Professional Documents
Culture Documents
Obesity and Metabolism
Obesity and Metabolism
Obesity and Metabolism
Series Editor
Volume Editor
Bibliographic Indices. This publication is listed in bibliographic services, including Current Contents® PubMed/MEDLINE.
Disclaimer. The statements, options and data contained in this publication are solely those of the individual authors and
contributors and not of the publisher and the editor(s). The appearance of advertisements in the book is not a warranty,
endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the
editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products
referred to in the content or advertisements.
Drug Dosage. The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this
text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research,
changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader
is urged to check the package insert for each drug for any change in indications and dosage and for added warnings and
precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug.
All rights reserved. No part of this publication may be translated into other languages, reproduced or utilized in any form or by
any means electronic or mechanical, including photocopying, recording, microcopying, or by any information storage and
retrieval system, without permission in writing from the publisher.
© Copyright 2008 by S. Karger AG, P.O. Box, CH–4009 Basel (Switzerland)
www.karger.com
Printed in Switzerland on acid-free and non-aging paper (ISO 9706) by Reinhardt Druck, Basel
ISSN 0301–3073
ISBN 978–3–8055–8429–6
Contents
VII Foreword
Grossman, A.B. (London)
IX Preface
Korbonits, M. (London)
VI Contents
Foreword
It has now become a truism that obesity is one of the most serious medical problems of
the developed world, and however unsettling this may be in the presence of famine and
starvation in parts of the developing world, it is nevertheless a situation which must be
faced. It has become difficult for endocrinologists to access all of the recent literature on
the genetics, metabolic phenotype and treatment of obesity, and Dr. Korbonits has been
stunningly successful in putting together an array of chapters from many of the fore-
most authorities and researchers in this field. With recent data indicating that the surgi-
cal therapy of gross obesity is associated with increased longevity, it is now an ideal time
to reveal what we know about the causes, metabolic disturbances and treatment of this
so common condition. I recommend this volume to all who are beginning research in
this area or are responsible for the clinical care of obese patients: this means the volume
will be relevant to almost all practising clinicians in 2008.
It was the discovery of leptin in 1994 that completely changed endocrinology’s atti-
tude to obesity. What had until then been a niche subject with little known patho-
physiological pathways, was, by the discovery of a hormone so profoundly affecting
weight, suddenly launched into the fast lane of scientific research.
In the current volume we have attempted to summarise key advances in many facets
of obesity: in the field of genetics, from the spectacular monogenic and syndromal
causes to the less dramatic but more common susceptibility genes, which have only
been recently identified; we have explored the effects of obesity in the pregnant mother,
in foetal life, in childhood and in old age, and have attempted to draw conclusions from
studies of periodic increased weight in animals. We have also mapped out the biochem-
ical and physiological background of the abnormal metabolism in obesity, by scrutiniz-
ing the hormones and enzymes most recently implicated in the development,
maintenance and consequences of obesity. The ‘traditional’ hormonal causes of obesity
are discussed, as they may occasionally cause a differential diagnostic challenge, and we
offer a practical update on clinical approach and treatment of obesity. Finally, we have
attempted to reflect the social aspects of obesity in society, and the view of the obese
body in art throughout the centuries. While a completely comprehensive overview of
the metabolism of obesity is beyond the scope of this book, we have aimed for a timely
and wide-ranging update which we believe will be both interesting and pertinent to all
endocrine clinicians and researchers.
Márta Korbonits, London
Korbonits M (ed): Obesity and Metabolism.
Front Horm Res. Basel, Karger, 2008, vol 36, pp 1–11
Abstract
We and others have identified several single gene defects that disrupt the molecules in the leptin-
melanocortin pathway causing severe obesity in humans. In this review, we consider these human
monogenic obesity syndromes and discuss how far the characterisation of these patients has informed
our understanding of the physiological role of leptin and the melanocortins in the regulation of human
body weight and neuroendocrine function. Copyright © 2008 S. Karger AG, Basel
The genetic contribution to body weight has been established through family studies,
investigating parent-offspring relationships, the study of twins and adopted children
[1, 2]. These studies consistently report heritability estimates of 40–70% [3]. As is the
case for height, where nutritional changes in the last 50 years have contributed to sub-
stantial increases in mean final height in many populations, environmentally-driven
changes in body weight in the population occur against a background of susceptibil-
ity to weight gain that is determined by genetic factors. Thus, genetic approaches can
be applied to understand both the molecular and physiological mechanisms involved
in human obesity.
We have explored the genetic basis of severe childhood obesity where we consid-
ered that major and more highly penetrant genetic effects were likely to be found. In
1997, we established the Genetics of Obesity Study (GOOS) to recruit patients with
severe obesity (body mass index standard deviation score, BMI SDS, ⬎3) of early
onset (⬍10 years). We were particularly interested in children with a strong family
history of obesity and those from consanguineous families. Our intention was to use
a candidate gene approach to look for mutations in genes thought to play a role in the
regulation of body weight based on evidence primarily from rodent models at the
time. With the help of colleagues throughout the world, we have to date recruited
over 2,500 patients to the GOOS cohort. In the past 9 years, several human disorders
of energy balance that arise from genetic defects have been described by ourselves
and others [4]. All of these are in molecules identical or similar to those known to
cause obesity in genetic and experimental syndromes of obesity in rodents and all
have been identified using a candidate gene approach. These mutations all result in
severe obesity in childhood without developmental pleiotropic features.
In 1997, we reported two severely obese cousins from a highly consanguineous fam-
ily of Pakistani origin [5]. Both children had undetectable levels of serum leptin and
were found to be homozygous for a frameshift mutation in the LEP gene (⌬G133),
which resulted in a truncated protein that was not secreted. We have since identified
5 further affected individuals from four other families [6, 7; unpubl. obs.] who are
also homozygous for the same mutation in the leptin gene. All the families are of
Pakistani origin but not known to be related over five generations. A large Turkish
family in which 3 adults carry a homozygous missense mutation (C→T substitution
at codon 105 resulting in Arg→Trp) in the LEP gene have also been described [8].
The first mutation in the leptin receptor gene was published in 1998 [9]. The muta-
tion was found in homozygous form in 3 severely obese adult siblings from a consan-
guineous family of Algerian origin. This mutation results in abnormal splicing of
leptin receptor transcripts and generates a mutant leptin receptor that lacks both trans-
membrane and intracellular domains. The mutant receptor circulates at high concen-
trations bound to leptin, resulting in very elevated serum leptin concentrations [10].
We recently sequenced the leptin receptor gene in a cohort of patients with severe,
early onset obesity in the absence of developmental delay and identified 8 unrelated
probands with homozygous or compound heterozygous loss of function mutations in
the leptin receptor gene [11]. The prevalence of pathogenic leptin receptor mutations
in this cohort was 3%. Six of the probands were from consanguineous families but 2
probands (including the compound heterozygote) were UK Caucasians whose parents
were unrelated. Although the prevalence of leptin receptor mutations is likely to be
higher amongst ethnic groups where consanguinity is common, leptin receptor defi-
ciency should be considered in all patients with hyperphagic obesity of early onset.
The clinical phenotypes associated with congenital leptin and leptin receptor defi-
ciencies are similar. Leptin and leptin receptor deficient subjects are of normal birth
weight but exhibit rapid weight gain in the first few months of life resulting in severe
obesity [6]. Body composition measurements show that leptin deficiency is charac-
terised by the preferential deposition of fat mass giving a distinct clinical appearance
with excessive amounts of subcutaneous fat over the trunk and limbs [6]. All patients
2 Farooqi
were hyperinsulinaemic, consistent with the severity of obesity and some adults have
developed type 2 diabetes in the 3rd–4th decade [6].
All subjects in these families are characterised by intense hyperphagia with food
seeking behaviour and aggressive behaviour when food is denied [6] and energy
intake at an ad libitum meal is markedly elevated [11].
In leptin deficient humans we found no detectable changes in resting metabolic
rate using indirect calorimetry or total energy expenditure using chamber calorime-
try [12]. However, Ozata et al. [13] reported abnormalities of sympathetic nerve func-
tion in leptin deficient adults consistent with defects in the efferent sympathetic limb
of thermogenesis.
Leptin and leptin receptor deficiency are associated with hypothalamic hypothy-
roidism and hypogonadotropic hypogonadism. Complete leptin deficiency is associ-
ated with a moderate degree of hypothalamic hypothyroidism characterised by low free
thyroxine and high serum thyroid-stimulating hormone (TSH) which is bioinactive. In
leptin deficient children, plasma free thyroxine concentrations are within the normal
range, but 4 children had significantly elevated TSH levels [6] and the pulsatility of
TSH secretion, studied in a single adult with congenital leptin deficiency, was charac-
terised by a markedly disorganised secretory pattern [14]. Two subjects homozygous
for a non-sense mutation in the leptin receptor were diagnosed with hypothyroidism in
childhood and thyroid hormone replacement therapy commenced [9].
Normal pubertal development does not occur in adults with leptin or leptin recep-
tor deficiency, with biochemical evidence of hypogonadotropic hypogonadism [8].
However, there is some evidence for the delayed but spontaneous onset of menses in
leptin and leptin receptor deficient adults [11, 13].
Leptin and leptin receptor deficient children have normal linear growth in child-
hood and normal IGF-1 levels. However, because of the absence of a pubertal growth
spurt the final height of adult subjects is reduced. In the first reported leptin receptor
deficient family, short stature and abnormal serum levels of GH, IGFBP3 were noted
in childhood. However, assessment of the GH/IGF axis is difficult in obese children
and adults as obesity itself is associated with abnormalities in basal and dynamic tests
of the GH/IGF axis. We conclude that while impaired linear growth has been
reported in some cases of LEPR deficiency, this does not appear to be a common
characteristic of this disease [11].
We demonstrated that children with leptin deficiency had profound abnormalities of
T cell number and function [6], consistent with high rates of childhood infection and a
high reported rate of childhood mortality from infection in obese Turkish subjects [13].
We have reported the dramatic and beneficial effects of daily subcutaneous injections
of recombinant human leptin leading to a reduction in body weight and fat mass in
Fig. 1. Effects of recombinant human leptin treatment in leptin deficiency. a A 3-year-old boy
weighing 42 kg. b The same boy at 7 years of age weighing 32 kg.
3 congenitally leptin deficient children [6, 12] (fig. 1). All children showed a response to
initial leptin doses that were designed to produce plasma leptin levels at only 10% of
those predicted by height and weight (i.e. approximately 0.01 mg/kg of lean body mass).
Leptin therapy has also been successfully used in the 3 Turkish leptin deficient adults
[15].
The major effect of leptin was on appetite with normalisation of hyperphagia.
Leptin therapy reduced energy intake during an 18MJ ad libitum test meal by up to
84% [6]. Leptin treatment was associated with reduced hunger scores with no change
in satiety in adults with leptin deficiency [15]. We were unable to demonstrate a
major effect of leptin on basal metabolic rate or free-living energy expenditure, but, as
weight loss by other means is associated with a decrease in basal metabolic rate, the
fact that energy expenditure did not fall in our leptin deficient subjects is notable.
The administration of leptin permitted progression of appropriately timed pubertal
development in the single child of appropriate age and did not cause the early onset of
puberty in the younger children [6]. In adults with leptin deficiency, leptin induced the
development of secondary sexual characteristics and pulsatile gonadotrophin secretion.
4 Farooqi
In the 3 previously reported children there were small, but sustained, increases in
free T4, free T3, and TSH that occurred within 1 month of leptin therapy. These obser-
vations are fully consistent with an effect of leptin at the hypothalamic level. A 4th
patient had substantial elevation of TSH before treatment, such that thyroxine ther-
apy was commenced [7]. However, replacement therapy was stopped when thyroid
function tests normalised after leptin treatment.
In 1998, Krude et al. [16] provided the first description of humans congenitally lack-
ing POMC gene products. One proband was a compound heterozygote for two non-
sense mutations and a 2nd patient was homozygous for a mutation in the
5⬘-untranslated region that introduced an additional out-of-frame start site, thus
interfering with POMC translational initiation. Subsequently, Krude et al. [16] have
reported 3 additional unrelated European children with congenital POMC deficiency
who were either homozygous or compound heterozygous for POMC mutations. We
have recently identified a 6th patient with complete POMC deficiency, being
homozygous for a complete loss of function mutation which results in the loss of all
POMC-derived peptides [17].
These patients all presented in early life with features of hypocortisolaemia sec-
ondary to ACTH deficiency, leading to hypoglycaemia, prolonged jaundice, suscepti-
bility to the effects of infection and in one case, neonatal death. The children
responded well to physiological replacement with glucocorticoids but all subse-
quently developed marked obesity in association with hyperphagia.
Notably, all children thus far reported have pale skin and red hair, features consis-
tent with the known role of POMC-derived peptides in the determination of the
phaeomelanin to eumelanin ratio in melanocytes. Our Turkish proband is the first
reported patient with POMC deficiency who does not have red hair [17]. It is likely
this can be explained by his differing genetic background as the other reported
patients were all white Caucasian subjects of European ancestry. The retention of
dark hair in this child and his similarly affected deceased sibling indicates that the
synthesis of eumelanin in humans is not absolutely dependent on the presence of
melanocortin peptides.
POMC Haploinsufficiency
Krude et al. [18] have previously attempted to assess the impact of loss of one POMC
allele in the parents and heterozygous relatives of their probands. They estimated the
maximum lifetime BMI SDS in adult POMC heterozygotes and suggested that most
had a maximum lifetime BMI SDS of 1, which is at the upper end of the normal range.
6 Farooqi
events, firstly within the endoplasmic reticulum and then within the secretory vesi-
cles of the regulated secretory pathway to generate a fully active 66-kDa isoform that
is stored in mature secretory granules.
We have previously reported an adult female with severe early-onset obesity,
hypogonadotropic hypogonadism, postprandial hypoglycaemia, hypocortisolaemia,
and evidence of impaired processing of POMC and proinsulin [23]. She was found to
be a compound heterozygote for PC1 mutations [24]. We have described the second
case of congenital PC1 deficiency, in a patient who was a compound heterozygote for
two loss of function mutations [25]. Intriguingly, this patient suffered from severe
small intestinal absorptive dysfunction as well as the characteristic severe early-onset
obesity, impaired prohormone processing, and hypocortisolaemia. We hypothesised
that the small intestinal dysfunction seen in this patient and, to a lesser extent, in the
1st patient we described may be the result of a failure of maturation of propeptides
within the enteroendocrine cells and nerves that express PC1 throughout the gut. The
finding of elevated levels of progastrin and proglucagon provided in vivo evidence
that prohormone processing in enteroendocrine cells was abnormal [25].
In 1998, two groups reported heterozygous mutations in the MC4 receptor in humans
which were associated with dominantly inherited obesity [26, 27]. Since then, het-
erozygous mutations in MC4R have been reported in obese humans from various
ethnic groups [28–30].
We have studied over 2,000 severely obese probands and found that approximately
5–6% have pathogenic MC4R mutations that are non-conservative in nature, not
found in control subjects from the background population and co-segregate with
obesity in families [31]. The prevalence of MC4R mutations has varied from 0.5% of
obese adults to 6% in patients with severe childhood obesity [31, 32]. Recent studies
provide an important indication of the true population prevalence of this disorder in
UK [33] and European populations [32].
While we found a 100% penetrance of early-onset obesity in heterozygous
probands, others have described obligate carriers who were not obese [29]. Given the
large number of potential influences on body weight, it is perhaps not surprising that
both genetic and environmental modifiers will have important effects in some pedi-
grees. Indeed we have now studied 6 families in whom the probands were homozy-
gotes and in all of these, the homozygotes were more obese than heterozygotes [31].
Interestingly, in these families, some heterozygous carriers were not obese. Taking
account of all of these observations, co-dominance, with modulation of expressivity
and penetrance of the phenotype, is the most appropriate descriptor for the mode of
inheritance. This finding is supported by the pattern of inheritance of obesity seen in
heterozygous and homozygous MC4R knockout mice [34].
50
40
30
20
10
0
Leptin Inactive Partial Treated Controls
deficiency MC4R mutations leptin
deficiency
Fig. 2. Genotype-phenotype correlations in human MC4R deficiency. Ad libitum food intake at an 18MJ
test meal for patients with leptin deficiency and complete and partial loss of function MC4R mutations.
We have now studied over 150 MC4R deficient subjects in our Clinical Research
Facility. The clinical features of MC4R deficiency include hyperphagia, which invari-
ably starts in the 1st year of life [31]. Alongside the increase in fat mass, MC4R-defi-
cient subjects also have an increase in lean mass and a marked increase in bone
mineral density, thus they often appear ‘big-boned’. They exhibit accelerated linear
growth in early childhood, which does not appear to be due to dysfunction of the GH
axis and may be a consequence of the disproportionate early hyperinsulinaemia seen
in these patients [31]. The accelerated linear growth and the disproportionate early
hyperinsulinaemia are consistent with observations in the MC4R KO mouse [35].
Although affected subjects are objectively hyperphagic, ad libitum energy intake at
a test meal is not as large as that seen with leptin deficiency [31]. Of particular note is
the finding that the severity of receptor dysfunction seen in in vitro assays can predict
the amount of food ingested at a test meal by the subject harbouring that particular
mutation (fig. 2).
We have studied in detail the signalling properties of many of these mutant receptors
and this information should help to advance the understanding of structure/function
relationships within the receptor [36]. Importantly, we have been unable to demonstrate
evidence for dominant negativity associated with these mutants, which suggests that
MC4R mutations are more likely to result in a phenotype through haplo-insufficiency
[36]. About 70% of missense mutations in MC4R are retained intracellularly [37].
8 Farooqi
While, at present, there is no specific therapy for MC4R deficiency, it is highly likely
that these subjects would respond well to pharmacotherapy that overcame the reduc-
tion in the hypothalamic melanocortinergic tone that exists in these patients. As most
patients are heterozygotes with one functional allele intact, it is possible that small mol-
ecule MC4R agonists might, in future, be excellent treatments for this disorder [38].
Recently, the concept that hypothalamic neuronal networks involved in energy home-
ostasis are ‘hardwired’ has been challenged. In mice, hypothalamic neurones projecting
from the arcuate nucleus to the paraventricular nucleus develop after birth and their
development is regulated by leptin [39]. In addition, synaptic plasticity in the mature
rodent brain has been identified as a component of the neuronal regulation of energy
homeostasis as leptin has been shown to acutely modulate excitatory and inhibitory
synaptic inputs at the level of first-order arcuate neurones [40]. However, it is difficult to
establish whether synaptic plasticity plays a role in the physiological regulation of energy
homeostasis in humans and whether under pathological conditions, hypothalamic neu-
ronal networks and plasticity may be impaired and contribute to human obesity.
Brain-derived neurotrophic factor (BDNF) regulates the development, survival and
differentiation of neurons through its high-affinity receptor, tropomyosin-related kinase
B (TrkB). Recently, BDNF has been implicated in the regulation of body weight, as its
expression is reduced by fasting [41] and BDNF administration causes weight loss in
wild-type mice through a reduction in food intake. BDNF has also been implicated in
memory and a range of behaviours using a number of conditional knockout models [42].
We previously reported a child with severe obesity, impaired short-term memory
and developmental delay who had a de novo missense mutation impairing the function
of TrkB, the tyrosine kinase receptor that mediates the effects of both BDNF and the
neurotrophin, NT4/5 [43]. We have also identified a patient with severe hyperphagia
and obesity and a complex neurobehavioural phenotype including impaired cognitive
function and memory as well as distinctive hyperactive behaviour. Interestingly, this
patient has a de novo paracentric inversion, 46,XX,inv(11)(p13p15.3), which encom-
passes the BDNF locus and disrupts BDNF expression [44]. Although to date only 2
such patients have been identified, understanding the mechanisms whereby BDNF reg-
ulates hypothalamic neuronal circuits may have potential therapeutic benefits for the
treatment of more common forms of human obesity.
Conclusions
In practical terms the discovery of these genetic disorders has helped de-stigmatise
human obesity and allow it to be seen as a medical condition. The genetic defects
References
1 Maes HH, Neale MC, Eaves LJ: Genetic and environ- 13 Ozata M, Ozdemir IC, Licinio J: Human leptin defi-
mental factors in relative body weight and human ciency caused by a missense mutation: multiple
adiposity. Behav Genet 1997;27:325–351. endocrine defects, decreased sympathetic tone, and
2 Stunkard AJ, Harris JR, Pedersen NL, McClearn GE: immune system dysfunction indicate new targets for
The body-mass index of twins who have been reared leptin action, greater central than peripheral resis-
apart. N Engl J Med 1990;322:1483–1487. tance to the effects of leptin, and spontaneous correc-
3 Barsh GS, Farooqi IS, O’Rahilly S: Genetics of body- tion of leptin-mediated defects. J Clin Endocrinol
weight regulation. Nature 2000;404:644–651. Metab 1999;84:3686–3695.
