insulin plays a chronic role in the regulation of leptin gene expression and production by WAT. Some studies have shown that hyperinsulinaemia increased plasma leptin concentrations and gene expression in WAT in both rodents and humans.207,208,209,210,211,212,213,214,215,216 Recent studies have shown that humans and rodents with insulinoma also have increased plasma leptin and leptin mRNA, and plasma leptin concentrations returned to normal after removal of the insulinoma.217,218,219 Streptozotocin-induced insulin deficiency results in reduced circulating concentrations of leptin and mRNA, and that suppression was rapidly reversed by treatment with insulin.71,220 Insulin appears to act directly at the level of the adipocyte by increasing leptin secretion and gene expression, perhaps due to increased glucose transport and metabolism.221 An in vivo study222 has shown that fa/fa rats have 40-fold higher leptin concentration compared with lean littermates, but euglycaemic hyperinsulinaema increased plasma leptin only in lean rats. The authors suggested that insulin may be a regulator of in vivo leptin secretion by adipose tissue of lean rats, whereas it is inactive in increasing plasma leptin in obese fa/fa rats. However, this lack of sensitivity may simply be a consequence of peripheral insulin resistance, a characteristic of this phenotype.
Leptin effects on insulin secretion and tissue sensitivity to insulin
In 1996, Kieffer et al102 reported that leptin receptors are expressed in the insulin producing β-cells within the pancreatic islets, suggesting that leptin might influence insulin secretion through a direct action on these cells. This hypothesis was investigated by several others,223,224,225,226,227 with apparently conflicting results.
Leptin has been shown to inhibit insulin secretion from isolated rodent217,225,227,228,229 and human217,230 islets, perfused rodent pancreas224,228 and in vivo in mice217 (Table 3). Emilsson et al228 demonstrated that supraphysiological concentrations of recombinant leptin (100 nM) inhibited insulin secretion either in perfused pancreas preparations with 5.6 mM glucose, or in isolated β-cells from ob/ob mice incubated in the presence of a high concentration (16.7 mM) of glucose. In contrast, Ceddia et al231 also indicated that in the presence of low (2.8 mM) or physiological (5.6 mM) glucose concentrations, leptin (50 nM) increased insulin secretion by perfused isolated islets. Similarly, Tanizawa231,232 showed that leptin stimulated insulin secretion in islets. However, there are also conflicting reports in the literature. Some studies using an isolated perfused rat pancreas model233,234 and isolated rodent islets217,235 found that leptin did not affect insulin secretion (Table 3). According to Leclercq-Meyer and co-workers234 normal physiological concentrations of leptin (1 nM) failed to influence the release of either insulin or glucagon in the isolated perfused rat pancreas.
It is important to note that the normal physiological range of plasma leptin concentrations is of the order of 1–10 ng/ml or approximately 0.06–0.6 nM. Furthermore, most of the in vitro studies reviewed here have found effects only at supraphysiological concentrations of leptin and thus these results should be considered with this caution. However, we have shown in saturation binding studies that leptin binds to abundant hydrophobic sites (Margetic et al, submitted), and it may be that, to obtain a biologically relevant concentration of leptin, sufficient leptin must be added to the system to saturate hydrophobic sites on the apparatus and in the tissues in the in vitro system. Furthermore, Cawthorne et al236 and LA Campfield (op. cit.) have suggested that recombinant leptin may be less potent than the native form. Thus, further investigation will be required to confirm or refute the effects of these factors.
In isolated rat adipocytes, (Table 4) Walder et al237 found that 125I-insulin binding was dose-dependently inhibited by leptin (1 nM to 2.5 µM). The study of Muller et al238 shows a range of effects in isolated rat adipocytes consistent with leptin inhibition of insulin binding. Leptin clearly impaired insulin-mediated stimulation of glucose transport, glycogen synthase activity, lipogenesis, inhibition of isoprenaline-induced lipolysis, protein kinase A activation and protein synthesis in a dose-dependent manner. These effects appeared to require preincubation with leptin for at least 2 h. Thus, in another study using a short leptin incubation (30 min), no leptin effects on insulin-mediated glucose uptake were observed.239 Similarly, Ranganathan et al240 found a negative result; however, this study incubated adipocytes at a sub-physiological glucose concentration, making interpretation of their findings more difficult. Thus, leptin inhibition of insulin-mediated glucose transport appears to be a direct (peripheral) relationship in adipocytes, however it appears to be centrally controlled in muscle tissue (see below).