4 Farooqi IS, O’Rahilly S: Monogenic obesity in humans. 14 Mantzoros CS, Ozata M, Negrao AB, et al: Synch-
Annu Rev Med 2005;56:44x3–458. ronicity of frequently sampled thyrotropin (TSH)
5 Montague CT, Farooqi IS, Whitehead JP, et al: Con- and leptin concentrations in healthy adults and lep-
genital leptin deficiency is associated with severe early- tin-deficient subjects: evidence for possible partial
onset obesity in humans. Nature 1997;387: 903–908. TSH regulation by leptin in humans. J Clin Endocrinol
6 Farooqi IS, Matarese G, Lord GM, et al: Beneficial Metab 2001;86:3284–3291.
effects of leptin on obesity, T cell hyporesponsive- 15 Licinio J, Caglayan S, Ozata M, et al: Phenotypic
ness, and neuroendocrine/metabolic dysfunction of effects of leptin replacement on morbid obesity, dia-
human congenital leptin deficiency. J Clin Invest betes mellitus, hypogonadism, and behavior in lep-
2002;110:1093–1103. tin-deficient adults. Proc Natl Acad Sci USA 2004;101:
7 Gibson WT, Farooqi IS, Moreau M, et al: Congenital 4531–4536.
leptin deficiency due to homozygosity for the 16 Krude H, Biebermann H, Luck W, Horn R, Brabant
Delta133G mutation: report of another case and G, Gruters A: Severe early-onset obesity, adrenal
evaluation of response to four years of leptin therapy. insufficiency and red hair pigmentation caused by
J Clin Endocrinol Metab 2004;89:4821–4826. POMC mutations in humans. Nat Genet 1998;19:
8 Strobel A, Issad T, Camoin L, Ozata M, Strosberg AD: 155–157.
A leptin missense mutation associated with hypogo- 17 Farooqi IS, Drop S, Clements A, et al: Heterozygosity
nadism and morbid obesity. Nat Genet 1998;18: for a POMC-null mutation and increased obesity risk
213–215. in humans. Diabetes 2006;55:2549–2553.
9 Clement K, Vaisse C, Lahlou N, et al: A mutation in 18 Krude H, Biebermann H, Schnabel D, et al: Obesity
the human leptin receptor gene causes obesity and due to proopiomelanocortin deficiency: three new
pituitary dysfunction. Nature 1998;392:398–401. cases and treatment trials with thyroid hormone and
10 Lahlou N, Landais P, De Boissieu D, Bougneres PF: ACTH4–10. J Clin Endocrinol Metab 2003;88:
Circulating leptin in normal children and during the 4633–4640.
dynamic phase of juvenile obesity: relation to body 19 Challis BG, Pritchard LE, Creemers JW, et al: A mis-
fatness, energy metabolism, caloric intake, and sex- sense mutation disrupting a dibasic prohormone
ual dimorphism. Diabetes 1997;46:989–993. processing site in pro-opiomelanocortin (POMC)
11 Farooqi IS, Wangensteen T, Collins S, et al: Clinical increases susceptibility to early-onset obesity through
and molecular genetic spectrum of congenital defi- a novel molecular mechanism. Hum Mol Genet 2002;
ciency of the leptin receptor. N Engl J Med 2007;356: 11:1997–2004.
237–247. 20 Lee YS, Challis BG, Thompson DA, et al: A POMC
12 Farooqi IS, Jebb SA, Langmack G, et al: Effects of variant implicates beta-melanocyte-stimulating hor-
recombinant leptin therapy in a child with congenital mone in the control of human energy balance. Cell
leptin deficiency. N Engl J Med 1999;341: 879–884. Metab 2006;3:135–140.
10 Farooqi
21 Biebermann H, Castaneda TR, van Landeghem F, 33 Alharbi KK, Spanakis E, Tan K, et al: Prevalence and
et al: A role for beta-melanocyte-stimulating hor- functionality of paucimorphic and private MC4R
mone in human body-weight regulation. Cell Metab mutations in a large, unselected European British
2006;3:141–146. population, scanned by meltMADGE. Hum Mutat
22 Seidah NG, Chretien M: Proprotein and prohormone 2007;28:294–302.
convertases: a family of subtilases generating diverse 34 Huszar D, Lynch CA, Fairchild-Huntress V, et al:
bioactive polypeptides. Brain Res 1999;848:45–62. Targeted disruption of the melanocortin-4 receptor
23 O’Rahilly S, Gray H, Humphreys PJ, et al: Brief report: results in obesity in mice. Cell 1997;88:131–141.
impaired processing of prohormones associated with 35 Fan W, Dinulescu DM, Butler AA, Zhou J, Marks DL,
abnormalities of glucose homeostasis and adrenal Cone RD: The central melanocortin system can
function. N Engl J Med 1995;333:1386–1390. directly regulate serum insulin levels. Endocrinology
24 Jackson RS, Creemers JW, Ohagi S, et al: Obesity and 2000;141:3072–3079.
impaired prohormone processing associated with 36 Yeo GS, Lank EJ, Farooqi IS, Keogh J, Challis BG,
mutations in the human prohormone convertase 1 O’Rahilly S: Mutations in the human melanocortin-4
gene [see comments]. Nat Genet 1997;16:303–306. receptor gene associated with severe familial obesity
25 Jackson RS, Creemers JW, Farooqi IS, et al: Small- disrupts receptor function through multiple molecu-
intestinal dysfunction accompanies the complex lar mechanisms. Hum Mol Genet 2003;12: 561–574.
endocrinopathy of human proprotein convertase 1 37 Lubrano-Berthelier C, Durand E, Dubern B, et al:
deficiency. J Clin Invest 2003;112:1550–1560. Intracellular retention is a common characteristic of
26 Yeo GS, Farooqi IS, Aminian S, Halsall DJ, Stanhope childhood obesity-associated MC4R mutations.
RG, O’Rahilly S: A frameshift mutation in MC4R Hum Mol Genet 2003;12:145–153.
associated with dominantly inherited human obesity 38 Nargund RP, Strack AM, Fong TM: Melanocortin-4
[letter]. Nat Genet 1998;20:111–112. receptor (MC4R) agonists for the treatment of obe-
27 Vaisse C, Clement K, Guy-Grand B, Froguel P: A frame- sity. J Med Chem 2006;49:4035–4043.
shift mutation in human MC4R is associated with a 39 Bouret SG, Draper SJ, Simerly RB: Trophic action of
dominant form of obesity. Nat Genet 1998;20: 113–114. leptin on hypothalamic neurons that regulate feed-
28 Farooqi IS, Yeo GS, Keogh JM, et al: Dominant and ing. Science 2004;304:108–110.
recessive inheritance of morbid obesity associated 40 Pinto S, Roseberry AG, Liu H, et al: Rapid rewiring of
with melanocortin 4 receptor deficiency. J Clin Invest arcuate nucleus feeding circuits by leptin. Science
2000;106:271–279. 2004;304:110–115.
29 Vaisse C, Clement K, Durand E, Hercberg S, Guy- 41 Xu B, Goulding EH, Zang K, et al: Brain-derived
Grand B, Froguel P: Melanocortin-4 receptor muta- neurotrophic factor regulates energy balance down-
tions are a frequent and heterogeneous cause of stream of melanocortin-4 receptor. Nat Neurosci
morbid obesity. J Clin Invest 2000;106:253–262. 2003;6:736–742.
30 Hinney A, Schmidt A, Nottebom K, et al: Several 42 Snider WD: Functions of the neurotrophins during
mutations in the melanocortin-4 receptor gene nervous system development: what the knockouts
including a nonsense and a frameshift mutation asso- are teaching us. Cell 1994;77:627–638.
ciated with dominantly inherited obesity in humans. J 43 Yeo GS, Connie Hung CC, Rochford J, et al: A de
Clin Endocrinol Metab 1999;84:1483–1486. novo mutation affecting human TrkB associated with
31 Farooqi IS, Keogh JM, Yeo GS, Lank EJ, Cheetham T, severe obesity and developmental delay. Nat
O’Rahilly S: Clinical spectrum of obesity and muta- Neurosci 2004;7:1187–1189.
tions in the melanocortin 4 receptor gene. N Engl J 44 Gray J, Yeo GS, Cox JJ, et al: Hyperphagia, severe
Med 2003;348:1085–1095. obesity, impaired cognitive function, and hyper-
32 Larsen LH, Echwald SM, Sorensen TI, Andersen T, activity associated with functional loss of one copy of
Wulff BS, Pedersen O: Prevalence of mutations and the brain-derived neurotrophic factor (BDNF) gene.
functional analyses of melanocortin 4 receptor vari- Diabetes 2006;55:3366–3371.
ants identified among 750 men with juvenile-onset
obesity. J Clin Endocrinol Metab 2005;90:219–224.
Abstract
Obesity results from the complex interaction of environmental factors that act on a genetic back-
ground that determines the susceptibility to obesity. The identification of such obesity susceptibility
genes can provide important insights into the mechanism underlying this condition. While candidate
gene approaches have not been tremendously successful in identifying relevant genetic contributors
to obesity, except PPAR γ , the advent of genome-wide strategies has recently revealed novel and unex-
pected genetic factors with strong associations with obesity and/or diabetes, i.e. FTO, TCF7L2, INSIG2,
ENPP1, or FASN (reviewed herein), although some of them are not undebated. Considering the function
of the encoded proteins, it will now be of interest to investigate the cellular and molecular mechanisms,
how these genetic variations affect body weight, energy metabolism and/or obesity-associated
morbidity. Copyright © 2008 S. Karger AG, Basel
Obesity results from an imbalance of energy expenditure and energy intake. There is
a great variety of factors affecting this fragile balance. It is obvious that major envi-
ronmental factors such as the accessibility to food and the degree of physical activity,
but also the psychosocial environment, the perinatal environment and many other
factors affect an individual’s body weight. The pandemic increase in obesity preva-
lence had been attributed to this increasingly urbanized and sedentary lifestyle with
convenient access to food, increased calorie intake and a reduction of energy expen-
diture in the industrialized world.
Nevertheless, over the last two decades, it has become clear that genetic factors
play an important role in the determination of body weight. First evidence for the
heritability of obesity came from early twin studies that observed a heritability for
body weight of 0.78–0.81% in monozygous twins [1–3] and similar values have
been obtained in subsequent studies analyzing the impact of the genetic back-
ground [4–7]. However, this high degree of heritability is rarely attributed to
monogenic forms of obesity [see the chapter by Farooqi, this vol., pp. 1–11], which
usually result in extreme and early-onset obesity and are usually accompanied by
additional phenotypic and endocrine abnormalities. Even though the discovery of
monogenic forms of obesity has allowed important insights into some of these
mechanisms by revealing a highly conserved pathway regulating mammalian body
weight, it is obvious that the pathology of obesity is far more complex.
Considering that the genetic pool has not dramatically changed over the last
decades, the secular trend of increased obesity prevalence is now regarded as the
interaction of the modern lifestyle factors that act on a genetic background that deter-
mines an individuals susceptibility to weight gain and obesity. According to this
‘thrifty gene’ hypothesis [8] individuals with a genetic disposition to accumulate
ample energy stores in times of good food availability were evolutionary more likely
to survive times of nutrient scarcity and to pass these genotypes to successive genera-
tions. For example, if feast and famine cycles characterized early human life, the
‘thrifty genotype’ was more likely to survive periods of food scarcity. This ancient
genetic selection to deposit fat efficiently is maladaptive in our modern obesogenic
environment with excess calorie intake and sedentary life style, and hence the same
genes now contribute to obesity.
It is now well acknowledged that a multitude of genetic polymorphisms and candi-
date regions scattered all over the genome regulate an individual’s susceptibility to
weight gain. Evolutionary concepts together with extensive population genetics to
characterize geographical and haplotypic structures of newly emerging biological as
well positional candidate genes will be inevitable to reveal whether these new genes
could indeed represent the ‘thrifty’ genes. These genetic studies provide valuable
insights as well as promoting new concepts into the mechanisms from the identifica-
tion of previously unsuspected genetic factors.
Complex diseases such as diabetes or obesity have genetic components, which due
to their polygenic nature can not easily be identified. Two basic approaches have
been used to identify susceptibility genes for complex diseases: candidate gene
approach and genomic approach (fig. 1). However, only limited success has been seen
so far.
Polygenic Obesity 13
Human Experimental model
Synteny
Chromosomal region
Disease gene
Fig. 1. Identifying genetic factors that underlie complex diseases. QTL ⫽ Quantitative trait locus.
Selection of candidate genes for obesity is usually based on their known physiological
role in pathways related to energy expenditure, food intake but also glucose and/or lipid
metabolism and hence requires some a priori knowledge of the pathophysiology of a
disease. In addition, candidate genes are selected on the basis of previous evidence on
association with obesity and/or diabetes in other populations or experimental animal
models. These genes are then analyzed for sequence variation that is associated or
linked with the disease. Even though advances in genotyping technology allied with our
knowledge of the human genome’s structure will lead to novel common gene variants
involved in susceptibility to human obesity, the candidate gene approach may still be
very powerful when it comes to identifying rare variants predisposing to obesity. This is
given by the fact that whereas coverage of common variation in genes by commercially
available single nucleotide polymorphism (SNP) panels (provided by Affymetrix and
Illumina) is generally comparable to the rest of the genome (see association studies
below), other focused classes of functional variants are captured poorly by SNP sets
aimed at common variation [9]. A large number of genes have been associated with the
development of obesity; they have been reviewed recently elsewhere [10, 11].
Genome-Wide Strategies
p = 0.00000001
Disease allele: strong association with obesity
Genetic marker
(SNP) A C TGA A G C
being replaced by
genome-wide
association studies.
with a phenotype [12]. In the last two decades, genetic studies have focused on the tech-
nique of genetic linkage. This study design proved to be efficient for identifying rare,
high-risk alleles, i.e. alleles that have a large population attributable fraction (PAF) in rare
single-gene mendelian diseases (cystic fibrosis) [13, 14] but do not appear to have large
PAF in common diseases. Using linkage studies, researchers attempt to find regions of
the genome with a higher than expected number of shared alleles among affected indi-
viduals within a family. Since the design is based on closely related individuals within a
family, and these individuals share larger regions of the genome, genotyping of a rela-
tively small number of polymorphic markers is sufficient to detect region of linkage.
Using positional cloning one may attempt to identify the gene(s), which reside within the
linkage region and segregate with the phenotype. Usually, genetic markers (e.g. SNPs) are
being selected to provide a high-density map (e.g. every 2 kb) within the region of link-
age. The SNPs are then being genotyped in study subjects and analyzed for association
with the disease. SNPs with strongest associations may be the causal disease risk variants
(direct association) or are in a close proximity and thus in high linkage disequilibrium
(non-random correlation between alleles at a pair of SNPs) and possibly indirectly
associated with the disease variants [12].
Polygenic Obesity 15
In contrast to linkage approaches, association analyses are expected to be more
powerful in identifying alleles that confer modest risk for developing a complex
disease. Common modest risk variants account for a larger PAF than do rare high-
risk alleles and this is often referred to as disease-common variant hypothesis [15].
The advantage of association analyses is based on the fact that for modest-risk alleles
the patterns of sharing between related affected individuals are less striking than pat-
terns of sharing between unrelated affected individuals [16]. This may partially
explain the limited success of linkage studies in identifying genes for common dis-
eases such as obesity. Another advantage of association analyses is that it is much eas-
ier to recruit large cohorts of unrelated individuals than collecting large pedigrees.
However, since the shared region among unrelated individuals is much smaller than
among the family members, association analyses require higher marker densities
than linkage analyses [16]. This seems to be given by recent advances in the field of
high throughput genotyping of SNPs, enabling a high-density coverage of the genome.
Besides the marker coverage, another point that requires consideration in association
analyses based on comparisons of the variants’ frequencies between cases and con-
trols is the power. Most studies are underpowered, especially when considering that
in order to detect associations with an odds ratio of 1.2 one would need 1,000 cases
and 1,000 controls [17]. Therefore, large collaborating networks enabling replications
of initial findings are crucial for successful association analyses. This was impressively
represented by a very recent genome-wide analysis involving research centers from
Europe and including more than 35,000 study subjects, which led to discovering FTO
as the strongest predictor of human polygenic obesity seen so far [18, 19].
Isolated Populations
Along with the polygenic nature and pathophysiologic complexity of diseases such as
obesity or diabetes, another major problem is the genetic heterogeneity of modern
populations. This means that healthy and afflicted subjects are likely to have very dif-
ferent sequences throughout the genome, not only in the area with the disease
gene(s). To reduce genetic heterogeneity, one can use crossing studies with experi-
mental models, which are inbred and so genetically uniform. Furthermore, their
environment can be standardized to overcome gene-environment interaction.
Linkage or quantitative trait analyses using experimental crosses may lead to chro-
mosomal regions associated with a phenotype of interest, which may then point to
syntenic regions in humans and so suggest novel human candidate genes. However,
very often it proves to be extremely difficult to find an experimental model com-
pletely resembling pathophysiological aspects/patterns of human diseases. Therefore,
one may attempt to reduce genetic heterogeneity by studying populations with
limited genetic variability [20]. Isolated populations have already contributed to iden-
tification of mendelian variants of complex diseases, such as Hirschsprung disease in
Studies in Children
Polygenic Obesity 17
strongest evidence for linkage with body mass index (BMI) was on chromosome
11q23–24 (LOD ⫽ 3.6) [32], while the strongest evidence for linkage with percentage
of body fat was on chromosome 17q25 (LOD ⫽ 1.9) in a multipoint sibling-based
variance component analysis [33]. The region on chromosome 17q35 seemed to be
particularly interesting since the human fatty acid synthase (FAS) gene was positioned
at 135 cM, within 7 cM of the peak of linkage to percentage of body fat [33]. The FAS
enzyme is necessary for the de novo synthesis of long-chain fatty acids from acetyl-
CoA, malonyl-CoA and NADPH [34]. Recent physiologic studies have shown that
inhibition of the FAS gene induces a rapid decline in fat stores in mice, suggesting a
role for FAS in energy homeostasis [35, 36]. Based on the chromosomal location and
known physiology of FAS, the FAS was investigated as a candidate gene for determin-
ing body weight and percentage of body fat in Pima Indians. In these studies, a novel
Val1483Ile polymorphism was identified, which was associated with percentage of
body fat and 24-hour substrate oxidation rates measured in a respiratory chamber
[37]. These findings indicate that the Val1483Ile substitution in FAS is protective
against obesity in Pima Indians, an effect possibly explained by the role of this gene in
the regulation of substrate oxidation. The effects of Val1483 on obesity have been
recently also investigated in Caucasian children and adolescents. In a cohort of 738
Caucasian children and adolescents and 205 obese children from Leipzig, Germany, a
significant interaction effect between gender and genotype was observed [38]. The
findings in Caucasian children suggest a gender specific protective effect of the
Val1483Ile polymorphism in FAS for obesity and lipid phenotypes in Caucasian boys.
The story of FAS is an example of how combining different approaches may ultimately
lead to novel genetic targets possibly involved in the pathophysiology of human obe-
sity. FAS is not only an excellent candidate gene based on its known biological func-
tion, but also a positional candidate mapped within a region of linkage to percentage of
body fat. In addition, studies in Pima Indians prove that genetically isolated popula-
tions may definitely help in the search for new obesity risk candidate genes.
Candidate Genes
Polygenic Obesity 19
PPAR␥-mediated adipocytokine release is involved. Adiponectin levels are higher in
subjects with the Ala/Ala genotype compared to the Pro/Ala and the Pro/Pro genotype
[54]. However, studies are not entirely coherent and an interaction of genetic and envi-
ronmental factors in the regulation of serum adiponectin concentrations is assumed
[40]. Adiponectin renders the liver more sensitive to the suppressive effect of insulin on
glucose production [40]. Since adiponectin release is under transcriptional control of
PPAR␥ [55] it represents a plausible candidate for mediating the effect of the Pro12Ala
polymorphism on hepatic insulin sensitivity and clearance. In support of this scenario,
in Pima Indians insulin suppression of glucose production was 40% more efficient in
carriers of the Ala allele while insulin-stimulated glucose uptake was not different [56].
In macrophages, PPAR␥ regulates production of inflammatory mediators, which mod-
ifies vascular inflammation and endothelial dysfunction [50] providing an alternative
scenario of how the polymorphism exerts its effects.
With the improved accessibility and availability of genome-wide scan platforms and
gene chips, this technique has increasingly been applied to identify new candidate
genes for obesity and diabetes in a relatively short time period over the last 2 years.
Polygenic Obesity 21
In contrast to the studies confirming the association of ENPP1 variants with T2D
and/or obesity [69–74], metabolic syndrome [77] or T2D and coronary artery disease
[78], there are a number of studies that failed to find these associations [79–83], or
even reported contrary results of the 121Q variant being associated with decreased
BMI or risk for T2D [84, 85] in different populations. However, in meta-analyses that
also included these negative studies, the effect of the 121Q risk allele on T2D was
retained [80, 83]. These divergences may be attributed to ethnic differences as has
been shown in US minority groups [86], although the comparison of ethnic popula-
tions did not reveal such differences [69, 84].