Other studies have found a direct link between plasma leptin concentration and measures of insulin sensitivity independent of the degree of body fatness.223,241,242 High levels of leptin were associated with insulin resistance independent of BMI. In humans, it is still not clear whether the hyperinsulinaemia that accompanies insulin resistance is associated with the higher leptin levels observed in some obese subjects with non-insulin-dependent diabetes mellitus (NIDDM)243 or polycystic ovary syndrome (PCOS).244
Whether leptin plays a role in the short-term response to food ingestion and/or macronutritient composition remains a matter for debate. It appears, however, that in ruminants in which nutrient flux from the rumen and small intestine (into the circulation) remains much more constant than in monogastrics, that leptin is not involved in signalling short-term changes to food intake.245 Furthermore, the mechanism underlying the contribution of leptin to insulin secretion as well as the physiological impact of leptin's nutrient stimulation effect remains to be studied in more detail.
Several studies have suggested that glucose is also an important regulator of leptin expression and secretion.246,247 Kamohara et al248 have demonstrated that i.v. (1 µg/h) and i.c.v. (5 ng/h) administered leptin into wild-type mice increased glucose turnover and glucose uptake, independent of increases in plasma insulin. However, given that leptin modifies insulin sensitivity of muscle and liver to glucose uptake (see below), probably via CNS control, it appears likely that glucose is regulated and is not a regulator.
Leptin and insulin sensitivity of skeletal muscle
It is worthy of note that skeletal muscle accounts for a large proportion of insulin-stimulated glucose uptake and whole body lipid oxidation and is the major contributing tissue to resting metabolic rate. Thus, muscle may be an important tissue for the expression of the juxtaposed interactions of leptin and insulin.
The recent study by Wang and co-workers17 found that leptin mRNA was expressed in skeletal muscle in Sprague–Dawley rats. The presence of Ob-Ra and Ob-Rb in mouse skeletal muscle has been reported previously.10,99,107 Although only 8% of total leptin receptor mRNA expression in skeletal muscle is the long form, both isoforms Ob-Ra and Ob-Rb appear to be able to mediate signal transduction.99,136
At the intracellular level, evidence for the intersection of the leptin and insulin pathways is supported as the effects of leptin treatment on muscle may be partially attenuated by synthetic blockade of phoshphatidylinositol (PI) 3-kinase activity.249 Furthermore, in hepatic cell models, leptin had profound effects on tyrosine phosphorylation of IRS-1, with no associated effect on insulin receptor phosphorylation;250 and, in addition, leptin-mediated recruitment of PI-3 kinase to IRS-2 has been shown. However, in this study the activity at PI-3 kinase did not translate to increased tyrosine phosphorylation of IRS-2.148 Most recently it was reported that leptin directly stimulates fatty-acid oxidation in muscle by activating the 5′-AMP-activated protein kinase, an enzyme that phosphorylates and subsequently inactivates CoA carboxylase.251 Thus a direct mechanism for the action of leptin on fatty acid oxidation consistent with the work of Muoio et al249,252 has now been defined.
Effect on glucose transport
Muoio et al252 showed that leptin (10 µg/ml) did not change insulin-stimulated metabolism in isolated mouse (female C57BL/6J mice) soleus and exterior digitorum longus (EDL) muscle. In agreement with this finding Ranganathan et al240 showed that leptin (0.5 µg/ml, for 24 h) had no effect on basal and insulin (10 nM)-stimulated glucose transport in cultured rat and human skeletal muscle cells. However, Burcelini et al253 found that i.v administration of leptin (40 ng/g/h) for 6 h to ob/ob mice (which are insulin-resistant) increased glucose turnover and stimulated glucose uptake in BAT, brain and heart. No increase in glucose turnover was observed in skeletal muscle or WAT. Kamohara et al248 suggested that leptin may acutely increase skeletal muscle glucose uptake by an insulin-independent mechanism. Infusion of murine leptin (1 µg/h) for 5 h into the femoral vein of conscious C57BL/6J wild-type mice led to an increase in glucose turnover and 2-deoxyglucose uptake into skeletal muscle and BAT increased whole-body glucose turnover and increased glucose oxidation, despite no change in plasma insulin or glucose concentrations. Furthermore, the leptin-induced increase in glucose uptake into edl and soleus muscles was attenuated by denervation. Taken together, all of these data indicate that the effects of acute leptin administration on glucose metabolism in muscle are mediated centrally.