Overall, there is evidence from genome-wide studies, from individual association
studies, from an animal model and from in vitro studies for a role of ENPP1 as a risk
factor for T2D and potentially obesity, but the lack of reproducibility in a consider-
able number of studies raises some caution regarding this hypothesis.
Polygenic Obesity 23
From Obesity to Diabetes: TCF7L2 – The Strongest Predictor of T2D
The TCF7L2 gene encodes for a transcription factor (Tcf-4) that is involved in the
regulation of cellular proliferation and differentiation [98].
Variants in the TCF7L2 gene have recently been associated with an increased risk for
T2D in a genome-wide analysis of the isolate population of Iceland [26]. The relative
risk for diabetes of 1.45 for heterozygous to 2.41 for homozygous carriers of the gene
variants was appreciably greater than for most identified genetic factors so far,
accounting for a population attributable risk of 21% [26]. The strongest associations
with T2D were reported for the rs7903146 variant with a clear gene dose effect [99].
Subsequent independent studies in several distinct ethnic populations, all in adults,
convincingly confirmed the initial findings, which was finally evidenced by a large
global meta-analysis [100]. This knowledge is extended by showing that these risk
alleles actually predict the progression from impaired glucose tolerance to diabetes
prospectively [101] and an increased severity of the disease [102] in adults.
Considering these data that suggest TCF7L2 as a major candidate gene for the pre-
disposition to T2D, one may also hypothesize that carriers of those at-risk variants
have an earlier age of onset than noncarriers. In our obese childhood cohort, three
risk variants (rs7901695, rs7903146, rs1225572) were significantly associated with
higher fasting and 120-min blood glucose levels. Although not statistically signifi-
cant, fasting and peak insulin levels and HOMA-IR appeared with a similar tendency
[103]. These data indicate that TCF7L2 variants confer a higher risk for early impair-
ment of glucose metabolism emerging as soon as in childhood and adolescence,
although we cannot directly predict the progression to overt diabetes from these data.
Studies in nondiabetic subjects may, however, help to understand the mechanisms
involved in this progression. While most studies in adults did not identify an associa-
tion of TCF7L2 variants with measures of insulin resistance [101, 104–106], several
studies did observe a defect in insulin secretion [101, 105, 107–109], possibly by
impaired -cell proinsulin processing [110], or a trend for decreased fasting insuline-
mia [109] in subjects carrying the risk alleles. The insulin secretion in type 1 diabetes
or MODY and neonatal diabetes forms are, however, not affected by TCF7L2 poly-
morphisms [111, 112]. The conclusion from these data was that the polymorphisms
affect the capacity of pancreatic -cells to secrete insulin rather than aggravating
insulin resistance. This was further supported by expression data suggesting a puta-
tive role of TCF7L2 in -cell differentiation [109]. A most recent implication exem-
plifies the potential applicability of genetic variants of TCF7L2 in the prediction of
drug response in the sense of pharmacogenomics. Pearson et al. [113] observed that
carriers of the rs12255372 risk allele were less likely to respond to sulfonylureas.
Considering that TCF7L2 as a transcription factor regulates genes involved in pro-
liferation and differentiation, one may hypothesize that early-onset obesity through
increased or dysregulated adipogenesis may constitute one underlying mechanism
[114, 115]. It has been shown that the TCF7L2 interacts with the Wnt/-catenin path-
way and may hence constitute a mediator of inhibition of adipogenesis by TNF␣
Polygenic Obesity 25
Notably, three major and most recent studies were intrigued by a high congruency
in the genes identified with an association with T2D.
CDKAL1 is a gene of yet unknown function that shares homology with a CDK5
regulatory subunit-associated protein, which is supposed to inhibit the activation of
cyclin-dependent kinase 5 (CDK5). An association of CDKAL1 with T2D was previ-
ously unknown but similarly identified by all three of these studies [87, 125, 126].
Another chromosomal region associated with T2D was identified near a cluster
involving CDKN2A (cyclin-dependent kinase inhibitor-2A) [87, 125, 126]. This gene,
also known as p16INK4a, was initially implicated in tumor susceptibility, particularly
for melanoma [127]. However, recently mice lacking p16INK4a were reported to
show enhanced islet proliferation and prolonged survival after -cell ablation [128].
The third new candidate with an association with T2D is IGF2BP2 [87, 125, 126].
IGF2BP2, also known as IMP2, is an IGF-II gene mRNA-binding protein that attaches
to the 5⬘ UTR from the translationally regulated IGF-II leader 3 mRNA and causes a
dose-dependent translational repression of IGF-II leader 3-luciferase mRNA [129].
IMPs are expressed in developing epithelia, muscle, and placenta in both mouse and
human embryos [129]. Hence, this gene is interesting with respect to the fundamen-
tal role of IGF-II in embryonic development. Another large-scale genome-wide asso-
ciation study also confirmed an association of TCF7L2, KCNJ11, and PPARG with
T2D [88].
Of particular note is that many of these genes, i.e. HHEX, SLC30A3, CDKN2A,
potentially TCF7L2, and others recently identified [124] can be implicated in the
physiology of pancreatic -cells through effects on embryonic development, regener-
ation or insulin processing.
Certainly, replication in more cohorts and the elucidation of the underlying mech-
anism remain mandatory. Nevertheless, the discovery of previously unsuspected
genes opens up new directions and potentially important insights into the pathogen-
esis of common diseases.
Perspectives: Interactions
Even though some recent approaches have been successful in identifying important
genetic contributors to obesity, these new genes are obviously not much more than just
a few pieces of the puzzle in the complex framework of pathogenic factors contributing
to obesity and it is obvious that many more genes await discovery. The successful iden-
tification of genetically and clinically significant and relevant genetic factors will
require study setups comprising of a large number of individuals with a defined and
well-characterized phenotype, the appropriate controls, the technological platforms
for high throughput analyses, and finally the statistical tools. For this, increasingly,
consortia that comprise a number of cohorts of distinct origin are being established to
identify genetic factors in genome-wide association studies. In order to separate true
Polygenic Obesity 27
Given the polygenic etiology of obesity, it has to be considered that the contribution
of the single genes is small. This challenge necessitates sensitive and powerful tools to
detect the significant but sometimes subtle effects of the genes. For future analyses, an
approach that combines the advantages of genome-wide association studies with the
rationale of candidate gene approaches, and further refined by adding information
from microarray and proteomic techniques may be promising for identifying novel
pathways and molecules as well as selecting the significant and relevant genes [139].
Combining the information from several risk factors can allow the identification
of high-risk groups of individuals and may hence have an important role in the initi-
ation of preventative measures for complex (polygenic) diseases such as obesity.
Similarly to what has recently been stated for diabetes, the genetics of obesity is
still a puzzle but no longer a nightmare [17]. Today we know that even the strongest
genetic predictors of obesity are not likely to exceed an odds ratio of 2.0. But we also
know that these predictors combined between themselves and/or with environment
are likely to explain most genetic factors controlling obesity. Identifying these mecha-
nisms remains to be a very exciting challenge.
Acknowledgments
This work was supported by grants from the Deutsche Forschungsgemeinschaft KFO 152:
‘Atherobesity’, projects KO 3512/1-1 (TP 1) and BE 1264/10-1 (TP5), the European Community
integrated project grant ‘PIONEER’, and by the Interdisciplinary Centre for Clinical Research
Leipzig at the Faculty of Medicine of the University of Leipzig (B27 and N06).
References
1 Feinleib M, Garrison RJ, Fabsitz R, Christian JC, 6 Katzmarzyk PT, Perusse L, Rao DC, Bouchard C:
Hrubec Z, Borhani NO, Kannel WB, Rosenman R, Familial risk of overweight and obesity in the
Schwartz JT, Wagner JO: The NHLBI twin study of Canadian population using the WHO/NIH criteria.
cardiovascular disease risk factors: methodology and Obes Res 2000;8:194–197.
summary of results. Am J Epidemiol 1977;106: 284–285. 7 Allison DB, Kaprio J, Korkeila M, Koskenvuo M,
2 Stunkard AJ, Foch TT, Hrubec Z: A twin study of Neale MC, Hayakawa K: The heritability of body
human obesity. JAMA 1986;256:51–54. mass index among an international sample of
3 Stunkard AJ, Harris JR, Pedersen NL, McClearn GE: monozygotic twins reared apart. Int J Obes Relat
The body-mass index of twins who have been reared Metab Disord 1996;20:501–506.
apart. N Engl J Med 1990;322:1483–1487. 8 Neel JV, Weder AB, Julius S: Type II diabetes, essen-
4 Pietilainen KH, Kaprio J, Rissanen A, Winter T, tial hypertension, and obesity as ‘syndromes of
Rimpela A, Viken RJ, Rose RJ: Distribution and heri- impaired genetic homeostasis’: the ‘thrifty genotype’
tability of BMI in Finnish adolescents aged 16y and hypothesis enters the 21st century. Perspect Biol Med
17y: a study of 4884 twins and 2509 singletons. Int J 1998;42:44–74.
Obes Relat Metab Disord 1999;23:107–115. 9 Barrett JC, Cardon LR: Evaluating coverage of
5 Koeppen-Schomerus G, Wardle J, Plomin R: A genome-wide association studies. Nat Genet 2006;38:
genetic analysis of weight and overweight in 4-year- 659–662.
old twin pairs. Int J Obes Relat Metab Disord 10 Mutch DM, Clement K: Genetics of human obesity.
2001;25:838–844. Best Pract Res Clin Endocrinol Metab 2006;20:
647–664.
Polygenic Obesity 29
29 Knowler WC, Bennett PH, Hamman RF, Miller M: 39 Mori H, Ikegami H, Kawaguchi Y, Seino S, Yokoi N,
Diabetes incidence and prevalence in Pima Indians: a Takeda J, Inoue I, Seino Y, Yasuda K, Hanafusa T,
19-fold greater incidence than in Rochester, Minnesota. Yamagata K, Awata T, Kadowaki T, Hara K, Yamada
Am J Epidemiol 1978;108:497–505. N, Gotoda T, Iwasaki N, Iwamoto Y, Sanke T, Nanjo
30 Bogardus C, Lillioja S, Howard BV, Reaven G, K, Oka Y, Matsutani A, Maeda E, Kasuga M: The
Mott D: Relationships between insulin secretion, Pro12 → Ala substitution in PPAR-gamma is associ-
insulin action, and fasting plasma glucose concentra- ated with resistance to development of diabetes in the
tion in nondiabetic and noninsulin-dependent dia- general population: possible involvement in impair-
betic subjects. J Clin Invest 1984;74:1238–1246. ment of insulin secretion in individuals with type 2
31 Lillioja S, Mott DM, Howard BV, Bennett PH, Yki- diabetes. Diabetes 2001;50:891–894.
Jarvinen H, Freymond D, Nyomba BL, Zurlo F, 40 Tonjes A, Stumvoll M: The role of the Pro12Ala poly-
Swinburn B, Bogardus C: Impaired glucose tolerance morphism in peroxisome proliferator-activated
as a disorder of insulin action. Longitudinal and receptor gamma in diabetes risk. Curr Opin Clin
cross-sectional studies in Pima Indians. N Engl J Nutr Metab Care 2007;10:410–414.
Med 1988;318:1217–1225. 41 Altshuler D, Hirschhorn JN, Klannemark M,
32 Hanson RL, Ehm MG, Pettitt DJ, Prochazka M, Lindgren CM, Vohl MC, Nemesh J, Lane CR,
Thompson DB, Timberlake D, Foroud T, Kobes S, Schaffner SF, Bolk S, Brewer C, Tuomi T, Gaudet D,
Baier L, Burns DK, Almasy L, Blangero J, Garvey Hudson TJ, Daly M, Groop L, Lander ES: The com-
WT, Bennett PH, Knowler WC: An autosomal mon PPARgamma Pro12Ala polymorphism is asso-
genomic scan for loci linked to type II diabetes melli- ciated with decreased risk of type 2 diabetes. Nat
tus and body-mass index in Pima Indians. Am J Genet 2000;26:76–80.
Hum Genet 1998;63:1130–1138. 42 Lohmueller KE, Pearce CL, Pike M, Lander ES,
33 Norman RA, Tataranni PA, Pratley R, Thompson DB, Hirschhorn JN: Meta-analysis of genetic association
Hanson RL, Prochazka M, Baier L, Ehm MG, Sakul H, studies supports a contribution of common variants
Foroud T, Garvey WT, Burns D, Knowler WC, Bennett to susceptibility to common disease. Nat Genet
PH, Bogardus C, Ravussin E: Autosomal genomic scan 2003;33:177–182.
for loci linked to obesity and energy metabolism in 43 McCarthy MI: Progress in defining the molecular
Pima Indians. Am J Hum Genet 1998;62:659–668. basis of type 2 diabetes mellitus through sus-
34 Semenkovich CF: Regulation of fatty acid synthase ceptibility-gene identification. Hum Mol Genet
(FAS). Prog Lipid Res 1997;36:43–53. 2004;13 Spec No 1:R33–R41. Epub 2004 Jan 13.
35 Loftus TM, Jaworsky DE, Frehywot GL, Townsend CA, 44 Tonjes A, Scholz M, Loeffler M, Stumvoll M:
Ronnett GV, Lane MD, Kuhajda FP: Reduced food Association of Pro12Ala polymorphism in peroxi-
intake and body weight in mice treated with fatty acid some proliferator-activated receptor gamma with
synthase inhibitors. Science 2000;288: 2379–2381. Pre-diabetic phenotypes: meta-analysis of 57 studies
36 Mobbs CV, Makimura H: Block the FAS, lose the fat. on nondiabetic individuals. Diabetes Care 2006;29:
Nat Med 2002;8:335–336. 2489–2497.
37 Kovacs P, Harper I, Hanson RL, Infante AM, 45 Masud S, Ye S: Effect of the peroxisome proliferator
Bogardus C, Tataranni PA, Baier LJ: A novel mis- activated receptor-gamma gene Pro12Ala variant on
sense substitution (Val1483Ile) in the fatty acid syn- body mass index: a meta-analysis. J Med Genet
thase gene (FAS) is associated with percentage of 2003;40:773–780.
body fat and substrate oxidation rates in nondiabetic 46 Andrulionyte L, Peltola P, Chiasson JL, Laakso M:
Pima Indians. Diabetes 2004;53:1915–1919. Single nucleotide polymorphisms of PPARD in com-
38 Körner A, Ma L, Franks PW, Kiess W, Baier LJ, bination with the Gly482Ser substitution of PGC-1A
Stumvoll M, Kovacs P: Sex-specific effect of the and the Pro12Ala substitution of PPARG2 predict
Val1483Ile polymorphism in the fatty acid synthase the conversion from impaired glucose tolerance to
gene (FAS) on body mass index and lipid profile in type 2 diabetes: the STOP-NIDDM trial. Diabetes
Caucasian children. Int J Obes (Lond) 2007;31: 2006;55:2148–2152.
353–358. 47 Florez JC, Jablonski KA, Sun MW, Bayley N,
Kahn SE, Shamoon H, Hamman RF, Knowler WC,
Nathan DM, Altshuler D: Effects of the type 2
diabetes-associated PPARG P12A polymorphism on
progression to diabetes and response to troglitazone.
J Clin Endocrinol Metab 2007;92:1502–1509.
Polygenic Obesity 31
66 Meyre D, Lecoeur C, Delplanque J, Francke S, Vatin 75 Meyre D, Bouatia-Naji N, Tounian A, Samson C,
V, Durand E, Weill J, Dina C, Froguel P: A genome- Lecoeur C, Vatin V, Ghoussaini M, Wachter C,
wide scan for childhood obesity-associated traits in Hercberg S, Charpentier G, Patsch W, Pattou F,
French families shows significant linkage on chro- Charles MA, Tounian P, Clement K, Jouret B, Weill J,
mosome 6q22.31-q23.2. Diabetes 2004;53:803–811. Maddux BA, Goldfine ID, Walley A, Boutin P, Dina
67 Maddux BA, Chang YN, Accili D, McGuinness OP, C, Froguel P: Variants of ENPP1 are associated with
Youngren JF, Goldfine ID: Overexpression of the childhood and adult obesity and increase the risk of
insulin receptor inhibitor PC-1/ENPP1 induces glucose intolerance and type 2 diabetes. Nat Genet
insulin resistance and hyperglycemia. Am J Physiol 2005;37:863–867.
Endocrinol Metab 2006;290:E746–E749. 76 Böttcher Y, Körner A, Reinehr T, Enigk B, Kiess W,
68 Pizzuti A, Frittitta L, Argiolas A, Baratta R, Goldfine Stumvoll M, Kovacs P: ENPP1 variants and haplotypes
ID, Bozzali M, Ercolino T, Scarlato G, Iacoviello L, predispose to early onset obesity and impaired glucose
Vigneri R, Tassi V, Trischitta V: A polymorphism and insulin metabolism in German obese children.
(K121Q) of the human glycoprotein PC-1 gene cod- J Clin Endocrinol Metab 2006;91:4948–4952.
ing region is strongly associated with insulin resis- 77 Tasic I, Milojkovic M, Sunder-Plassmann R, Lazarevic
tance. Diabetes 1999;48:1881–1884. G, Tasic NM, Stefanovic V: The association of PC-1
69 Abate N, Chandalia M, Satija P, Adams-Huet B, (ENPP1) K121Q polymorphism with metabolic syn-
Grundy SM, Sandeep S, Radha V, Deepa R, Mohan V: drome in patients with coronary heart disease. Clin
ENPP1/PC-1 K121Q polymorphism and genetic sus- Chim Acta 2007;377:237–242.
ceptibility to type 2 diabetes. Diabetes 2005;54: 78 Bacci S, Ludovico O, Prudente S, Zhang YY, Di Paola
1207–1213. R, Mangiacotti D, Rauseo A, Nolan D, Duffy J, Fini G,
70 Frittitta L, Baratta R, Spampinato D, Di Paola R, Salvemini L, Amico C, Vigna C, Pellegrini F,
Pizzuti A, Vigneri R, Trischitta V: The Q121 PC-1 Menzaghi C, Doria A, Trischitta V: The K121Q poly-
variant and obesity have additive and independent morphism of the ENPP1/PC-1 gene is associated with
effects in causing insulin resistance. J Clin Endocrinol insulin resistance/atherogenic phenotypes, including
Metab 2001;86:5888–5891. earlier onset of type 2 diabetes and myocardial infarc-
71 Costanzo BV, Trischitta V, Di Paola R, Spampinato D, tion. Diabetes 2005;54:3021–3025.
Pizzuti A, Vigneri R, Frittitta L: The Q allele variant 79 Chen MP, Chung FM, Chang DM, Tsai JC, Huang
(GLN121) of membrane glycoprotein PC-1 interacts HF, Shin SJ, Lee YJ: ENPP1 K121Q polymorphism is
with the insulin receptor and inhibits insulin signal- not related to type 2 diabetes mellitus, features of
ing more effectively than the common K allele vari- metabolic syndrome, and diabetic cardiovascular
ant (LYS121). Diabetes 2001;50: 831–836. complications in a Chinese population. Rev Diabet
72 Kubaszek A, Pihlajamaki J, Karhapaa P, Vauhkonen I, Stud 2006;3:21–30.
Laakso M: The K121Q polymorphism of the PC-1 80 Grarup N, Urhammer SA, Ek J, Albrechtsen A,
gene is associated with insulin resistance but not with Glumer C, Borch-Johnsen K, Jorgensen T, Hansen T,
dyslipidemia. Diabetes Care 2003;26:464–467. Pedersen O: Studies of the relationship between the
73 Willer CJ, Bonnycastle LL, Conneely KN, Duren WL, ENPP1 K121Q polymorphism and type 2 diabetes,
Jackson AU, Scott LJ, Narisu N, Chines PS, Skol A, insulin resistance and obesity in 7,333 Danish white
Stringham HM, Petrie J, Erdos MR, Swift AJ, Enloe subjects. Diabetologia 2006;49:2097–2104.
ST, Sprau AG, Smith E, Tong M, Doheny KF, Pugh 81 Keshavarz P, Inoue H, Sakamoto Y, Kunika K,
EW, Watanabe RM, Buchanan TA, Valle TT, Tanahashi T, Nakamura N, Yoshikawa T, Yasui N,
Bergman RN, Tuomilehto J, Mohlke KL, Collins FS, Shiota H, Itakura M: No evidence for association of
Boehnke M: Screening of 134 single nucleotide poly- the ENPP1 (PC-1) K121Q variant with risk of type 2
morphisms (SNPs) previously associated with type 2 diabetes in a Japanese population. J Hum Genet
diabetes replicates association with 12 SNPs in nine 2006;51:559–566.
genes. Diabetes 2007;56:256–264. 82 Lyon HN, Florez JC, Bersaglieri T, Saxena R, Winckler
74 Bochenski J, Placha G, Wanic K, Malecki M, W, Almgren P, Lindblad U, Tuomi T, Gaudet D,
Sieradzki J, Warram JH, Krolewski AS: New poly- Zhu X, Cooper R, Ardlie KG, Daly MJ, Altshuler D,
morphism of ENPP1 (PC-1) is associated with Groop L, Hirschhorn JN: Common variants in the
increased risk of type 2 diabetes among obese indi- ENPP1 gene are not reproducibly associated with dia-
viduals. Diabetes 2006;55:2626–2630. betes or obesity. Diabetes 2006;55:3180–3184.
83 Weedon MN, Shields B, Hitman G, Walker M,
McCarthy MI, Hattersley AT, Frayling TM: No evi-
dence of association of ENPP1 variants with type 2
diabetes or obesity in a study of 8,089 U.K.
Caucasians. Diabetes 2006;55:3175–3179.
Polygenic Obesity 33
106 Scott LJ, Bonnycastle LL, Willer CJ, Sprau AG, 115 Ross SE, Hemati N, Longo KA, Bennett CN, Lucas
Jackson AU, Narisu N, Duren WL, Chines PS, PC, Erickson RL, MacDougald OA: Inhibition of
Stringham HM, Erdos MR, Valle TT, Tuomilehto J, adipogenesis by Wnt signaling. Science 2000;289:
Bergman RN, Mohlke KL, Collins FS, Boehnke M: 950–953.
Association of transcription factor 7-like 2 (TCF7L2) 116 Cawthorn WP, Heyd F, Hegyi K, Sethi JK: Tumour
variants with type 2 diabetes in a Finnish sample. necrosis factor-alpha inhibits adipogenesis via a
Diabetes 2006;55:2649–2653. beta-catenin/TCF4(TCF7L2)-dependent pathway.
107 Munoz J, Lok KH, Gower BA, Fernandez JR, Hunter Cell Death Differ 2007;14:1361–1373.
GR, Lara-Castro C, De Luca M, Garvey WT: 117 Korinek V, Barker N, Morin PJ, van Wichen D, de
Polymorphism in the transcription factor 7-like 2 Weger R, Kinzler KW, Vogelstein B, Clevers H:
(TCF7L2) gene is associated with reduced insulin Constitutive transcriptional activation by a beta-
secretion in nondiabetic women. Diabetes 2006;55: catenin-Tcf complex in APC-/- colon carcinoma.
3630–3634. Science 1997;275:1784–1787.
108 Wang J, Kuusisto J, Vanttinen M, Kuulasmaa T, 118 Mariadason JM, Bordonaro M, Aslam F, Shi L,
Lindstrom J, Tuomilehto J, Uusitupa M, Laakso M: Kuraguchi M, Velcich A, Augenlicht LH: Down-
Variants of transcription factor 7-like 2 (TCF7L2) regulation of beta-catenin TCF signaling is linked to
gene predict conversion to type 2 diabetes in the colonic epithelial cell differentiation. Cancer Res
Finnish Diabetes Prevention Study and are associ- 2001;61:3465–3471.
ated with impaired glucose regulation and impaired 119 Jung C, Kim RS, Lee SJ, Wang C, Jeng MH: HOXB13
insulin secretion. Diabetologia 2007;50:1192–1200. homeodomain protein suppresses the growth of
109 Cauchi S, Meyre D, Dina C, Choquet H, Samson C, prostate cancer cells by the negative regulation of T-
Gallina S, Balkau B, Charpentier G, Pattou F, cell factor 4. Cancer Res 2004;64: 3046–3051.
Stetsyuk V, Scharfmann R, Staels B, Fruhbeck G, 120 Burwinkel B, Shanmugam KS, Hemminki K, Meindl
Froguel P: Transcription Factor TCF7L2 Genetic A, Schmutzler RK, Sutter C, Wappenschmidt B,
Study in the French Population: Expression in Kiechle M, Bartram CR, Frank B: Transcription fac-
human {beta}-cells and adipose tissue and strong tor 7-like 2 (TCF7L2) variant is associated with
association with type 2 diabetes. Diabetes 2006;55: familial breast cancer risk: a case-control study.
2903–2908. BMC Cancer 2006;6:268.
110 Loos RJ, Franks PW, Francis RW, Barroso I, Gribble 121 Duval A, Rolland S, Tubacher E, Bui H, Thomas G,
FM, Savage DB, Ong KK, O’Rahilly S, Wareham NJ: Hamelin R: The human T-cell transcription factor-4
TCF7L2 polymorphisms modulate proinsulin levels gene: structure, extensive characterization of alter-
and {beta}-cell function in a British Europid popu- native splicings, and mutational analysis in colorec-
lation. Diabetes 2007. tal cancer cell lines. Cancer Res 2000;60:3872–3879.
111 Field SF, Howson JM, Smyth DJ, Walker NM, 122 Duval A, Gayet J, Zhou XP, Iacopetta B, Thomas G,
Dunger DB, Todd JA: Analysis of the type 2 diabetes Hamelin R: Frequent frameshift mutations of the
gene, TCF7L2, in 13,795 type 1 diabetes cases and TCF-4 gene in colorectal cancers with microsatellite
control subjects. Diabetologia 2007;50:212–213. instability. Cancer Res 1999;59:4213–4215.
112 Cauchi S, Vaxillaire M, Choquet H, Durand E, 123 Korinek V, Barker N, Moerer P, van Donselaar E,
Duval A, Polak M, Froguel P: No major contribu- Huls G, Peters PJ, Clevers H: Depletion of epithelial
tion of TCF7L2 sequence variants to maturity onset stem-cell compartments in the small intestine of
of diabetes of the young (MODY) or neonatal dia- mice lacking Tcf-4. Nat Genet 1998;19:379–383.
betes mellitus in French white subjects. Diabetologia 124 Sladek R, Rocheleau G, Rung J, Dina C, Shen L, Serre
2007;50:214–216. D, Boutin P, Vincent D, Belisle A, Hadjadj S, Balkau
113 Pearson ER, Donnelly LA, Kimber C, Whitley A, B, Heude B, Charpentier G, Hudson TJ, Montpetit A,
Doney AS, McCarthy MI, Hattersley AT, Morris Pshezhetsky AV, Prentki M, Posner BI, Balding DJ,
AD, Palmer CN: Variation in TCF7L2 influences Meyre D, Polychronakos C, Froguel P: A genome-
therapeutic response to sulfonylureas: A GoDARTs wide association study identifies novel risk loci for
study. Diabetes 2007;56:2178–2182. type 2 diabetes. Nature 2007;445: 881–885.
114 Singh R, Artaza JN, Taylor WE, Braga M, Yuan X,
Gonzalez-Cadavid NF, Bhasin S: Testosterone inhibits
adipogenic differentiation in 3T3-L1 cells: nuclear
translocation of androgen receptor complex with
beta-catenin and T-cell factor 4 may bypass canonical
Wnt signaling to down-regulate adipogenic transcrip-
tion factors. Endocrinology 2006;147:141–154.
Polygenic Obesity 35
133 Vockley J, Rinaldo P, Bennett MJ, Matern D, 137 Franks PW, Luan J, Browne PO, Harding AH,
Vladutiu GD: Synergistic heterozygosity: disease O’Rahilly S, Chatterjee VK, Wareham NJ: Does per-
resulting from multiple partial defects in one or oxisome proliferator-activated receptor gamma
more metabolic pathways. Mol Genet Metab 2000; genotype (Pro12ala) modify the association of phys-
71:10–18. ical activity and dietary fat with fasting insulin
134 Phillips C, Lopez-Miranda J, Perez-Jimenez F, level? Metabolism 2004;53:11–16.
McManus R, Roche HM: Genetic and nutrient 138 Jirtle RL, Skinner MK: Environmental epigenomics
determinants of the metabolic syndrome. Curr and disease susceptibility. Nat Rev Genet 2007;8:
Opin Cardiol 2006;21:185–193. 253–262.
135 Luan J, Browne PO, Harding AH, Halsall DJ, 139 Challis BG, Yeo GS: Past, present and future strate-
O’Rahilly S, Chatterjee VK, Wareham NJ: Evidence gies to study the genetics of body weight regulation.
for gene-nutrient interaction at the PPARgamma Brief Funct Genomic Proteomic 2002;1:290–304.
locus. Diabetes 2001;50:686–689.
136 Pisabarro RE, Sanguinetti C, Stoll M, Prendez D:
High incidence of type 2 diabetes in peroxisome
proliferator-activated receptor gamma2 Pro12Ala
carriers exposed to a high chronic intake of trans
fatty acids and saturated fatty acids. Diabetes Care
2004;27:2251–2252.
In a recent study, the putative role of the Fto protein has been unraveled. By means of bioinformatic
analysis [1, 2] and confirmed [1] by functional in vitro experiments, Fto has been shown to act as
an Fe(II) and 2-oxoglutarate oxygenase. With these enzymatic properties, Fto may be involved in
DNA repair, fatty acid metabolism and posttranslational modifications. Further support that this
enzymatic function may be relevant to the regulation of body weight derives from the observation
that Fto mRNA levels were reduced by fasting and not rescued by leptin supplementation.
1 Gerken T, Girard CA, Tung YC, Webby CJ, Saudek V, 2 Sanchez-Pulido L, Andrade-Navarro MA: The FTO
Hewitson KS, Yeo GS, McDonough MA, Cunliffe S, (fat mass and obesity associated) gene codes for a
McNeill LA, Galvanovskis J, Rorsman P, Robins P, novel member of the non-heme dioxygenase super-
Prieur X, Coll AP, Ma M, Jovanovic Z, Farooqi IS, family. BMC Biochem 2007;8:23.
Sedgwick B, Barroso I, Lindahl T, Ponting CP, Ashcroft
FM, O’Rahilly S, Schofield CJ: The obesity-associated
FTO gene encodes a 2-oxoglutarate dependent nucleic
acid demethylase. Science 2007; Epub ahead of print.
Abstract
There are numerous reports of multi-system genetic disorders with obesity. Many have a characteristic pre-
sentation and several, an overlapping phenotype indicating the likelihood of a shared common underlying
mechanism or pathway. By understanding the genetic causes and functional perturbations of such syn-
dromes we stand to gain tremendous insight into obesogenic pathways. In this review we focus particularly
on Bardet-Biedl syndrome, whose molecular genetics and cell biology has been elucidated recently, and
Prader-Willi syndrome, the commonest obesity syndrome due to loss of imprinted genes on 15q11–13. We
also discuss highlights of other genetic obesity syndromes including Alstrom syndrome, Cohen syndrome,
Albright’s hereditary osteodystrophy (pseudohypoparathyroidism), Carpenter syndrome, MOMO syn-
drome, Rubinstein-Taybi syndrome, cases with deletions of 6q16, 1p36, 2q37 and 9q34, maternal uni-
parental disomy of chromosome 14, fragile X syndrome and Börjeson-Forssman-Lehman syndrome.
Copyright © 2008 S. Karger AG, Basel
Bardet-Biedl Syndrome
c e
Fig. 1. BBS. a, b Post-axial polydactyly in hands and feet of the same child. c Brachydactyly in hands
of an adult. Note the postaxial scars. d High arched palate is common in BBS. e Dental anomalies
frequently include crowded dentition with hypodontia and short roots.
Twelve genes have now been identified for which there is little evidence of any phe-
notype-genotype correlation. The BBS genes (BBS1–12) have few sequence similarities
to each other or other protein groups. Three, BBS6, BBS10 and BBS12 have strong
homology with the type II group of chaperones and account for around 30% of all muta-
tions [2–4]. Only BBS3/ARL6 (a member of the Ras superfamily of small GTP-binding
proteins) and BBS11/TRIM32 (an E3 ubiquitin ligase) encode known proteins [5–7].
Recent evidence suggests that BBS is probably caused by dysfunction of primary
cilia and the intraflagellar transport (IFT) process. All BBS proteins studied thus far
localise to the cilium/basal body/centrosome complex (fig. 2). In mammalian cul-
tured cells several BBS proteins localise either to the basal body and pericentriolar
region or the ciliary axoneme [7–13].
Studies indicate that many BBS proteins function in microtubular processes such
as IFT as demonstrated in several Bbs mouse mutants, in which each develops severe
retinal degeneration similar to patients [11, 14–17]. In photoreceptors, rhodopsin
relies on IFT for transport to the outer segment – in Bbs mutants rhodopsin accumu-
lates in the cell body triggering apoptosis [11, 17]. Anosmia was recently reported in
Bbs1 and Bbs4 mutants arising from depletion of olfactory proteins in the ciliary layer
of olfactory neurones [15]. Subsequently, anosmia was demonstrated in BBS patients,
38 Goldstone · Beales
PCM-1
CHE-11
OSM-5 (polaris)
IFT particle
Transition fibres
BBS7 BBS8
BBS4
Cargo
Dynein/dynactin
Kinesin
BBS8
BBS7
BBS4
BBS7
BBS8
BBS4
BBS1 Transition zone (basal body)
BBS6
BBS2
BBS3 BBS5
BBS4
Centriolar
satellites Microtubules
BBS4
Centrosome
Centrioles
Fig. 2. BBS proteins. BBS proteins are found in the basal body, centrosome and occasionally the cil-
iary axoneme. Many are directly involved with IFT, a process dependent on the molecular motors,
dynein and kinesin. BBS4 is thought to behave as an adaptor protein, facilitating loading of cargo
prior to dynein (retrograde) transport.
Alström Syndrome
Alström syndrome (ALS, OMIM 203800) is a rare recessive disorder typically pre-
senting with early-onset obesity, hyperinsulinaemia (often with acanthosis nigricans)
and type 2 diabetes mellitus, dilated cardiomyopathy, short stature and male hypogo-
nadism. Infants usually display nystagmus and photophobia, eventually progressing
40 Goldstone · Beales
to cone and rod photoreceptor degeneration making it a key differential diagnosis
with the BBS [27, 28]. It may be associated with hepatic dysfunction, hepatic steatosis
and hyperlipidaemia. In addition, ALS patients develop variable sensorineural hear-
ing loss owing to cochlear neuronal degeneration [28, 29].
Rapid weight gain occurs during infancy but tends to plateau in adolescence with
a truncal distribution. Despite the improvement in adiposity and BMI with age, the
insulin resistance continues to increase and thus ALS may represent a monogenic
model of the metabolic syndrome [29]. There are numerous associated endocrine
disturbances including growth hormone (GH) deficiency and hyper- or hypogo-
nadotrophic hypogonadism in males [30, 31]. Amongst females with ALS, hirsutism,
precocious puberty, and amenorrhoea have been reported. Like BBS up to 50% will
have renal and/or urological dysfunction.
The underlying gene, ALMS1 was discovered in 2002 on 2p13 [32, 33]. Neither the
predicted gene nor protein sequence has similarity to any other genes, although there
are several conserved sequence motifs of limited functional significance. Of interest
are the presence of a large 8-kb exon containing a tandem-repeat domain and in exon
1, a polyglutamic acid/polyalanine tract, the length of which does not appear to impact
on the AS phenotype. There do not appear to be any phenotype-genotype correlations.
ALMS1 is ubiquitously expressed throughout all organ tissues [32], and is a compo-
nent protein of the centrosome with basal body localisation suggesting involvement in
ciliary function and perhaps explaining the phenotypic overlap with BBS [33].
Common variations in the ALMS1 gene were not associated with type 2 diabetes
mellitus in two studies of a Dutch and UK population [34, 35].
Cohen Syndrome
Carpenter Syndrome
42 Goldstone · Beales
hypoparathyroidism secondary to end organ resistance to parathyroid hormone, a
component of which they termed ‘pseudo-hypoparathyroidism’ (PHP). Since then
several cases of AHO without end-organ resistance have also be reported and termed
‘pseudo-pseudohypoparathyroidism’ (PPHP). PHP is further subdivided into types Ia,
Ib, Ic and type II. The PHP type Ia and PPHP forms of AHO are caused by inactivating
mutations in the tissue-specifically imprinted gene GNAS1 resulting in reduction of
the encoded Gs protein [53].
Despite normal or even low birth weights, 50–65% of AHO patients develop gen-
eralised obesity. The aetiology of the obesity is far from clear but there are a number
of possible mechanisms. The melanocortin receptor (MC4R), mutations in which are
one of the most common causes of genetic obesity, is transduced by Gs, as are many
of the other G-protein-coupled seven transmembrane receptors that mediate anorex-
igenic signals from hormones and other neurotransmitters. Loss of such anorexigenic
signals such as through MC4R should produce hyperphagia, but this has not been
widely studied in obese AHO individuals [54].
The observations that Gs represses differentiation of fibroblasts (3T3L1 pre-
adipocytes) into adipocytes, and patients with PHP1a may have reduced cAMP
responses to -adrenergic stimulation in fat cell membranes, reduced basal and
adrenergic-stimulated glycerol production and reduced circulating levels of nora-
drenaline suggest that increased adipogenesis and reduced lipolysis and sympathetic
activity may also contribute to obesity in AHO [53].
Rubinstein-Taybi Syndrome
Prader-Willi Syndrome
PWS AS Non-imprinted
B
A
38
HB 437
HB -436
38
HB II-13
II-4
II-4
GABR
II-
HB
HB
P
ND 2
MA 3
3 5 3
L
1
UBE3A
RN
1
2
GE
FIP
P5
PA
PA
cen
MK
GC
CY
NI
NI
tel
BP1 BP2
exons 1-3 4 -10 13-20 21 52 62 63 142 144 148 ATP10C BP3
SNURF SmN PAR-5 PAR-7 IPW PAR-1 PAR-4
UBE3A-AS
PWCR1
SNURF-SNRPN
44 Goldstone · Beales
behavioural problem with temper tantrums, obsessive compulsive behaviours, skin
picking, stubbornness, rigidity, stealing and lying; central and obstructive sleep
apnoea; eye abnormalities such as esotropis and myopia; thick viscous saliva and
speech articulation defect; high pain threshold; decreased vomiting; altered tempera-
ture sensitivity; scoliosis or kyphosis [57–60].
PWS Genetics
PWS arises from the lack of expression of genes within the paternally derived chro-
mosome 15q11-q13 which are silenced (imprinted) on the maternally derived chro-
mosome [see 59, 61 for references]. These include NDN, MAGEL2, MKRN3
(previously called ZFP127), the SNURF-SNRPN locus which extends over 460 kb
encoding at least 148 exons and several repeating intronic C/D box small nucleolar
RNAs (snoRNAs) including HBII-52 and HBII-85 (fig. 3b).
The finding of patients with smaller microdeletions and balanced translocations
has permitted narrowing of the critical PWS region from 4.5 MB to less than 4.3 kb,
spanning the promoter and exon 1 of the SNRPN gene and demonstrated the impor-
tance of some snoRNAs in the PWS phenotype [61]. The promotor and first exon of
the SNURF-SNRPN gene locus is an integral part of the imprinting centre (IC) in the
PWS chromosomal region (fig. 3b). The mouse PWS gene homologues are particu-
larly expressed throughout the developing brain, particularly the hypothalamus, and
there is also embryonic and post-natal expression of Ndn and Magel2 outside the
brain [62].
The molecular biology of the genes within the PWS critical region has not been
fully established. The SmN product of the SNRPN locus (exons 4–10) is involved in
RNA alternative splicing, but the role of other transcripts from the SNRPN locus,
including IPW, PAR-1, -4 and -7, are poorly defined [61] (fig. 3b). The locus also
encodes an anti-sense transcript for the paternally-imprinted UBE3A gene involved
in Angelman syndrome.
Necdin has a pivotal role in neuronal differentiation and survival, prevention of
apoptosis and axonal growth, and may interact with several neurotrophic and cell
cycle-regulatory transcription factors and hence proapoptotic genes such as TrkA,
p75, E2F1, Cdc2, p53, hnRNP U, and NEFA [59, 63]. Necdin and Magel2 proteins
can both bind to and prevent proteasomal degradation of Fez1, a fasciculation and
elongation protein implicated in axonal outgrowth and kinesin-mediated transport,
which also binds to the BBS protein BBS4 at or near centrosomes [64]. Ndn-
deficient mice exhibit neonatal lethality with respiratory distress; an abnormal res-
piratory rhythm-generating centre in the medulla; increased skin-scraping activity;
improved spatial learning; hypothalamic structural abnormalities with reduced
oxytocin and LHRH cell number (but preserved fertility); abnormal axonal out-
growth and fasciculation in embryos, including serotoninergic, noradrenergic, sym-
pathetic (e.g. diaphragmatic and superior cervical ganglion), retinal ganglion cell,
46 Goldstone · Beales
hyperphagia and severe obesity in PWS compared to monogenic causes of obesity
such as leptin deficiency or melanocortin-4 receptor mutations is unknown.
Obesity management involves early institution of a low-calorie, well-balanced diet,
with regular exercise, rigorous supervision, restriction of access to food and money with
appreciation of legal and ethical obligations, appropriate psychological and behavioural
counselling of the patient and family [77]. Group homes specifically designed for indi-
viduals with PWS, where available, have been particularly successful in management of
these problems during adulthood. Anecdotally, pharmacological treatment, including
available anorexigenic agents, has not been of benefit in treating hyperphagia, though
there are few published control studies. Restrictive bariatric surgery, such as gastric
banding or bypass, have not been shown to reduce hyperphagia or achieve long-term
weight reduction and are associated with unacceptable morbidity and mortality, but
some of the reports using biliopancreatic diversion which produces intestinal malab-
sorption have reported successful weight loss though with frequent complications [78].