Effect on lipid partitioning
Muscle plays an important role in clearance of serum free fatty acids (FFA) and triacylglycerol (TAG) and in whole-body FFA oxidation. In contrast to their findings of a lack of leptin-mediated glucose uptake by isolated skeletal muscle, Muoio et al252 found that leptin (10 µg/ml) increased FFA oxidation in isolated mouse soleus muscle by 42%, whereas insulin (10 mU/ml) decreased soleus muscle FFA oxidation by 40%. When both hormones were administered, leptin attenuated both the antioxidative and the lipogenic effects of insulin by 50%. Less pronounced hormonal effects were observed in mouse edl. Also, their study showed that the effect of leptin on lipid partitioning was dose-dependent in both soleus and edl (Table 4). These data are consistent with the findings of Solini et al,254 who showed that, during euglycaemic, hyperinsulinaemic clamp, lipid oxidation was higher in obese than in lean subjects. However, this study also found that during the clamp glucose turnover was lower in obese subjects. One interpretation of these data links the apparent differing responses of muscle and adipose tissue to the interaction between leptin and insulin: obese subjects, having proportionately greater fat mass, may show a direct peripheral effect of leptin suppressing insulin-mediated glucose transport in adipose tissue, while the lipid oxidation effects are mostly manifested in muscle tissue. On the other hand, in lean subjects, having much lower fat mass, most of the effects on turnover are manifested in muscle tissue, and the relativity of the circulating concentrations of leptin and insulin, being lower, would favour lower FFA oxidation in muscle of lean compared to obese subjects. This relationship would be further complicated by the generally increased insulin resistance of obese subjects.
Muoio et al249 also studied the acute (60–90 min) effects of leptin and insulin on glucose and oleate metabolism in muscles isolated from lean and obese ob/ob mice with insulin resistance. In ob/ob soleus, leptin decreased glycogen synthesis by 36–46%, increased oleate oxidation by 26%, decreased oleate incorporation into TAG by 32%, and decreased the oleate partitioning ratio by 44%. Insulin decreased oleate oxidation by 31%, increased oleate incorporation into TAG by 46%, and increased the partitioning ratio by 125%. Addition of leptin (10 µg/ml) diminished insulin's antioxidative, lipogenic effects, recapitulating this group's earlier findings in isolated, normal mouse muscle.252 Thus, leptin opposes insulin's promotion of TAG accumulation in lean and ob/ob muscle, and indicates that skeletal muscle is critical in mediating leptin's effects on fuel homeostasis, weight regulation and adiposity. Importantly, these studies indicate that leptin attenuates insulin's antioxidative, lipogenic actions on muscle FFA metabolism via a peripheral mechanism, while leptin mediation of insulin-stimulated glucose disposal appears to occur via a central mechanism. Furthermore, as skeletal muscle represents approximately 40% of the total body weight,255 leptin-stimulated FFA oxidation in skeletal muscle may represent a very important anti-obesity mechanism independent of the central nervous system.
Leptin interaction with insulin-mediated glycogen synthesis
Defective glycogen synthesis in muscle is one of the earliest manifestations of insulin resistance. Liu et al256 found that recombinant mouse leptin at 10 and 100 nM inhibited glycogen synthesis by 35 and 45%, respectively, in soleus muscle of ob/ob mice in the presence of insulin (100 µU/ml). Leptin did not have an inhibitory effect on glycogen synthesis in soleus muscle from either wild-type (Table 4) or db/db mice. Ob-Rb was expressed in soleus muscle of both ob/ob and wild-type mice with no detectable difference in expression level, suggesting that the ob/ob leptin receptor may be linked to a more sensitive down-stream pathway. In contrast, two other studies have shown that leptin increased glycogen synthesis in cultured C2C12 muscle cells.257,258 A possible factor contributing to these contradictory results may be the difference between muscle cell lines and freshly isolated skeletal muscle. C2C12 cells represent an homogeneous cell population, whereas skeletal muscle is heterogeneous with respect to fibre and cell composition. Thus, the effect in C2C12 cells may not be truly representative of leptin effects on muscle in vivo.