Body composition studies show both increased body fat and reduced muscle in
PWS [79]. Magnetic resonance imaging has found that PWS adults of both sexes have
less visceral adiposity than expected for their overall adiposity [80, 81] (fig. 4a). This
may explain the relative hypoinsulinaemia and lower triglyceride levels with preser-
vation of insulin sensitivity and protective elevation in adiponectin levels in patients
with PWS despite their overall obesity [80, 82, 83] (fig. 4b, c).
Obesity, hypersomnolence and persistent poor muscle strength contribute to
reduced physical activity in PWS. Resting metabolic rate is reduced relative to body
size, as a result of the abnormal body composition, which further contributes to a
reduction in 24-hour energy expenditure [79]. Increased physical activity and exer-
cise programs improve body composition in PWS.
Spontaneous or pharmacologically stimulated GH secretion and IGF-I levels are
reduced in PWS children and adults, and the GH deficiency is independent of obesity
[60]. In PWS children, GH therapy is now licensed and significantly improves height
velocity and final height [60, 84]. GH significantly decreases total body fat, increases
lean body mass, lipolysis and resting energy expenditure, and improves physical
strength and agility in children and infants with PWS, and may also have neurodevel-
opmental benefits [60, 84, 85]. There may also be a potential benefit of lower GH
doses to improve body composition in PWS adults.
60 5 1,200
Percent body fat
50 b
4 a, b a, b 1,000
Ghrelin (pM)
b b
HOMA-IR
40 800
3
30 a 600 a, b a, b
2 a
20 400
10 1 200
0 0 0
b NO OB CRHO PWS c NO OB CRHO PWS d NO OB CRHO PWS
Fig. 4. Reduced visceral adiposity, preserved insulin sensitivity and hyperghrelinaemia in PWS.
a A T1-weighted MRI scan at the level of the lower abdomen, showing less visceral adiposity in a PWS
adult female (total adipose tissue, AT, volume 86.9 l, MRI total body fat 54.4%, visceral AT 4.4% of total
AT, visceral AT:subcutaneous AT ratio 0.048) compared to a similarly obese control adult female (total
AT volume 86.4 l, MRI total body fat 50.5%, visceral AT 9.3% of total AT, visceral AT:subcutaneous AT
ratio 0.111). With permission from Goldstone et al. [80]. b–d Mean SEM values for percent body fat
(b), homeostasis model insulin resistance index (HOMA-IR; c) and fasting plasma ghrelin levels (d) in
non-obese (NO, n
15) and obese (OB, n
16) controls, craniopharyngioma subjects with hypo-
thalamic obesity (CRHO, n
9) and subjects with PWS (n
26). a p 0.01 vs. PWS, b p 0.01 vs. NO.
Despite similar degrees of obesity, subjects with PWS have increased fasting plasma ghrelin and pre-
served insulin sensitivity, compared to the other two obese groups. With permission from Goldstone
et al. [82].
Subjects with PWS have marked elevations in the stomach-derived orexigenic hor-
mone ghrelin for their obesity, and increased density of ghrelin immunostaining in
the stomach, though plasma levels do fall appropriately after food [82, 88, 89] (fig. 4d).
This may be explained at least partly by their relative hypoinsulinaemia (fig. 4). A
primary importance for hyperghrelinaemia is questioned by the overlap of ghrelin
levels in PWS with lean subjects despite the former’s near ubiquitous hyperphagia,
and the inability of acute normalisation of ghrelin levels in PWS using somatostatin
to reduce appetite [90]. However somatostatin will also suppress secretion of a wide
variety of anorexigenic gut hormones that might counteract any beneficial effect of
lowering orexigenic ghrelin [90]. Fasting and post-prandial levels of the anorexigenic
48 Goldstone · Beales
hormone pancreatic polypeptide are also reduced in PWS children and adults which
may also contribute to hyperphagia [91, 92]. It seems likely that in addition to these
hormonal abnormalities in PWS, there are overriding brain defects, including hypo-
thalamic, which lead to resistance to peripheral satiety signals [59]. Infusion of PP to
subjects with PWS has only a small anorexigenic effect [93]. The possibility of thera-
peutic avenues for reducing hyperphagia in PWS may depend on the existence of rel-
ative rather than absolute resistance to peripheral satiety signals.
Deletion 6q16
Deletion 1p36
Obesity and/or hyperphagia has also been reported in 23% of 14 evaluated patients
with monosomy 1p36, in whom developmental delay, hypotonia, growth delay, feed-
ing difficulties in infancy, epilepsy, hearing loss, hypermetropia, orofacial clefting
abnormalities, structural heart defects and dilated cardiomyopathy, micro- and
50 Goldstone · Beales
brachycephaly, deep set eyes, flat nasal bridge and nose, pointed chin, thickened ear
helices, asymmetric ears and short fifth finger are also reported [121].
Deletion 2q37
Deletion 9q34.3
Fragile X Syndrome
Börjeson-Forssman-Lehmann Syndrome
52 Goldstone · Beales
Obesity Syndromes without Identified Genetic Cause
Moretti-Ferreira et al. [139] reported 2 unrelated children with similar clinical fea-
tures comprising truncal obesity, mental retardation and retinal coloboma and nys-
tagmus (MOMO, OMIM 157980). Facially, features were unremarkable with
hypertelorism, downslanting palpebral fissures, a prominent forehead, and a broad
nasal root. A potential third case was reported by Zannolli et al. [140] describing a 5-
year-old girl with mild learning impairment, morbid obesity, macrocephaly, right
optic disc coloboma and left choroidal coloboma, and recurvation of the femur. The
cause is unknown.
Conclusions
References
1 Beales PL, Elcioglu N, Woolf AS, Parker D, Flinter 2 Katsanis N, Beales PL, Woods MO, Lewis RA, Green
FA: New criteria for improved diagnosis of Bardet- JS, Parfrey PS, Ansley SJ, Davidson WS, Lupski JR:
Biedl syndrome: results of a population survey. J Med Mutations in MKKS cause obesity, retinal dystrophy
Genet 1999;36:437–446. and renal malformations associated with Bardet-
Biedl syndrome. Nat Genet 2000;26: 67–70.
54 Goldstone · Beales
20 Eichers ER, Abd-El-Barr MM, Paylor R, Lewis RA, 31 Alter CA, Moshang T Jr: Growth hormone defi-
Bi W, Lin X, Meehan TP, Stockton DW, WU SM, ciency in two siblings with Alstrom syndrome. Am J
Lindsay E, Justice MJ, Beales PL, Katsanis N, Lupski Dis Child 1993;147:97–99.
JR: Phenotypic characterization of Bbs4 null mice 32 Collin GB, Marshall JD, Ikeda A, So WV, Russell-
reveals age-dependent penetrance and variable expres- Eggitt I, Maffei P, Beck S, Boerkoel CF, Sicolo N,
sivity. Hum Genet 2006. Martin M, Nishina PM, Naggert JK: Mutations in
21 Croft JB, Morrell D, Chase CL, Swift M: Obesity in ALMS1 cause obesity, type 2 diabetes and neurosen-
heterozygous carriers of the gene for the Bardet- sory degeneration in Alstrom syndrome. Nat Genet
Biedl syndrome. Am J Med Genet 1995;55:12–15. 2002;31:74–78.
22 Croft JB, Swift M: Obesity, hypertension, and renal 33 Hearn T, Renforth GL, Spalluto C, Hanley NA, Piper
disease in relatives of Bardet-Biedl syndrome sibs. K, Brickwood S, White C, Connolly V, Taylor JF,
Am J Med Genet 1990;36:37–42. Russell-Eggitt I, Bonneau D, Walker M, Wilson DI:
23 Reed DR, Ding Y, Xu W, Cather C, Price RA: Human Mutation of ALMS1, a large gene with a tandem
obesity does not segregate with the chromosomal repeat encoding 47 amino acids, causes Alstrom syn-
regions of Prader-Willi, Bardet-Biedl, Cohen, drome. Nat Genet 2002;31:79–83.
Borjeson or Wilson-Turner syndromes. Int J Obes 34 Patel S, Minton JA, Weedon MN, Frayling TM,
Relat Metab Disord 1995;19:599–603. Ricketts C, Hitman GA, McCarthy MI, Hattersley
24 Andersen KL, Echwald SM, Larsen LH, Hamid YH, AT, Walker M, Barrett TG: Common variations in
Glumer C, Jorgensen T, Borch-Johnsen K, Andersen T, the ALMS1 gene do not contribute to susceptibility
Sorensen TI, Hansen T, Pedersen O: Variation of the to type 2 diabetes in a large white UK population.
McKusick-Kaufman gene and studies of relationships Diabetologia 2006;49:1209–1213.
with common forms of obesity. J Clin Endocrinol 35 Hart LM, Maassen JA, Dekker JM, Heine RJ: Lack of
Metab 2005;90:225–230. association between gene variants in the ALMS1
25 Fan Y, Rahman P, Peddle L, Hefferton D, Gladney N, gene and Type 2 diabetes mellitus. Diabetologia 2003;
Moore SJ, Green JS, Parfrey PS, Davidson WS: 46:1023–1024.
Bardet-Biedl syndrome 1 genotype and obesity in the 36 Chandler KE, Kidd A, Al Gazali L, Kolehmainen J,
Newfoundland population. Int J Obes Relat Metab Lehesjoki AE, Black GC, Clayton-Smith J: Diagnostic
Disord 2004;28:680–684. criteria, clinical characteristics, and natural history of
26 Benzinou M, Walley A, Lobbens S, Charles MA, Cohen syndrome. J Med Genet 2003; 40:233–241.
Jouret B, Fumeron F, Balkau B, Meyre D, Froguel P: 37 Kivitie-Kallio S, Norio R: Cohen syndrome: essential
Bardet-Biedl syndrome gene variants are associated features, natural history, and heterogeneity. Am J
with both childhood and adult common obesity in Med Genet 2001;102:125–135.
French Caucasians. Diabetes 2006;55:2876–2882. 38 Cohen MM Jr, Hall BD, Smith DW, Graham CB,
27 Russell-Eggitt IM, Clayton PT, Coffey R, Kriss A, Lampert KJ: A new syndrome with hypotonia, obe-
Taylor DS, Taylor JF: Alstrom syndrome. Report of sity, mental deficiency, and facial, oral, ocular, and
22 cases and literature review. Ophthalmology 1998; limb anomalies. J Pediatr 1973;83:280–284.
105:1274–1280. 39 Fryns JP, Legius E, Devriendt K, Meire F, Standaert L,
28 Marshall JD, Bronson RT, Collin GB, Nordstrom AD, Baten E, van den Berghe H: Cohen syndrome: the
Maffei P, Paisey RB, Carey C, Macdermott S, Russell- clinical symptoms and stigmata at a young age. Clin
Eggitt I, Shea SE, Davis J, Beck S, Shatirishvili G, Genet 1996;49:237–241.
Mihai CM, Hoeltzenbein M, Pozzan GB, Hopkinson 40 Seifert W, Holder-Espinasse M, Spranger S,
I, Sicolo N, Naggert JK, Nishina PM: New Alstrom Hoeltzenbein M, Rossier E, Dollfus H, Lacombe D,
syndrome phenotypes based on the evaluation of 182 Verloes A, Chrzanowska KH, Maegawa GH, Chitayat
cases. Arch Intern Med 2005; 165:675–683. D, Kotzot D, Huhle D, Meinecke P, Albrecht B,
29 Minton JA, Owen KR, Ricketts CJ, Crabtree N, Mathijssen I, Leheup B, Raile K, Hennies HC, Horn
Shaikh G, Ehtisham S, Porter JR, Carey C, Hodge D, D: Mutational spectrum of COH1 and clinical het-
Paisey R, Walker M, Barrett TG: Syndromic obesity erogeneity in Cohen syndrome. J Med Genet 2006;
and diabetes: changes in body composition with age 43:e22.
and mutation analysis of ALMS1 in 12 United 41 Norio R, Raitta C, Lindahl E: Further delineation of
Kingdom kindreds with Alstrom syndrome. J Clin the Cohen syndrome; report on chorioretinal dystro-
Endocrinol Metab 2006;91:3110–3116. phy, leukopenia and consanguinity. Clin Genet
30 Maffei P, Munno V, Marshall JD, Scandellari C, 1984;25:1–14.
Sicolo N: The Alstrom syndrome: is it a rare or 42 Kivitie-Kallio S, Summanen P, Raitta C, Norio R:
unknown disease? Ann Ital Med Int 2002;17: Ophthalmologic findings in Cohen syndrome. A long-
221–228. term follow-up. Ophthalmology 2000;107:1737–1745.
56 Goldstone · Beales
68 Andrieu D, Meziane H, Marly F, Angelats C, 79 Goldstone AP, Brynes AE, Thomas EL, Bell JD, Frost
Fernandez PA, Muscatelli F: Sensory defects in G, Holland A, Ghatei MA, Bloom SR: Resting meta-
Necdin deficient mice result from a loss of sensory bolic rate, plasma leptin concentrations, leptin recep-
neurons correlated within an increase of develop- tor expression, and adipose tissue measured by
mental programmed cell death. BMC Dev Biol 2006; whole-body magnetic resonance imaging in women
6:56. with Prader-Willi syndrome. Am J Clin Nutr 2002;
69 Kishore S, Stamm S: The snoRNA HBII-52 regulates 75:468–475.
alternative splicing of the serotonin receptor 2C. 80 Goldstone AP, Thomas EL, Brynes AE, Bell JD, Frost
Science 2006;311:230–232. G, Saeed N, Hajnal JV, Howard JK, Holland A, Bloom
70 Vitali P, Basyuk E, Le Meur E, Bertrand E, Muscatelli SR: Visceral adipose tissue and metabolic complica-
F, Cavaille J, Huttenhofer A: ADAR2-mediated edit- tions of obesity are reduced in Prader-Willi syn-
ing of RNA substrates in the nucleolus is inhibited by drome female adults: evidence for novel influences
C/D small nucleolar RNAs. J Cell Biol 2005;169: on body fat distribution. J Clin Endocrinol Metab
745–753. 2001;86:4330–4338.
71 Boer H, Holland A, Whittington J, Butler J, Webb T, 81 Goldstone AP, Unmehopa UA, Thomas EL et al:
Clarke D: Psychotic illness in people with Prader Hypothalamic neuropeptides and regulation of fat
Willi syndrome due to chromosome 15 maternal mass in Prader-Willi syndrome; in Eiholzer U,
uniparental disomy. Lancet 2002;359:135–136. l’Allemand D, Zipf W (eds): Prader-Willi Syn-
72 Chai JH, Locke DP, Greally JM, Knoll JH, Ohta T, drome as a Model for Obesity. Basel, Karger, 2003,
Dunai J, Yavor A, Eichler EE, Nicholls RD: Identifi- pp 31–43.
cation of four highly conserved genes between break- 82 Goldstone AP, Patterson M, Kalingag N, Ghatei MA,
point hotspots BP1 and BP2 of the Prader-Willi/ Brynes AE, Bloom SR, Grossman AB, Korbonits M:
Angelman syndromes deletion region that have Fasting and post-prandial hyperghrelinemia in
undergone evolutionary transposition mediated Prader-Willi syndrome is partially explained by
by flanking duplicons. Am J Hum Genet 2003;73: hypoinsulinemia, and is not due to peptide YY 3–36
898–925. deficiency or seen in hypothalamic obesity due to
73 Butler MG, Bittel DC, Kibiryeva N, Talebizadeh Z, craniopharyngioma. J Clin Endocrinol Metab 2005;
Thompson T: Behavioral differences among subjects 90:2681–2690.
with Prader-Willi syndrome and type I or type II 83 Kennedy L, Bittel DC, Kibiryeva N, Kalra SP, Torto R,
deletion and maternal disomy. Pediatrics 2004; Butler MG: Circulating adiponectin levels, body
113:565–573. composition and obesity-related variables in Prader-
74 Dudley O, Muscatelli F: Clinical evidence of Willi syndrome: comparison with obese subjects. Int
intrauterine disturbance in Prader-Willi syndrome, a J Obes (Lond) 2006;30:382–387.
genetically imprinted neurodevelopmental disorder. 84 Allen DB, Carrel AL: Growth hormone therapy for
Early Hum Dev 7 A.D.;83:471–478. Prader-Willi syndrome: a critical appraisal. J Pediatr
75 Stevenson DA, Heinemann J, Angulo M, Butler MG, Endocrinol Metab 2004;17(suppl 4):1297–1306.
Loker J, Rupe N, Kendell P, Clericuzio CL, Scheimann 85 Myers SE, Whitman BY, Carrel AL, Moerchen V,
AO: Deaths due to choking in Prader-Willi syn- Bekx MT, Allen DB: Two years of growth hormone
drome. Am J Med Genet A 2007;143: 484–487. therapy in young children with Prader-Willi syn-
76 Wharton R, Wang T, GraemeCook F, Briggs S, Cole drome: Physical and neurodevelopmental benefits.
R: Acute idiopathic gastric dilatation with gastric Am J Med Genet 2007;143:443–448.
necrosis in individuals with Prader-Willi syndrome. 86 Holland AJ, Treasure J, Coskeran P, Dallow J, Milton
Am J Med Genet 1997;73:437–441. N, Hillhouse E: Measurement of excessive appetite
77 Eiholzer U: A comprehensive approach to limiting and metabolic changes in Prader-Willi syndrome. Int
weight gain and to normalising body composition in J Obes 1993;17:527–532.
Prader-Willi syndrome; in Eiholzer U, l’Allemand D, 87 Choe YH, Jin DK, Kim SE, Song SY, Paik KH, Park HY,
Zipf W (eds): Prader-Willi syndrome as a Model for Oh YJ, Kim AH, Kim JS, Kim CW, Chu SH, Kwon EK,
Obesity. Basel, Karger, 2003, pp 211–221. Lee KH: Hyperghrelinemia does not accelerate gastric
78 Papavramidis ST, Kotidis EV, Gamvros O: Prader- emptying in Prader-Willi syndrome patients. J Clin
Willi syndrome-associated obesity treated by bil- Endocrinol Metab 2005;90: 3367–3370.
iopancreatic diversion with duodenal switch. Case 88 Cummings DE, Clement K, Purnell JQ, Vaisse C,
report and literature review. J Pediatr Surg 2006;41: Foster KE, Frayo RS, Schwartz MW, Basdevant A,
1153–1158. Weigle DS: Elevated plasma ghrelin levels in Praderh
Willi syndrome. Nat Med 2002;8:643–644.
58 Goldstone · Beales
112 O’Neill MA, Farooqi IS, Wevrick R: Evaluation 123 Wilson LC, Leverton K, Oude Luttikhuis ME, Oley
of Prader-Willi Syndrome gene MAGEL2 in CA, Flint J, Wolstenholme J, Duckett DP, Barrow
severe childhood-onset obesity. Obes Res 2005;13: MA, Leonard JV, Read AP: Brachydactyly and men-
1841–1842. tal retardation: an Albright hereditary osteodystro-
113 Meyre D, Lecoeur C, Delplanque J, Francke S, Vatin phy-like syndrome localized to 2q37. Am J Hum
V, Durand E, Weill J, Dina C, Froguel P: A genome- Genet 1995;56:400–407.
wide scan for childhood obesity-associated traits in 124 Shrimpton AE, Braddock BR, Thomson LL, Stein CK,
French families shows significant linkage on chro- Hoo JJ: Molecular delineation of deletions on 2q37.3
mosome 6q22.31-q23.2. Diabetes 2004;53:803–811. in three cases with an Albright hereditary osteodys-
114 Feitosa MF, Borecki IB, Rich SS, Arnett DK, trophy-like phenotype. Clin Genet 2004; 66:537–544.
Sholinsky P, Myers RH, Leppert M, Province MA: 125 Polityko A, Maltseva O, Rumyantseva N, Khurs O,
Quantitative-trait loci influencing body-mass index Seidel J, Claussen U, Weise A, Liehr T, Starke H:
reside on chromosomes 7 and 13: the National Two further AHO-like syndrome patients with
Heart, Lung, and Blood Institute Family Heart deletion of glypican 1 gene region in 2q37.2-q37.3.
Study. Am J Hum Genet 2002;70:72–82. Int J Mol Med 2004;14:977–979.
115 Varela MC, Simoes-Sato AY, Kim CA, Bertola DR, 126 Cormier-Daire V, Molinari F, Rio M, Raoul O, De
de Castro CI, Koiffmann CP: A new case of intersti- Blois MC, Romana S, Vekemans M, Munnich A,
tial 6q16.2 deletion in a patient with Prader-Willi- Colleaux L: Cryptic terminal deletion of chromo-
like phenotype and investigation of SIM1 gene some 9q34: a novel cause of syndromic obesity in
deletion in 87 patients with syndromic obesity. Eur J childhood? J Med Genet 2003;40:300–303.
Med Genet 2006;49:298–305. 127 Neas KR, Smith JM, Chia N, Huseyin S, St Heaps L,
116 Holder JLJ, Butte NF, Zinn AR: Profound obesity Peters G, Sholler G, Tzioumi D, Sillence DO, Mowat
associated with a balanced translocation that dis- D: Three patients with terminal deletions within the
rupts the SIM1 gene. Hum Mol Genet 2000;9: subtelomeric region of chromosome 9q. Am J Med
101–108. Genet A 2005;132:425–430.
117 Michaud JL, Boucher F, Melnyk A, Gauthier F, 128 Harada N, Visser R, Dawson A, Fukamachi M,
Goshu E, Levy E, Mitchell GA, Himms-Hagen J, Fan Iwakoshi M, Okamoto N, Kishino T, Niikawa N,
CM: Sim1 haploinsufficiency causes hyperphagia, Matsumoto N: A 1-Mb critical region in six patients
obesity and reduction of the paraventricular nucleus with 9q34.3 terminal deletion syndrome. J Hum
of the hypothalamus. Hum Mol Genet 2001;10: Genet 2004;49:440–444.
1465–1473. 129 Cotter PD, Kaffe S, McCurdy LD, Jhaveri M,
118 Holder Jr JL, Zhang L, Kublaoui BM, DiLeone RJ, Willner JP, Hirschhorn K: Paternal uniparental dis-
Oz OK, Bair CH, Lee YH, Zinn AR: Sim1 gene omy for chromosome 14: a case report and review.
dosage modulates the homeostatic feeding response Am J Med Genet 1997;70:74–79.
to increased dietary fat in mice. Am J Physiol 130 Kurosawa K, Sasaki H, Sato Y, Yamanaka M,
Endocrinol Metab 2004;287:E105–E113. Shimizu M, Ito Y, Okuyama T, Matsuo M, Imaizumi
119 Kublaoui BM, Holder JL Jr, Tolson KP, Gemelli T, K, Kuroki Y, Nishimura G: Paternal UPD14 is
Zinn AR: SIM1 overexpression partially rescues responsible for a distinctive malformation complex.
agouti yellow and diet-induced obesity by normaliz- Am J Med Genet 2002;110:268–272.
ing food intake. Endocrinol 2006;147:4542–4549. 131 Mitter D, Buiting K, von Eggeling F, Kuechler A,
120 Kublaoui BM, Holder JL Jr, Gemelli T, Zinn AR: Sim1 Liehr T, Mau-Holzmann UA, Prott EC, Wieczorek
haploinsufficiency impairs melanocortin-mediated D, Gillessen-Kaesbach G: Is there a higher incidence
anorexia and activation of paraventricular nucleus of maternal uniparental disomy 14 [upd(14) mat]?
neurons. Mol Endocrinol 2006;20: 2483–2492. Detection of 10 new patients by methylation-spe-
121 D’Angelo CS, Da Paz JA, Kim CA, Bertola DR, cific PCR. Am J Med Genet A 2006;140: 2039–2049.
Castro CI, Varela MC, Koiffmann CP: Prader-Willi- 132 Fryns JP, Haspeslagh M, Dereymaeker AM, Volcke P,
like phenotype: investigation of 1p36 deletion in 41 van den BH: A peculiar subphenotype in the fra(X)
patients with delayed psychomotor development, syndrome: extreme obesity-short stature-stubby
hypotonia, obesity and/or hyperphagia, learning hands and feet-diffuse hyperpigmentation. Further
disabilities and behavioral problems. Eur J Med evidence of disturbed hypothalamic function in the
Genet 2006;49:451–460. fra(X) syndrome? Clin Genet 1987;32: 388–392.
122 Phelan MC, Rogers RC, Clarkson KB, Bowyer FP, 133 de Vries BB, Fryns JP, Butler MG, Canziani F,
Levine MA, Estabrooks LL, Severson MC, Dobyns Wesby-van Swaay E, van Hemel JO, Oostra BA,
WB: Albright hereditary osteodystrophy and del(2) Halley DJ, Niermeijer MF: Clinical and molecular
(q37.3) in four unrelated individuals. Am J Med studies in fragile X patients with a Prader-Willi-like
Genet 1995;58:1–7. phenotype. J Med Genet 1993;30:761–766.
Anthony P. Goldstone
Senior Clinician Scientist and Consultant Endocrinologist
MRC Clinical Sciences Centre, Imperial College London
Hammersmith Hospital Campus
Du Cane Road, London W12 0NN (UK)
Tel. 44 20 8383 1029, Fax 44 20 8743 5409, E-Mail tony.goldstone@imperial.ac.uk
60 Goldstone · Beales
Korbonits M (ed): Obesity and Metabolism.
Front Horm Res. Basel, Karger, 2008, vol 36, pp 61–72
Abstract
Evolutionary and developmental perspectives add considerably to our understanding of the aetiology of
obesity and its related disorders. One pathway to obesity represents the maladaptive consequences of an
evolutionarily preserved mechanism by which the developing mammal monitors nutritional cues from its
mother and adjusts its developmental trajectory accordingly. Prediction of a nutritionally sparse environ-
ment leads to a phenotype that promotes metabolic parsimony by favouring fat deposition, insulin resis-
tance, sarcopenia and low energy expenditure. But this adaptive mechanism evolved to accommodate
gradual changes in nutritional environment; rapid transition to a situation of high energy density results in
a mismatch between predicted and actual environments and increased susceptibility to metabolic disease.
This pathway may also explain why breast and bottle feeding confer different risks of obesity. We discuss
how early environmental signals act through epigenetic mechanisms to alter metabolic partitioning, glu-
cocorticoid action and neuroendocrine control of appetite. A second pathway involves alterations in fetal
insulin levels, as seen in gestational diabetes, leading to increased prenatal fat mass which is subsequently
amplified by postnatal factors. Both classes of pathway may coexist in an individual. This developmental
approach to obesity suggests that potential interventions will vary according to the target population.
Copyright © 2008 S. Karger AG, Basel
A large body of research has focused on the environmental, physiological and bio-
chemical mechanisms that underlie the susceptibility of humans to obesity [1]. At the
most general level, obesity results from an imbalanced energy budget, where energy
intake chronically exceeds energy expenditure. This may result from poor nutritional
environment (imbalanced diet), weak or deranged satiety signals (overeating), effi-
cient energy storage or reduced energy output (disinclination to exercise or reduced
adaptive thermogenesis [2]).
Over recent years, there has been a growing trend towards considering obesity and
related medical problems in the context of human evolutionary ecology. In an early
application of this approach – the ‘thrifty genotype hypothesis’ – Neel [3] proposed
that regional differences in the propensity of human populations to store ingested
energy are a consequence of evolutionary adaptation to different ancestral environ-
ments. In this hypothesis, obesity results from a gene-environment interaction, such
that in hypocaloric ancestral environments humans accumulated ‘thrifty’ alleles of
genes associated with energy partitioning and in modern hypercaloric environments
these alleles cause excess energy storage. Although individual variation in obesity
does have a strong genetic component [4], several lines of evidence now suggest that
the thrifty genotype hypothesis cannot account for the temporal and spatial variation
in the incidence of human obesity [5, 6].
Current information thus suggests that there is a substantial developmental compo-
nent to obesity, such that humans are genetically predisposed to develop susceptible phe-
notypes in response to some developmental circumstances but not others. Those with a
developmentally determined susceptible phenotype will develop obesity if exposed to
hypercaloric environments, whereas those with a non-susceptible phenotype will not
become obese in the same environments. Such explanations do not diminish the
explanatory role for evolution, but refine it to consider also those aspects of human
development that favour energy accumulation and disfavour energy expenditure.
Two major developmental pathways can lead to obesity. First is an adaptive path-
way by which signals in early life act through the processes of developmental plastic-
ity to increase the adipogenic potential of exposure of the organism to a high
postnatal nutritional load. We term this the ‘mismatch pathway’, and it may have mul-
tiple forms. Second is a pathway that is likely to be pathogenic in its origin, namely
the effects of fetal hyperinsulinaemia leading to increased fetal fat mass which is then
amplified after birth. This is most typically seen in infants of diabetic mothers. It is
less clear whether maternal adiposity without diabetes induces a similar biology.
These pathways may coexist in individuals and across generations.
In this chapter, we primarily focus on the mismatch pathway and explore the concept
that obesity and metabolic syndrome are the maladaptive consequences of an evolution-
arily preserved mechanism by which environmental influences early in development set
the life course of the organism. We will focus on how such influences can programme
contributors to obesity such as altered energy partitioning and altered neuroendocrine
control of appetite. We will consider environmental signals arising from the fetal envi-
ronment (both undernutrition and overnutrition) and from the neonatal environment.
Although humans are the fattest mammals at birth, averaging about 15% fat [7], there
is substantial variability that is related to intrauterine nutrition. Importantly, the
Infants of diabetic mothers are born relatively obese as a result of maternal hypergly-
caemia, which causes fetal hyperglycaemia and in turn fetal hyperinsulinaemia which
then promotes fat deposition in the fetus and macrosomia at birth. This overweight
tracks through into childhood and is exacerbated by maternal obesity [42]. The
resulting metabolic derangements persist into adulthood and are transmitted to the
next generation via disturbance of the intrauterine environment. There is experimen-
tal and epidemiological evidence for such transgenerational effects [43–45].
The mismatch pathway may well lead to offspring who, because of the inherent
insulin resistance of pregnancy, are predisposed to develop gestational diabetes, and
in turn their own offspring may have obesity because of the hyperinsulinaemia path-
way [46]. The two pathways can coexist in such circumstances, and this appears to be
frequent in the Indian subcontinent where small maternal size drives maternal con-
straint but with better gestational nutrition and rising rates of gestational diabetes.
The consequence is that maternal diabetes is associated with functional fetal macro-
somia at birth weights considered normal in Caucasian populations [47, 48].
Epidemiological evidence suggests that both excessive weight gain early in life in
infants of normal birth weight [49–51] and accelerated weight gain (so-called ‘catch-
up’ or ‘compensatory’ growth) in infants of low birth weight [52] may predispose to
later obesity. Such accelerated weight gain represents an increment in central adipos-
ity rather than musculoskeletal growth [53]; the rise in central adiposity continues in
early childhood even after normalisation of weight-for-age in late infancy and is asso-
ciated with the development of insulin resistance [21]. The apparent protective effect
of breastfeeding against obesity in later life [54] may reflect the lower energy density
of human breast milk compared with formulations based on cows’ milk.
These phenomena may represent additional forms of the mismatch pathway – in
which the postnatal environment is inappropriate for the environment determined by
the level of maternal constraint and where the level of nutritional exposure is in
excess of that predicted in utero.
Mechanistic Considerations
Basic Mechanisms
There is considerable evidence that the mechanisms of developmental plasticity are
underpinned by developmental epigenetic processes, as reviewed in depth elsewhere [55,
56]. These are the processes by which access of transcription factors to DNA is altered by
Metabolic Partitioning
The correlation between an adverse fetal environment and the subsequent develop-
ment of insulin resistance is well established [60], and hyperinsulinaemia subsequent
to insulin resistance will promote fat deposition. In turn, the marked susceptibility of
central adipose tissue to lipolysis will result in release of free fatty acids, which
increases insulin resistance [61]. A parallel study in low birth weight humans and
growth-restricted rats showed reduced muscle expression of specific insulin-sig-
nalling proteins, including protein kinase C- and the glucose transporter GLUT4, in
affected individuals [62], providing a potential mechanism for the insulin resistance
seen in this setting. Dulloo [53] has proposed that preferential deposition of ‘catch-up
fat’ is an evolved mechanism for rebuilding adipose stores after nutritional restriction
and that such redistribution of glucose from skeletal muscle to adipose tissue is medi-
ated by suppression of muscle thermogenesis. Experimentally, prevention of early
catch-up growth reversed the development of glucose intolerance and obesity in a
mouse model of diabetes caused by maternal undernutrition [63].
Glucocorticoids
Resetting of the hypothalamic-pituitary-adrenal axis has been proposed to underlie
many of the phenomena caused by an adverse fetal environment [64], and indeed alter-
ations in basal and stimulated cortisol secretion can be correlated with birth weight in
humans [65]. Although adipose tissue is responsive to glucocorticoids and excessive
glucocorticoid exposure causes central obesity, dyslipidaemia and insulin resistance,
circulating glucocorticoid levels are often not significantly raised in obesity [66].
However, tissue exposure to glucocorticoids is controlled not only by circulating hor-
mone levels but also by tissue levels of the glucocorticoid receptor and tissue activity of
the two enzymes that interconvert active and inactive glucocorticoids, namely 11-
hydroxysteroid dehydrogenase (11HSD) types 1 and 2, respectively. Visceral adipose
tissue has increased levels of 11HSD1, suggesting that local glucocorticoid action may
be amplified in this tissue, and experimental manipulation of levels of 11HSD1 affects
propensity to diet-induced obesity [66]. These effects are mimicked by nutritional
restriction during gestation, which increases levels of the glucocorticoid receptor and of
11HSD1 and decreases levels of 11HSD2 in the offspring [67], enhancing adipose
tissue sensitivity to glucocorticoids in association with greater adiposity.
An Evolutionary Coda
Environmental change, resulting from natural events or migration, was a key driver
of human evolution, and there is paleoclimatic evidence that such change would have
been felt across several generations rather than over a short time-frame [77]. In such
an environment, ‘adaptive versatility’ [77] or ‘forecasting’ [78] using well-established
mechanisms of developmental plasticity would have been valuable in allowing more
rapid adaptation to changing nutritional conditions than possible by genetic means
alone, and integration of the cue across several generations [31] would tend to
increase the fidelity of the prediction. An adaptive mechanism to counter a predicted
sparse environment would be likely to have effects not only on metabolism, for exam-
ple a tendency to deposit fat and reduce energy expenditure, but also on other aspects
of the individual’s life history, such as reproductive strategy and longevity, and such
effects can be observed [30]. But if the prediction is faulty and the individual cannot
adjust, as would occur if there is a major shift in the environment between early child-
hood and adulthood, then the mismatch between phenotype and environment will
lead to disease. Such effects are seen most clearly when individuals move rapidly
between environments [14, 79] or when their society undergoes rapid nutritional
transition [80]. If the later environment does match the prediction, even at a low
plane of adult nutrition [81], then the risk of disease is less. But a continued high
References
1 Björntorp P: International Textbook of Obesity. 9 Wells JCK: The evolution of human fatness and sus-
Chichester, John Wiley & Sons Inc., 2001. ceptibility to obesity: an ethological approach. Biol
2 van Baak M: Adaptive thermogenesis during Rev 2006;81:183–205.
over- and underfeeding in man. Br J Nutr 2004;91: 10 Benyshek DC, Watson JT: Exploring the thrifty
329–330. genotype’s food-shortage assumptions: a cross-
3 Neel JV: Diabetes mellitus: a ‘thrifty’ genotype ren- cultural comparison of ethnographic accounts of
dered detrimental by ‘progress’? Am J Hum Genet food security among foraging and agricultural soci-
1962;14:353–362. eties. Am J Phys Anthropol 2006;131:120–126.
4 Mutch DM, Clement K: Genetics of human obesity. 11 Prentice AM, Whitehead RG, Roberts SB, Paul AA:
Best Pract Res Clin Endocrinol Metab 2006;20: Long-term energy balance in childbearing African
647–664. women. Am J Clin Nutr 1981;34:2790–2799.
5 Gluckman PD, Hanson MA: The developmental ori- 12 Godfrey K: The ‘developmental origins’ hypothesis:
gins of the metabolic syndrome. Trends Endocrinol epidemiology; in Gluckman PD, Hanson MA (eds):
Metab 2004;15:183–187. Developmental Origins of Health and Disease.
6 Speakman JR: Thrifty genes for obesity and the Cambridge, Cambridge University Press, 2006, pp
metabolic syndrome – time to call off the search? 6–32.
Diabet Vasc Dis Res 2006;3:7–11. 13 Kensara O, Wootton S, Phillips D, Patel M, Jackson
7 Kuzawa CW: Adipose tissue in human infancy and A, Elia M, the Hertfordshire Study Group: Fetal pro-
childhood: an evolutionary perspective. Yearbook gramming of body composition: relation between
Physical Anthropol 1998;41:177–209. birth weight and body composition measured with
8 Harrington TAM, Thomas EL, Frost G, Modi N, Bell dual-energy X-ray absorptiometry and anthropo-
JD: Distribution of adipose tissue in the newborn. metric methods in older Englishmen. Am J Clin Nutr
Pediatr Res 2004;55:437–441. 2005;82:980–987.
Abstract
Obesity and its sequelae may prove to be the greatest threat to human lifestyle and health in the devel-
oped world this century. The so called obesity epidemic has seen the incidence of obesity and over-
weight almost double in Western societies and the trend is mirrored in nations that are transitioning to
first world economies. There is no doubt that much of the rise in obesity can be attributed to lifestyle fac-
tors such as the excess consumption of energy-dense foods and the decline in physical activity.
However, the ‘fetal origins’ hypothesis, first proposed by Barker and colleagues and elaborated by several
groups over the past 15 years to be termed the ‘Developmental Origins of Adult Health and Disease’
(DOHaD), provides an alternative explanation for the rising rates of obesity. The DOHaD hypothesis states
that exposure to an unfavourable environment during development (either in utero or in the early post-
natal period) programmes changes in fetal or neonatal development such that the individual is then at
greater risk of developing adulthood disease. This chapter discusses the effects of maternal obesity on
fetal development and birth outcomes as well as the manner in which DOHaD may contribute to the
obesity epidemic. Copyright © 2008 S. Karger AG, Basel
Health care systems around the globe are beginning to recognise the risk that obesity
poses to human health and many programmes are now being put into place in an
effort to reduce the burden of obesity and its related diseases. Current definitions of
obesity are based on the ratio of bodyweight (in kg) and height squared (in m2) and
expressed as body mass index (BMI) with a normal BMI defined as 20–24.9, moder-
ate overweight between 25–29.9 and obesity as ⱖ30. In 2000, the World Health
Organisation released the following statement: ‘Obesity is a chronic disease, prevalent
in both developed and developing countries, and affecting children as well as adults.
Indeed it is now so common that it is replacing the more traditional public health
concerns, including under-nutrition and infectious disease as one of the most signifi-
cant contributors to ill health’ [1]. At the turn of the millennium and the time of pub-
lication of the WHO report, the incidence of obesity in the United States was 30.5%
(compared with 22.9% in 1994) and 64.5% of the population were overweight (com-
pared with 55.9% in 1994) [2]. More recent statistics suggest that the incidence of
obesity and overweight is rising, not falling, in spite of the apparent efforts of govern-
ments and health care agencies. This shift in body mass has occurred over the past
one to two generations and as such it is unlikely that genetic drift is the cause of the
current obesity epidemic. Rather, a change in lifestyle, compounded by epigenetic or
developmental programming of an obese phenotype are the likely causative factors.
Obesity statistics from the United States are most often quoted, perhaps because
they give the greatest impact; however, scientific studies conducted in other nations
emphasise the fact that obesity is a worldwide problem. A study of cause of death in
South Korea illustrates this fact. In 1938, cardiovascular disease accounted for
approximately 1% of deaths in South Korea whilst infectious diseases were the cause
of approximately 23% of deaths. By 1993, this trend had reversed; approximately 30%
of deaths were attributable to cardiovascular disease whereas only 3% of deaths were
caused by infection. Certainly such statistics are affected both by the vast improve-
ments in anti-microbial medication and sanitation in that 60-year period; however,
the fact remains that obesity-related illness is the next public health hurdle.
Obesity may not, in itself, be a great risk to human health. Indeed, there are some
individuals who are overweight or obese but do not show any other signs of disease or
ill health. However, for the vast majority, increased body fat is associated with a range
of other, more serious conditions. These include increased blood pressure, insulin
resistance and diabetes mellitus, atherogenic plasma lipid profiles, and increased lev-
els of vascular inflammatory markers. Collectively, this spectrum of conditions is
termed the ‘metabolic syndrome’ and clinical diagnosis is based on the presence of 3
or more of the above signs. Endothelial dysfunction and leptin resistance are also
likely to contribute to the metabolic syndrome [3].
The rise of obesity is certainly due to the increased availability of food, and the
preponderance of energy dense (high fat and simple carbohydrate) foods that are reg-
ularly consumed in developing and developed societies. Moreover, the industrial era
has produced all manner of labour saving devices that has ultimately seen a reduction
in the physical activity quotient over time [4]. However, despite the obvious impor-
tance of food intake and energy expenditure during adulthood, there is now evidence
that adult lifestyle may not be the only factor at play in determining obesity [5]. The
environment encountered during the in utero and early postnatal periods may also
act to ‘programme’ an individual to have a greater risk of developing obesity and the
metabolic syndrome.
The developing fetus and neonate orchestrates its growth and development to best
meet the environmental conditions encountered at any given period. Where environ-
mental challenges or stimuli span a period of organogensis or developmental plastic-
ity, the adaptations made may be permanent.
Epidemiological data from Norway provide the first of many reports that that
environment encountered in early life may affect later health outcomes with regard to
cardiovascular disease [6]. The concept of ‘programming’ was introduced by Lucas
[7] and provides a conceptual framework for the observations made by Barker and
Osmond [8] with regard to an individual’s birth weight and the later risk of disease in
United Kingdom cohort studies. These early Barker studies focused on the relation-
ship between the weight at birth, and subsequently fetal nutrition, with death from
coronary heart disease [9, 10].