In contrast, leptin actions on insulin binding in rat liver (a major storage depot for glycogen) were studied by Nowak et al.259 These workers found that administration of a single dose of mouse leptin (1 and 2 nM) to rats decreased Bmax but did not change the affinity of insulin for its high affinity binding sites in the liver. These results show that leptin may antagonize some functions of insulin by the attenuation of insulin receptor capacity. Furthermore, this action of leptin is consistent with the hypothesis that leptin tends to mobilize body fuel reserves and increase energy expenditure. However, Zhao and co-workers260 showed that leptin, like insulin, induced an intracellular signalling pathway in primary rat hepatocytes, probably mediated through the short forms of the leptin receptor. This pathway involves stimulation of phosphatidylinositol 3-kinase (PI3K) binding to insulin receptor substrate-1 and insulin receptor substrate-2, activation of PI3K and protein kinase B, and PI3K-dependent activation of cyclic nucleotide phosphodiesterase 3B, a cAMP-degrading enzyme. Leptin also suppressed glucagon-induced cAMP elevation in a PI3K-dependent manner. Indeed, Ahren and Havel261 suggested that cAMP-protein kinase A signal transduction pathway is a target for leptin to inhibit insulin secretion in insulin-producing cells.
Leptin interaction with insulin-mediated hexosamine synthesis
Several studies have demonstrated a clear interaction between intracellular hexosamine concentrations (in muscle and in fat) and insulin resistance, in the rodent model, using in vivo17,262,263,264,265 and in vitro17,266 methods. This interaction has also been demonstrated in vitro in both human myogenic cells267 and adipocytes.262 Importantly, in the rodent model, an increase in leptin gene expression was associated with insulin resistance and intracellular glucosamine concentrations.17,266,268
The hexosamine biosynthetic pathway is the primary source of the substrates for the glycosylation of (glyco-) proteins,269 and is hypothesized to be a cellular ‘sensor’ of energy availability.17,262,263 However, quantitatively, only 1–3% of incoming glucose ultimately enters the hexosamine pathway.264,269 An alternative perspective to this proposed model is that, in vivo, it is increased leptin which signals excess energy stores and that leptin-mediated damping of insulin actions (particularly in muscle) on both glucose and in this case hexosamines reduces the activity in these storage pathways. In this alternative model, as leptin concentration increases in response to increasing fat mass (muscle has already become loaded with glycogen and hexosamines), the actions of insulin on glucose and hexosamine storage become increasingly damped. Simultaneously, TAG and thus, fatty acids are preferentially utilized as an energy source. This leads to a brief discussion of leptin interaction with insulin resistance.
Leptin and insulin resistance in muscle
A vast majority of studies of insulin resistance concentrate upon insulin interactions with glucose, while less attention is paid to insulin-mediated fatty acid uptake, for which mechanisms are now becoming clarified.270,271,272 However, the inter-relationship of the classes of substrate utilized for energy, as first proposed by Randle,273 has been long established. Although insulin actively increases free fatty acid uptake by muscle,270,271,272 it clearly inhibits oxidation of free fatty acids, an action which is suppressed by leptin.249,252,274,275 This partitioning of insulin actions, some of which appear to be synergistic with leptin actions, such as increased uptake of free fatty acids by a direct mechanism and uptake of glucose possibly via a CNS-mediated mechanism (above), and others of which may appear to be antagonistic to leptin actions, such as inhibition of insulin-mediated suppression of fatty acid oxidation,249,252,274,275 are difficult to resolve into a single model. However, it may be stated in broad terms that insulin is anabolic and tends to increase storage of fuels and amino acids, while leptin appears to mobilize TAG and modulate preferential utilization of fatty acids. Simultaneously, leptin inhibits insulin-mediated mechanisms of long-term glucose storage and oxidation. It is this sub-set of insulin actions upon which many investigations of insulin resistance have focussed, and it may be that the typical, many-fold increases in circulating leptin concentrations observed in obesity have profound effects only upon glucose-metabolism-linked peripheral insulin resistance through the mechanisms proposed here. Thus, more thorough investigations of the interaction of leptin and insulin in regulating both glucose and fatty acid oxidation are of paramount importance in clarifying the progression of obesity and the development of peripheral insulin resistance.