The direct relationship between maternal nutrition and later offspring obesity
was revealed in a study of conditions encountered by a discrete population during
the Dutch Hunger Winter of 1944–1945. During the World War II military opera-
tions by Allied forces to liberate The Netherlands, the occupying Nazi forces block-
aded areas of Holland over the winter of 1944–1945 and official rations were cut to
300–500 kcal per day. Later study of adults (at 50 years of age) who were in utero
during this defined period of famine indicate that exposure to famine during the
first half of pregnancy predisposed individuals to the development of obesity [11]
and coronary heart disease [12] and that famine exposure later in gestation resulted
in glucose intolerance and insulin resistance [13]. This direct evidence for the role of
maternal nutrition in the programming of adulthood obesity and disease in the off-
spring was followed by studies from UK cohorts with several studies showing an
inverse association between birth weight and BMI in adulthood [14] as well as
insulin resistance. BMI is used as an indicator of obesity because of the simplicity by
which it can be measured in large trials or retrieved retrospectively from records.
However, use of the BMI parameter as an indicator of obesity has been challenged,
and there are suggestions that it is a measure of heaviness rather than obesity per se.
Consistent with this, studies of monozygotic twin pairs [15] found that lower birth
weight was associated with increased waist–hip ratio, skin fold thickness and
reduced muscle mass, but not necessarily BMI. Nonetheless, body fat measurement
by dual energy X-ray absorptiometry in adult men born of low birth weight shows
these individuals to have a 5% increase in fat mass compared with subjects born of
normal birth weight [16].
These studies formed the basis for various experimental animal models of mater-
nal undernutrition in which to study the Developmental Origins of Adult Health and
Disease (DOHaD) hypothesis and these models support the hypothesis that fetal or
neonatal undernutrition results in aberrant development of the endocrine pancreas,
liver, kidney and cardiovascular systems, such that the offspring born from protein or
Given the overall rates of obesity within the general population of many societies, it is
not surprising that rates of maternal obesity are rising. A retrospective analysis of
287,213 pregnancies in the United Kingdom reports that 27.5% of pregnant women
in the cohort were moderately obese (BMI 25–29.5) and a further 10.9% were very
obese (BMI ⱖ30) [28]. This trend is also seen in the US population, where recent
studies estimate that between 18 and 35% of pregnant women are obese [29].
Maternal obesity is associated with a range of adverse effects that directly affect
maternal health and pregnancy outcome. These include, pre-eclampsia, post-partum
haemorrhage, hypertensive disorders and complications of delivery [30]. The finan-
cial cost of obesity is also an important consideration. As health systems have finite,
and often, constrained budgets with which to provide care, any increase in the burden
of cost will further limit the abilities of health care systems to cope with the obesity
epidemic. In the year 2000, a French study estimated that complications arising from
obesity in pregnancy (pre-gravid BMI ⬎30) resulted in a 5.4- to 16.2-fold increase in
The factors present in the maternal or fetoplacental milieu that may instigate alter-
ations in organogenesis or induce morphometric changes in the fetus and thus result
The previous sections of this chapter have focussed on the effect of maternal obesity
and offspring exposure to obesity in the in utero period, and the following chapter
considers in detail the determinants of childhood obesity. Nonetheless, when consid-
ering the developmental origins of obesity, it is important to consider both in utero
and early neonatal environments as the hypothalamic appetite centres ultimately
responsible for the maintenance and determination of body weight set-points mature
in early life [5, 42]. This section will consider the role of maternal obesity and fat
intake in both the in utero and postnatal suckling periods on the developmental pro-
gramming of obesity.
The current dietary intakes in many Western populations are high in saturated
fatty acids, and clearly the intake of energy-dense and potentially pro-inflammatory
fats (those that are prone to becoming oxidised low-density lipoprotein cholesterols)
impacts on adult obesity. However, fat intake and obesity may also be of great impor-
tance in the developmental programming of disease. Despite the obvious reflection
on the food intake in Western societies, there is still a paucity of experimental animal
studies considering the role of maternal overnutrition in the developmental program-
ming of adult disease.
Early studies in baboons showed that overnutrition in the suckling period resulted in
permanent increases in plasma cholesterol concentrations [43] and adipocyte size [44].
Rats exposed to a lard-rich diet during gestation and suckling (via maternal diet),
then weaned onto a control diet, are obese, hypertensive, insulin resistant, dyslipi-
daemic, and demonstrate blunted endothelium-dependent vasodilatation [45–47].
Interestingly, the type of fatty acid predominating in the maternal diet impacts on the
programming of offspring disease. High-fat diets that are also rich in –3 polyunsat-
urated fatty acids have been shown to result in a largely normal offspring phenotype
despite the increased total fat and caloric load of the diet [48, 49]. Moreover, neonatal
rat pups suckled by dams fed fat-rich diets supplemented with –3 and –6 essential
fatty acids show reduced adipose tissue weight and reduced plasma leptin concentra-
tions [50]. Therefore, it appears that exposure to maternal diets that are rich in satu-
rated fatty acids is deleterious to later offspring health, whereas even relatively high
essential –3 and –6 fatty acid intake in the diet results in a normal offspring phe-
notype. There is also emerging evidence in humans that lowering total fat intake
whilst increasing the proportion of –3 and –6 essential fatty acid may have benefi-
cial effects on offspring and maternal health [51].
Conclusions
There are a myriad of socioeconomic, lifestyle and dietary behavioural factors that
have contributed and continue to add to the worldwide obesity epidemic.
Complications of maternal obesity or type 2 diabetes are summarised in table 1. The
developmental programming of adult disease appears to be one more factor that war-
rants serious consideration and one that should be tackled by public health initiatives.
Because of the alarming rate of obesity in young children across many societies, obe-
sity-related disease will not be eradicated in the near future; however, it is vital that
not only is the population informed of the risks and consequences of obesity to
maternal health, but also of the potential risks to the health of future generations.
References
1 WHO: Obesity: Preventing and managing the global 7 Lucas A: Programming by early nutrition in man.
epidemic. Technical Support Series 894. Geneva, World The childhood environment and adult disease. Ciba
Health Organisation, 2002, pp 1–4. Foundation Symposium 156. G. Bock and J. Whelan.
2 Ogden CL, Flegal KM, Carroll MD, Johnson CL: Chichester, John Wiley and Sons, 1991, pp 38–50.
Prevalence and trends in overweight among US chil- 8 Barker DJ, Osmond C: Childhood respiratory infec-
dren and adolescents, 1999–2000. JAMA 2002;288: tion and adult chronic bronchitis in England and
1728–1732. Wales. Br Med J (Clin Res Ed) 1986;293:1271–1275.
3 Armitage JA, Khan IY, Taylor PD, Nathanielsz PW, 9 Barker DJ, Gluckman PD, Godfrey KM, Harding JE,
Poston L: Developmental programming of metabolic Owens JA, Robinson JS: Fetal nutrition and car-
syndrome by maternal nutritional imbalance; how diovascular disease in adult life. Lancet 1993;341:
strong is the evidence from experimental models in 938–941.
animals. J Physiol 2004;561:355–377. 10 Barker DJ: Fetal origins of coronary heart disease.
4 Choi BC, Hunter DJ, Tsou W, Sainsbury P: Diseases of BMJ 1995;311:171–174.
comfort: primary cause of death in the 22nd century. 11 Ravelli GP, Stein ZA, Susser MW: Obesity in young
J Epidemiol Community Health 2005;59:1030–1034. men after famine exposure in utero and early infancy.
5 Levin B: The obesity epidemic: metabolic imprinting N Engl J Med 1976;295:349–353.
on genetically susceptible neural circuits. Obes Res 12 Roseboom TJ, van der Meulen JH, Osmond C,
2000;8:342–347. Barker DJ, Ravelli AC, Schroeder-Tanka JM, van
6 Forsdahl A: Are poor living conditions in childhood Montfrans GA, Michels RP, Bleker OP: Coronary
and adolescence an important risk factor for arte- heart disease after prenatal exposure to the Dutch
riosclerotic heart disease? Br J Prev Soc Med 1977;31: famine, 1944–45. Heart 2000;84:595–598.
91–95.
Childhood Obesity
M.A. Sabina,b ⭈ J.P.H. Shieldb
a
Royal Children’s Hospital and Murdoch Childrens Research Institute,
Melbourne, Australia; bUniversity of Bristol, Bristol, UK
Abstract
The prevalence of childhood obesity continues to increase worldwide. Its presence is associated with
significant adverse effects on health including an increased propensity to type II diabetes, cardiovascular,
respiratory, and liver disease. In the vast majority of children, obesity is lifestyle-related, yet there is a
dearth of evidence on how to best develop effective prevention and treatment strategies. This review
outlines the importance of childhood and adolescent growth on long-term health, the definitions used
to define obesity in children (along with up-to-date prevalence data), causes and consequences, and
aspects of prevention and management. Copyright © 2008 S. Karger AG, Basel
Most countries throughout the world, excepting certain areas of Sub-Saharan Africa have
witnessed a continued increase in the prevalence of obesity over the last two decades [1].
This carries major public health implications as there is little doubt that obese children
are at increased risk of developing long-term morbidity and eventual mortality sec-
ondary to increasing adiposity. Despite this, there remain very few effective prevention or
treatment strategies with which to halt this escalating epidemic. Understanding the fac-
tors that regulate adiposity in childhood, and the development of co-morbidities, would
allow us to focus prevention and treatment strategies to those most at risk.
Growth throughout childhood has profound implications for later health. There is now
considerable evidence indicating that size at birth, and growth in the first years of life,
affect the long-term risk of obesity and insulin resistance [2]. The risk associated with
birth weight has been defined as a ‘U-shaped curve’ with infants at both ends of the birth
weight spectrum being at increased risk. Infants born small for gestational age or with
lower ponderal indices [(weight in g/length in cm3) ⫻100], who subsequently show
rapid growth, are at greater risk of type 2 diabetes (T2DM) and cardiovascular disease in
later life, with evidence earlier in life of increased insulin resistance. Children born large
for gestational age are also at risk of subsequent obesity and T2DM, although the degree
to which maternal gestational diabetes accounts for these associations is still poorly char-
acterised. There is evidence that rapid growth in small for gestational age babies is specif-
ically associated with an increased propensity to abdominal obesity.
The transition from childhood to adolescence is fundamentally important in the
development of obesity, as puberty has a profound effect on regional body composi-
tion. During puberty, females accumulate a higher proportion of their total adult fat
mass, than their total adult lean tissue mass. The increased subcutaneous fat that is
seen in women develops during puberty, suggesting that oestrogen may preferentially
promote subcutaneous adipose deposition. During adolescence, boys characteristi-
cally maintain a relatively fixed absolute fat mass, whereas girls display an increase in
fat mass, accruing fat at ⬎1 kg per year [3]. The remodelling of fat mass and distribu-
tion appears to be central to the development of the physiological insulin resistance
that develops during pubertal growth [4].
Childhood adiposity can be measured in numerous ways although body mass index
(BMI) remains the most commonly used. Whilst BMI is a relatively simple tool with
which to assess body mass, it is a relatively poor predictor of actual body composition in
both adults and children. Due to its ease of determination however, along with a good cor-
relation with body fat, it has remained the accepted method to define obesity in children
based on current expert opinion [5]. Another important measure is waist circumference
which has been validated as a surrogate marker of visceral adiposity in children [6].
In adults, a BMI ⬎25 and ⬍30 corresponds to ‘overweight’, whereas a BMI ⬎30
identifies those with obesity. These cut-off points correspond to an increased risk of
cardiovascular disease and diabetes in adults. In children, however, BMI changes with
normal longitudinal growth, as shown in figure 1. Therefore, it is inappropriate to sim-
ply express raw BMI in children, without adjusting for age and sex. Instead, BMI stan-
dard deviation scores or z-scores (BMI SDS – representing increases or decreases
around the 50th centile for age and sex) are used to determine which children are rela-
tively ‘overweight’ or ‘obese’. For national statistics, BMI levels of ⬎95th, ⬎97th, or
⬎98th percentile have been used to identify the ‘fattest’ children within different popula-
tions. These limits have high specificity and moderate sensitivity, and allow temporal
changes within countries to be assessed. For across-country comparisons, international
cut-off points must be used and these have been based on the extrapolation of adult
cut-off points back into childhood [7]. The international cut-off points tend to greatly
underestimate obesity prevalence when used for determining prevalence rates within a
specific country [8]. However, BMI SDS may not be the best tool with which to
86 Sabin · Shield
Fig. 1. Boys BMI chart, demonstrating changes in BMI with age. Similar charts for girls are also avail-
able. Reproduced with permission from the Child Growth Foundation. Printed copies and further
information are available from www.healthforallchildren.co.uk.
Childhood Obesity 87
assess longitudinal changes in adiposity in children enrolled into weight management
programs, as the within-child variability over time depends upon the child’s levels of adi-
posity. Under these circumstances, age-adjusted BMI (calculated by subtracting the sex-
and age-specific median BMI) may be a better tool [9]. However, the normative data nec-
essary to make these calculations are currently unavailable so most workers continue to
use BMI SDS when reporting their results.
There has been a dramatic rise in the number of children who fulfil the criteria neces-
sary for the diagnosis of obesity [10]. Data from the CDC in the USA
(http://www.cdc.gov) demonstrate an increase in the prevalence of children aged 6–19
years old who were considered to be overweight (ⱖ95th percentile) from 4–5% in
1963–1970 to 15% in 1999–2000. Using similar criteria in the majority of cases, the
International Obesity Task Force have inspected the prevalence of obesity in children
aged around 10 years old from data derived from 21 European countries between 1992
and 2001 and found levels to vary between 10 and 36% [11]. Data from the Health
Survey for England in 2002 indicated that 8.5% of all 6-year-olds and 15% of all 15-
year-olds satisfied the criteria for obesity, and similar data from 2003 found that the
prevalence of obesity in children aged 2–10 had increased from 9.9 to 13.7% from 1995
(www.dh.gov.uk). In the non-Westernised world, there is also evidence that obesity in
general is increasing, especially amongst urban populations [12]. For example, China
(a country previously defined as one of the world’s leanest populations) has witnessed
a dramatic recent rise in childhood overweight and obesity prevalence [13].
Endocrine and single gene disorders causing obesity in childhood are rare, account-
ing for 1–2% of obese children seen in a tertiary care setting. Nevertheless, an under-
standing of these disorders is required to recognise rare but treatable causes of
childhood obesity. A thorough description of these conditions is beyond the scope of
this review and can be found elsewhere [14].
The majority of cases, however, arise from a simple interaction between host factors
that enhance susceptibility and environmental factors which increase food intake and
decrease energy expenditure [15]. Factors causing the imbalance in energy intake and
energy expenditure are numerous, simply reflecting the components of the obesogenic
environment in which we live. Factors important in excessive energy intake include the
consumption of energy-dense foods, increased portion sizes, between-meal snacking
and regular intake of sugar-sweetened beverages and fruit juices. Decreased energy
expenditure is often due to the coupling of increased sedentary activities, such as TV and
computer games, alongside decreased physical activity. There are also very significant
88 Sabin · Shield
parental [16] and socioeconomic [1, 17] contributions to obesity risk as demonstrated by
a recent study which showed that while 89% of parents of overweight 5- to 6-year-olds
were unaware that their child was overweight, 71% were not concerned, with less edu-
cated parents being less likely to take action [18]. Ethnicity also significantly impacts
upon obesity risk and the development of co-morbidities. For example, data from several
countries show that black children have a higher prevalence of obesity than white chil-
dren [17] while obese children from certain ethnic groups (e.g. South Asia) appear to
exhibit higher rates of complications like T2DM for a given level of obesity [19].
Addressing these complex demographic and lifestyle interactions remains central to the
development of effective prevention and treatment programs for childhood obesity.
Childhood obesity tracks into adult life with a high degree of predictability – 68% of
children with a BMI consistently 85th percentile grow up to be obese adults [20].
However, childhood obesity independently has its own short- and long-term adverse
effects on health.
Endocrine Complications
The major endocrine consequence of obesity is impaired glucose tolerance and subse-
quent T2DM – a process that develops in individuals with insulin resistance and relative
(rather than absolute) insulin deficiency. Significant numbers of children and adoles-
cents with obesity seen within specialised clinics have impaired glucose tolerance when
challenged with an oral glucose load. The figures vary across centres but have been put as
high as 25% in the USA, lower in the UK with levels around 10% and possibly even lower
in mainland Europe [21]. Predictions from the USA imply that obesity-associated
T2DM is likely to become the commonest form of newly diagnosed diabetes in adoles-
cent youth with evidence suggesting a global spread in the condition (although actual
incidence data are currently sparse [22]). Various centres in the USA have recorded dra-
matic increases in the number of children diagnosed with T2DM. For example, a 10-fold
increase was recorded from a centre in New York from 1990 to 2000 with 50% of all new
cases of diabetes having T2DM [23]. In Japan, researchers have documented a rise in the
annual incidence of T2DM in children from 1.73/100,000 to 2.76/100,000 over 20 years
[24], whilst evidence is also emerging of an increase in urban South-Asian children. Data
from Europe are scarce: a population based study in Austria established an incidence of
0.25/100,000 children [25], whilst a report from France indicated relatively low but
increasing numbers of children presenting with T2DM [26]. In the UK, Ehtisham et al.
[27] estimated a crude prevalence of T2DM in patients under 16 years of 0.21/100,000,
whilst a recent report reviewing first hospital admissions with a diagnosis of T2DM in
patients under 18 years indicated a significant rise between 1996–7 to 2003–4 [28].
Childhood Obesity 89
The emergence of T2DM in adolescence has important implications for both the
health of the individual and health service resources. Treatment compliance [29] and
psychological health [30] is often poor in childhood T2DM. Various studies have
implied an accelerated risk of nephropathy [31] and retinopathy [32] compared to
young people with type 1 diabetes, whilst recent data indicate early signs of cardio-
vascular disease in youth with T2DM [33]. The only currently available longitudinal
data make for worrying reading: of 79 children re-contacted up to 15 years after the
diagnosis of T2DM, 9% had died and 6% were on dialysis [34].
Respiratory Complications
Obesity has long been associated with sleep disorders, now termed ‘obstructive sleep
apnoea syndrome’ [44]. This causes decreased overnight oxygen saturation and
repetitive arousal prohibiting qualitative REM sleep, leading to daytime lethargy and
90 Sabin · Shield
somnolence. This not only affects physical activity levels (further compounding obe-
sity), but also school performance and concentration [45].
Interestingly, several studies have also now identified that reduced sleep duration
may be associated with an increased propensity to obesity, probably through changes
in orexegenic hormone levels, suggesting that encouraging longer sleep in childhood
may be a useful public health tool [46].
Gastrointestinal Complications
Obesity causes non-alcoholic fatty liver disease (NAFLD) in childhood [47]. NAFLD
refers to a spectrum of fatty infiltration of the liver, without alcohol consumption,
that encompasses simple fatty deposits seen in the liver, fatty liver with necro-inflam-
mation (termed non-alcoholic steatohepatitis) and ‘cryptogenic’ cirrhosis. Significant
numbers of obese children, particularly those with associated insulin resistance, have
been identified as suffering from this condition. It has a male preponderance and is
intimately linked to insulin resistance. The classical feature is of raised serum amino-
transferases in the absence of significant alcohol ingestion. Weight management
remains the mainstay of therapy for this condition [48], although alternative treat-
ments undergoing investigation in childhood include metformin and vitamin E.
Other Systems
Other complications of childhood obesity include slipped femoral capital epiphysis,
renal abnormalities in a condition now termed ‘obesity-related glomerulopathy’ and
idiopathic intracranial hypertension. The psychological effects of obesity should not
be forgotten with many obese children seen in the clinical setting, lacking self-esteem,
with reduced quality of life scores comparable to those of children diagnosed with
cancer. Interestingly, however, some workers have found that while obese children
may have higher levels of body dissatisfaction than their normal-weight peers, there
is little evidence to suggest that they have significantly reduced self-esteem or higher
levels of depression [49].
Obesity prevention is ‘the only feasible solution for developed and developing countries’
[1], although no prevention strategies have yet been developed which show long-term
efficacy. It is likely that numerous strategies have to be established ranging from popu-
lation-based initiatives to targeted interventions aimed at those most at risk. A recent
report by the British Medical Association stated that ‘prevention strategies will require a
coordinated effort between the medical community, health administrators, teachers,
parents, food producers and processors, retailers and caterers, advertisers and the
Childhood Obesity 91
media, recreation and sport planners, urban architects, city planners, politicians and
legislators’ (www.bma.org.uk/ap.nsf/content/childhoodobesity), highlighting the com-
plexity involved in developing effective prevention programs for childhood obesity.
A comprehensive overview of prevention trials for childhood obesity is outside the
scope of this review and the reader is directed to other sources on the subject [50].