Leptin and glucocorticoids
Adipocyte culture studies have shown that glucocorticoids stimulate leptin gene expression.276,277 Glucocorticoid secretion is linked with meal times in normal humans and rodents,278 and is increased in Cushing's disease or by glucocorticoid administration to normal volunteers.279 In humans, glucocorticoids stimulate leptin gene expression and secretion independently of effects on food intake, although increases in insulin or lipogenesis associated with food intake may contribute to leptin production.280,281 The recent study of Solano and Jacobson282 indicated that leptin infusion (0.5 µg/h) significantly decreased food intake and body weight in adrenalectomized mice. Glucocorticoid replacement increased food intake without reversing leptin inhibition of hypothalamic neuropeptide Y (NPY) mRNA levels. These investigators have concluded that glucocorticoid levels within the physiological range interfere with leptin action and that glucocorticoid effects are at least partly independent of NPY. Dexamethasone administration (10 mg/day) in obese humans caused increases in leptin levels from 16.3 ng/ml at baseline to 27.6 ng/ml after 5 days of treatment.280 Similar stimulatory effects were also obtained in obese children with a lower dose (1 mg), although the effects were less pronounced.283
The mechanism(s) of glucocorticoid stimulation of plasma leptin is still unknown. Wabitsch and co-workers277 have postulated that glucocorticoids may influence leptin gene expression directly and independently of their differentiation-promoting effects on adipocytes, or they can induce changes in plasma insulin sensitivity. De Vos et al have shown similar effects on leptin expression in rats,284 however, a later study by the same group285 indicates that the leptin gene promoter region does not contain a binding site for the glucocorticoid receptor, thus the effect does not rely on the classical molecular mechanism of glucocorticoid receptor action. However, another possible mechanism includes modulation via the CNS, mediated by NPY.286
Leptin interaction with the growth hormone axis and with other cytokines
Although most of the interactions of leptin with growth hormone are likely to be centrally mediated, interactions of leptin with major components of the growth hormone axis, such as insulin-like growth factor-1 (IGF-1), appear to be more direct (see below).
Leptin and growth hormone
Several groups have reported that leptin regulates growth hormone (GH) secretion in humans,287,288 rodents,289,290,291 sheep,292,293,294 and pigs.295 However, when isolated rat adipocytes were incubated with either GH or insulin-like growth factor-1 (IGF-1) alone, there was no effect on leptin secretion.74 Recently, Houseknecht et al296 have extended the work of Hardie et al74 to include incubation of bovine adipose tissue explants with insulin, dexamethasone and GH alone and in combination. After 24 h incubation, GH alone had no effect on leptin expression in bovine subcutaneous adipose tissue. However, when incubated in combination, GH attenuated the insulin or dexamethasone stimulation of leptin expression. In addition, the same investigators reported apparently contrasting results in an in vivo experiment as part of the same study. This study used young castrated male cattle treated with GH for 3 days, and found that GH treatment increased adipose tissue leptin and IGF-1 mRNA concentrations. Furthermore, leptin levels were highly correlated with adipose tissue IGF-1 mRNA in GH-treated animals. In contrast, other in vivo studies have shown an inhibitory effect of GH on leptin gene expression.297,298 Each of these studies was conducted in a growth hormone-deficient model, growth hormone-deficient children and fatty Zucker rats, respectively. These results are consistent with the in vitro studies cited above.296,299,300
Leptin and IGF-1
Studies of the effects of IGF-1 on serum leptin and leptin expression are also contradictory. Isozaki et al298 showed that IGF-1 did not change the percentage body fat or leptin mRNA in visceral fat tissue in Zucker rats, while Boni-Schnetzler and co-workers301 found that IGF-1 decreased leptin mRNA in epidydimal fat tissue of hypophysectomized rats. Boni-Schnetzler et al302 in a recent study showed that 6 days subcutaneous (s.c.) administration of human IGF-1 (1 mg/day) decreased leptin mRNA in epididymal fat pads in normal rats (by 38.8%) and reduced serum leptin (by 51.6%) in comparison to control rats. Fat pad weight was also reduced (by 60.3%). These authors concluded that the effect of IGF-1 on the reduction of fat pad mass is most likely due to a suppression of insulin secretion, and thus enhancement of fat mobilization and fatty acid (FA) oxidation. As a consequence, adipose tissue leptin mRNA and serum leptin concentrations decrease as well.
Leptin and other cytokines
The influence of cytokines on leptin mRNA expression and circulating concentrations has been investigated in human subjects.303,304,305,306 Interleukin-1 (IL-1) was found to induce leptin levels directly or indirectly, by increasing the activity of the hypothalamic–pituitary axis.303,307 Mantzoros and co-workers306 reported a positive and independent association between tumour necrosis factor-α (TNF-α) levels and circulating leptin concentrations, suggesting that TNF-α may directly induce leptin gene expression in humans, as it does in rodents.307,308 These studies may implicate a role for leptin in the pathogenesis of cachexia that is accompanied by increased levels of cytokine in advanced stages of AIDS and cancer.309
Interaction of leptin with other metabolites and hormones
Recent studies have shown that free fatty acids (FFA) decrease leptin mRNA levels in adipocytes,22,310 suggesting that FFA are involved in the regulation of leptin production in adipocytes. However, regulation of FFA in vivo involves complex mechanisms mediated by a range of hormones including insulin, growth hormone and leptin. Thus, it is likely that hormones mediating FFA are the drivers of this interaction.
Stumvoll et al85 showed that, in humans, isoprenoline acutely suppressed leptin levels independently of increased FFA, indicating a direct effect via β-adrenergic receptors on leptin secretion. These authors concluded that a signalling mechanism other than that mediated by cAMP must be involved in modulating leptin secretion.
There is conflicting evidence as to the effects of thyroid hormone on leptin production, with suggestions that tri-iodothyronine (T3) has inhibitory,311 stimulatory312 or no effect313,314 on leptin levels. The study of Fain et al311 showed that administration of T3 to hypothyroid male rats resulted in a 40% decrease in leptin mRNA at 8 h. This decrease in leptin mRNA was associated with a parallel decline in circulating leptin levels of about 50% at 24 h. In contrast, Sesmilo et al314 investigated patients with hypothyroidism and hyperthyroidism returning to normal thyroid function. Hypothyroid patients were treated with L-T4 and hyperthyroid patients with methimazole. This study showed that there appeared to be no correlation between serum leptin and thyroid hormone (free thyroxine, thyrotrophin and T3) levels at any stage during the study.
Although the proposed roles of leptin in satiety and energy expenditure were originally thought to be centrally focussed, more recent studies have begun to reveal direct effects of leptin in the periphery and via interactions with other peripherally acting hormones, one of the most important of which is insulin.
It is clear that one of the principal roles of leptin is as a lipostat, signalling to other systems the energy reserves available to the body, mediating fuel utilization and mobilization and, consequently, energy expenditure. In some cases, this extends to signalling that the (juvenile) body has sufficient reserves to move into a potentially more metabolically demanding phase—puberty. Puberty in both humans and animals potentially leads to secondary consequences which are metabolically demanding. In the female, pregnancy, lactation and nurture and, in the male, competitive behaviours in the quest for a suitable mate(s) and paternal duties. In this context leptin's actions are complex and mediated both centrally and peripherally.
During pregnancy, leptin production by the foetus and the placenta may signal foetal nutrient status. However, it may also provide a mechanism whereby maternal fuel reserves are more readily mobilized, favouring utilization by the foetus rather than building maternal reserves such as glycogen in muscle and in the liver. In the case where maternal leptin concentrations are low owing to limiting nutrition, the proportional importance of foetal and placental leptin in mobilizing fuel for use by the foetus may be greater.
Obesity is a state characterized by hyperleptinaemia. Although plasma leptin concentration is correlated with level of obesity, many other factors interact. These include developmental stage, sex, pregnancy, multiple hormonal factors, present nutritional status and nutritional history of the individual. Thus, leptin is not simply an indicator of proportional body fat or lean mass. In some in vitro studies, in which leptin has been used at concentrations typical of plasma concentrations in obese subjects, leptin has had some unexpected effects which may be mediated through a mechanism independent of the Ob-R. This observation may suggest a pathophysiological effect of hyperleptinaemia associated with obesity. Clearly, this mechanism is not at work in all cell types, as many studies using supraphysiological concentrations of leptin have reported no aberrant effects.
Leptin is still only a newly discovered hormone. Despite extensive investigations, worldwide since 1994, we have much to learn about the leptin axis and its interactions both centrally and in the periphery.