In brief, different prevention programs have included school-based initiatives,
individual and family-based interventions, and antenatal/postnatal/preschool
approaches. None has proven to be particularly effective and it is likely that a combi-
nation of strategies (within a wider context of governmental policy change, changes
in healthcare and service provision to children, and less child-centred advertising) is
required to have a demonstrable effect.
It is clear that the development of well-designed studies to assess which interven-
tions are most effective is now critical if we are to overcome the mismatch between
the prevalence of childhood obesity and the knowledge base from which to inform
preventative activity [51].
Management Considerations
The most recent Cochrane review of obesity therapy in childhood concludes that
‘there is a limited amount of quality data on the components of programs to treat
childhood obesity that favour one program over another’ [52]. In general, it appears
that behaviour modification centred on parent education, by which parents take on
the primary responsibility for change in lifestyle, might be of some benefit [53–55].
This appears to be most likely to work in younger children rather than adolescents, as
shown in figure 2 [56]. Furthermore reducing sedentary behaviour and increasing
exercise might prove useful although the evidence is still very limited [57, 58].
Given the limited resources and ill-defined framework for obesity therapy in child-
hood, it is sensible to consider who is most likely to benefit if weight management is
successful. There is evidence that some racial groups such as Black, African-
Americans and Europeans and South Asians have both an increased prevalence of
obesity and risk for T2DM with some authors suggesting that these groups should be
targeted as a priority [59]. Another priority group should be those with a family his-
tory of T2DM as there is good evidence that obese children from these families are a
greater risk. Given these considerations, it is still important to establish when referral
for obesity management should be undertaken, given that other medical consequences
of profound obesity may not be entirely centred on the above high risk groups.
The most important consideration is the need for the family of the child concerned
to want to address the issue of obesity. Without an acceptance that changes in lifestyle
are required, it is highly unlikely that any progress can be made. It has been demon-
strated that many parents underestimate their child’s weight and the health care risks
associated with obesity. It is therefore important within any consultation in primary
92 Sabin · Shield
1.0
0.0
⫺0.5
⫺1.0
⫺1.5
2 4 6 8 10 12 14 16 18
Age at entry to clinic (years)
Fig. 2. Change in BMI SDS versus age at entry to a hospital-based multidisciplinary paediatric
weight management clinic, demonstrating that younger children achieve better results with this
type of intervention. Reproduced from reference [56].
care or a community setting that the issue of a child’s weight is addressed, but this
should never be at the expense of embarrassing either the child or family. If after con-
sultation, the family accepts that changes need to be made, then an assessment and
plan for action can be developed.
In older children, where simple modification of lifestyle tends to prove more diffi-
cult, there is now some evidence for the use of pharmacological treatments in resis-
tant individuals. Orlistat (a gastrointestinal lipase inhibitor) and sibutramine (a
selective serotonin reuptake inhibitor) are the 2 main anti-obesity medications used
in adults and evidence from the USA suggests that each may be associated with bene-
ficial changes in BMI within the context of appropriate lifestyle modification pro-
grammes. Both orlistat and sibutramine have been tested in randomised controlled
trials in adolescence with some apparent benefit [60, 61]. It is still unclear which ado-
lescents require pharmacological treatment, although it would seem that the best
approach is to stratify treatment by the severity of obesity/co-morbid conditions and
reserve drug therapy for those in whom behavioural therapy has failed [62]. Current
evidence suggests, however, that both medications should only be prescribed from
specialist adolescent weight management centres and as part of an overall lifestyle
modification programme [63]. Rimonabant, an endocannabinoid receptor antago-
nist, is a new anti-obesity medication that has recently been licensed for use in adults.
Although there have been no studies to date that have examined its use in adoles-
cents, data from adult studies do not suggest any significant side effects that would
necessarily preclude its use in this age group in the future [64].
Childhood Obesity 93
In those in whom lifestyle change and pharmacotherapy have failed, especially
where severe obesity is associated with co-morbid conditions in childhood, then
surgery remains an alternative treatment approach [65]. Surgical procedures have
involved either gastric restrictive or malabsorptive methods, although the laparo-
scopic placement of an adjustable gastric band has become the most widely adopted
procedure. Results of a large randomised controlled trial assessing this surgical
approach versus a modified lifestyle approach in adolescents with severe obesity is
keenly awaited from the Centre for Obesity Research and Education in Australia.
Conclusions
Childhood obesity now affects at least 10% of the world’s children. It is associated
with numerous health consequences that will have a significant impact on healthcare
provision over the next few decades. Despite this, we have very little evidence on the
most appropriate prevention and treatment strategies needed to tackle the problem.
A recent report assessing the cost effectiveness of interventions in childhood obesity
concluded that ‘the greatest health benefit is likely to be achieved by the ‘‘reduction of
TV advertising of high fat and/or high sugar foods and drinks to children’’, ‘‘laparo-
scopic adjustable gastric banding’’ and the ‘‘multi-faceted school-based programme
with an active physical education component’’’ [66].
References
1 Lobstein T, Baur L, Uauy R: Obesity in children and 7 Cole TJ, Bellizzi MC, Flegal KM, Dietz WH:
young people: a crisis in public health. Obes Rev Establishing a standard definition for child over-
2004;5(suppl 1):4–104. weight and obesity worldwide: international survey.
2 Yajnik CS: Early life origins of insulin resistance and BMJ 2000;320:1240–1243.
type 2 diabetes in India and other Asian countries. 8 Reilly JJ: Descriptive epidemiology and health conse-
J Nutr 2004;134:205–210. quences of childhood obesity. Best Pract Res Clin
3 Cowell CT, Briody J, Lloyd-Jones S, Smith C, Moore Endocrinol Metab 2005;19:327–341.
B, Howman-Giles R: Fat distribution in children and 9 Cole TJ, Faith MS, Pietrobelli A, Heo M: What is
adolescents–the influence of sex and hormones. the best measure of adiposity change in growing
Horm Res 1997;48(suppl 5):93–100. children: BMI, BMI %, BMI z-score or BMI centile?
4 Roemmich JN, Clark PA, Lusk M, Friel A, Weltman Eur J Clin Nutr 2005;59:419–425.
A, Epstein LH, Rogol AD: Pubertal alterations in 10 Hedley AA, Ogden CL, Johnson CL, Carroll MD,
growth and body composition. VI. Pubertal insulin Curtin LR, Flegal KM: Prevalence of overweight and
resistance: relation to adiposity, body fat distribution obesity among US children, adolescents, and adults,
and hormone release. Int J Obes Relat Metab Disord 1999–2002. JAMA 2004;291:2847–2850.
2002;26:701–709. 11 Lobstein T, Frelut ML: Prevalence of overweight
5 Dietz WH, Bellizzi MC: Introduction: the use of among children in Europe. Obes Rev 2003;4:
body mass index to assess obesity in children. Am J 195–200.
Clin Nutr 1999;70:S123–S125. 12 Yajnik CS: Obesity epidemic in India: intrauterine
6 Lee S, Bacha F, Gungor N, Arslanian SA: Waist cir- origins? Proc Nutr Soc 2004;63:387–396.
cumference is an independent predictor of insulin 13 Wu Y: Overweight and obesity in China. BMJ 2006;
resistance in black and white youths. J Pediatr 333:362–363.
2006;148:188–194.
94 Sabin · Shield
14 Farooqi IS: Genetic and hereditary aspects of child- 29 Kawahara R, Amemiya T, Yoshino M, Miyamae M,
hood obesity. Best Pract Res Clin Endocrinol Metab Sasamoto K, Omori Y: Dropout of young non-
2005;19:359–374. insulin-dependent diabetics from diabetic care.
15 Dietz W: Factors associated with childhood obesity. Diabetes Res Clin Pract 1994;24:181–185.
Nutrition 1991;7:290–291. 30 Katz LSS, Abraham M, Murphy KM, Jawad AF,
16 Unit JHS: Forecasting obesity to 2010. In: Department McKnight-Menci H, Berkowitz R: Neuropsychiatric
of Health, 2006, London. disorders at the presentation of type 2 diabetes melli-
17 Wardle J, Brodersen NH, Cole TJ, Jarvis MJ, Boniface tus in children. Pediatr Diabetes 2005;6:79–83.
DR: Development of adiposity in adolescence: five 31 Yokoyama H, Okudaira M, Otani T, Sato A, Miura J,
year longitudinal study of an ethnically and socioe- Takaike H, Yamada H, Muto K, Uchigata Y, Ohashi Y,
conomically diverse sample of young people in Iwamoto Y: Higher incidence of diabetic nephropathy
Britain. BMJ 2006;332:1130–1135. in type 2 than in type 1 diabetes in early-onset diabetes
18 Crawford D, Timperio A, Telford A, Salmon J: in Japan. Kidney Int 2000;58: 302–311.
Parental concerns about childhood obesity and the 32 Yoshida Y, Hagura R, Hara Y, Sugasawa G, Akanuma
strategies employed to prevent unhealthy weight gain Y: Risk factors for the development of diabetic
in children. Public Health Nutr 2006;9: 889–895. retinopathy in Japanese type 2 diabetic patients.
19 Ehtisham S, Barrett TG: The emergence of type 2 Diabetes Res Clin Pract 2001;51:195–203.
diabetes in childhood. Ann Clin Biochem 2004;41: 33 Gungor N, Thompson T, Sutton-Tyrrell K, Janosky J,
10–16. Arslanian S: Early signs of cardiovascular disease in
20 Freedman DS, Khan LK, Serdula MK, Dietz WH, youth with obesity and type 2 diabetes. Diabetes Care
Srinivasan SR, Berenson GS: Racial differences in the 2005;28:1219–1221.
tracking of childhood BMI to adulthood. Obes Res 34 Dean H, Flett B: Natural history of Type 2
2005;13:928–935. diabetes diagnsoed in childhood: long term follow up
21 Invitti C, Gilardini L, Pontiggia B, Morabito F, in young adult years (abstract). Diabetes 2002;
Mazzilli G, Viberti G: Period prevalence of abnormal 51:A24.
glucose tolerance and cardiovascular risk factors 35 Tounian P, Aggoun Y, Dubern B, Varille V, Guy-
among obese children attending an obesity centre in Grand B, Sidi D, Girardet JP, Bonnet D: Presence of
Italy. Nutr Metab Cardiovasc Dis 2006;16: 256–262. increased stiffness of the common carotid artery and
22 Pinhas-Hamiel O, Zeitler P: The global spread of endothelial dysfunction in severely obese children: a
type 2 diabetes mellitus in children and adolescents. J prospective study. Lancet 2001;358: 1400–1404.
Pediatr 2005;146:693–700. 36 Berenson GS, Srinivasan SR, Bao W, Newman WP,
23 Grinstein G, Muzumdar R, Aponte L, Vuguin P, 3rd, Tracy RE, Wattigney WA: Association between
Saenger P, DiMartino-Nardi J: Presentation and 5- multiple cardiovascular risk factors and athero-
year follow-up of type 2 diabetes mellitus in African- sclerosis in children and young adults. The Bogalusa
American and Caribbean-Hispanic adolescents. Heart Study. N Engl J Med 1998;338: 1650–1656.
Horm Res 2003;60:121–126. 37 Freedman DS, Dietz WH, Tang R, Mensah GA, Bond
24 Urakami T, Kubota S, Nitadori Y, Harada K, Owada MG, Urbina EM, Srinivasan S, Berenson GS: The rela-
M, Kitagawa T: Annual incidence and clinical char- tion of obesity throughout life to carotid intima-media
acteristics of type 2 diabetes in children as detected thickness in adulthood: the Bogalusa Heart Study. Int J
by urine glucose screening in the Tokyo metropoli- Obes Relat Metab Disord 2004;28: 159–166.
tan area. Diabetes Care 2005;28:1876–1881. 38 Golley RK, Magarey AM, Steinbeck KS, Baur LA,
25 Rami B, Schober E, Nachbauer E, Waldhor T: Type 2 Daniels LA: Comparison of metabolic syndrome
diabetes mellitus is rare but not absent in children prevalence using six different definitions in overweight
under 15 years of age in Austria. Eur J Pediatr pre-pubertal children enrolled in a weight mana-
2003;162:850–852. gement study. Int J Obes (Lond) 2006;30: 853–860.
26 Ortega-Rodriguez E, Levy-Marchal C, Tubiana N, 39 Cook S, Weitzman M, Auinger P, Nguyen M,
Czernichow P, Polak M: Emergence of type 2 dia- Dietz WH: Prevalence of a metabolic syndrome phe-
betes in an hospital based cohort of children with notype in adolescents: findings from the third
diabetes mellitus. Diabetes Metab 2001;27:574–578. National Health and Nutrition Examination Survey,
27 Ehtisham S, Hattersley AT, Dunger DB, Barrett TG: 1988–1994. Arch Pediatr Adolesc Med 2003;157:
First UK survey of paediatric type 2 diabetes and 821–827.
MODY. Arch Dis Child 2004;89:526–529. 40 Adult Treatment Panel I: Third report of the national
28 Dr Foster’s case notes: Obesity and type 2 diabetes in cholesterol education program (NCEP) expert panel
children, 1996–7 to 2003–4. BMJ 2005;331:1167. on detection, evaluation, and treatment of high
blood cholesterol in adults (Adult Treatment Panel
III) final report. Circulation 2002;106:3143–3421.
Childhood Obesity 95
41 Duncan GE, Li SM, Zhou XH: Prevalence and 55 Golan M, Kaufman V, Shahar DR: Childhood obesity
trends of a metabolic syndrome phenotype among treatment: targeting parents exclusively v. parents
U.S. adolescents, 1999–2000. Diabetes Care 2004;27: and children. Br J Nutr 2006;95:1008–1015.
2438–2443. 56 Sabin M, Ford A, Hunt L, Jamal R, Crowne E, Shield
42 Weiss R, Dziura J, Burgert TS, Tamborlane WV, J: Which factors are associated with a successful out-
Taksali SE, Yeckel CW, Allen K, Lopes M, Savoye M, come in a weight management programme for obese
Morrison J, Sherwin RS, Caprio S: Obesity and the children? J Eval Clin Prac 2007;13:364–368.
metabolic syndrome in children and adolescents. 57 Epstein LH, Valoski AM, Vara LS, McCurley J,
N Engl J Med 2004;350:2362–2374. Wisniewski L, Kalarchian MA, Klein KR, Shrager
43 Sabin MA, Ford AL, Holly JM, Hunt LP, Crowne EC, LR: Effects of decreasing sedentary behavior and
Shield JP: Characterisation of morbidity in a UK, increasing activity on weight change in obese chil-
hospital based, obesity clinic. Arch Dis Child dren. Health Psychol 1995;14:109–115.
2006;91:126–130. 58 Epstein LH, Goldfield GS: Physical activity in the
44 Dietz WH: Health consequences of obesity in youth: treatment of childhood overweight and obesity: cur-
childhood predictors of adult disease. Pediatrics rent evidence and research issues. Med Sci Sports
1998;101:518–525. Exerc 1999;31:S553–S559.
45 Wing YK, Hui SH, Pak WM, Ho CK, Cheung A, Li 59 Freedman DS, Khan LK, Serdula MK, Ogden CL,
AM, Fok TF: A controlled study of sleep related dis- Dietz WH: Racial and ethnic differences in secular
ordered breathing in obese children. Arch Dis Child trends for childhood BMI, weight, and height.
2003;88:1043–1047. Obesity (Silver Spring) 2006;14:301–308.
46 Chaput JP, Brunet M, Tremblay A: Relationship 60 Chanoine JP, Hampl S, Jensen C, Boldrin M,
between short sleeping hours and childhood over- Hauptman J: Effect of orlistat on weight and body
weight/obesity: results from the ‘Quebec en Forme’ composition in obese adolescents: a randomized
Project. Int J Obes (Lond) 2006;30:1080–1085. controlled trial. JAMA 2005;293:2873–2883.
47 Roberts EA: Non-alcoholic fatty liver disease (NAFLD) 61 Berkowitz RI, Fujioka K, Daniels SR, Hoppin AG,
in children. Front Biosci 2005;10: 2306–2318. Owen S, Perry AC, Sothern MS, Renz CL, Pirner
48 Marion AW, Baker AJ, Dhawan A: Fatty liver disease MA, Walch JK, Jasinsky O, Hewkin AC, Blakesley
in children. Arch Dis Child 2004;89:648–652. VA: Effects of sibutramine treatment in obese adoles-
49 Wardle J, Cooke L: The impact of obesity on psycho- cents: a randomized trial. Ann Intern Med 2006;145:
logical well-being. Best Pract Res Clin Endocrinol 81–90.
Metab 2005;19:421–440. 62 Dietz WH: What constitutes successful weight man-
50 Ells LJ, Campbell K, Lidstone J, Kelly S, Lang R, agement in adolescents? Ann Intern Med 2006;145:
Summerbell C: Prevention of childhood obesity. Best 145–146.
Pract Res Clin Endocrinol Metab 2005;19:441–454. 63 Joffe A: Pharmacotherapy for adolescent obesity: a
51 Campbell K, Waters E, O’Meara S, Kelly S, weighty issue. JAMA 2005;293:2932–2934.
Summerbell C: Interventions for preventing obesity 64 Padwal RS, Majumdar SR: Drug treatments for obe-
in children. Cochrane Database Syst Rev:CD001871, sity: orlistat, sibutramine, and rimonabant. Lancet
2002. 2007;369:71–77.
52 Summerbell CD, Ashton V, Campbell KJ, Edmunds 65 Barnett SJ, Stanley C, Hanlon M, Acton R, Saltzman
L, Kelly S, Waters E: Interventions for treating obesity DA, Ikramuddin S, Buchwald H: Long-term follow-
in children. Cochrane Database Syst Rev: CD001872, up and the role of surgery in adolescents with morbid
2003. obesity. Surg Obes Relat Dis 2005;1:394–398.
53 Golan M, Fainaru M, Weizman A: Role of behaviour 66 Haby MM, Vos T, Carter R, Moodie M, Markwick A,
modification in the treatment of childhood obesity Magnus A, Tay-Teo KS, Swinburn B: A new approach
with the parents as the exclusive agents of change. Int J to assessing the health benefit from obesity interven-
Obes Relat Metab Disord 1998;22:1217–1224. tions in children and adolescents: the assessing cost-
54 Golan M, Crow S: Parents are key players in the pre- effectiveness in obesity project. Int J Obes (Lond)
vention and treatment of weight-related problems. 2006;30:1463–1475.
Nutr Rev 2004;62:39–50.
96 Sabin · Shield
Korbonits M (ed): Obesity and Metabolism.
Front Horm Res. Basel, Karger, 2008, vol 36, pp 97–106
Abstract
Many older people in developed countries are overweight or obese. The prevalence is increasing as
more people reach old age already overweight. Obesity in old age is associated with increased morbid-
ity and a reduction in quality of life. The relative increase in mortality is less in older than young adults
and the body weight associated with maximal survival increases with advancing age. Although inten-
tional weight loss by overweight older people is probably safe and beneficial, caution should be exer-
cised in recommending weight loss to overweight older people on the basis of body weight alone.
Methods of achieving weight loss in older adults are the same as in younger adults. Weight loss diets
should be combined with an exercise program to preserve muscle mass, as dieting results in loss of mus-
cle as well as fat, and older people have reduced skeletal muscle mass in any case. Weight loss drugs
have not been extensively studied in older people, and there is the potential for drug side effects and
interactions. Weight loss surgery appears to be safe and effective, albeit slightly less so than in younger
adults, but little is known about the outcomes of such surgery in those over 65 years.
Copyright © 2008 S. Karger AG, Basel
Prevalence
In most developed countries a substantial minority, and in some countries the major-
ity, of older people are overweight according to standard body weight criteria.
According to recent large surveys [1, 2], approximately 71% of Americans 60 years or
older and 60% of those 65 years or older were overweight (body mass index, BMI
ⱖ25 kg/m2), while approximately 32% of those 60 years or older and 20% of those 65
years or older were obese (BMI ⱖ30). Similarly, 29% of 55- to 64-year-olds in
England were obese [3], while 43% of Australians over 65 were overweight and 25%
of Australians aged 65–74 years and 14.4% over 75 years were obese [4].
Not only are many older people overweight or obese, but the rates are increasing
rapidly, in parallel with the dramatic increase over recent years in rates in younger
adults. For example, the prevalence of obesity (BMI ⱖ30) among people in the USA
over 60 years increased from 20 to 32% between 1988–1994 and 1999–2000 [5] and
among those over 70 years from 11.4 to 15.5% between 1991 and 2000 [6]. There have
been similar increases in other countries [4].
An understanding of the causes and consequences of excess weight in older people
is aided by an understanding of the changes in appetite, food intake, energy expendi-
ture and body composition that occur with ageing.
On average, adults become less hungry and eat less as they get older, even if healthy
[7]. This physiological, age-related reduction in appetite and energy intake has been
termed ‘the anorexia of aging’ [8] and appears to have many causes [9]. Average daily
energy intake decreases by up to 30% between 20 and 80 years [7]. Most of the age-
related decrease in energy is probably a response to the decline in overall energy
expenditure that also occurs as people get older. Changes in body weight and body
composition reflect the balance of these two declines. As indicated below, body
weight tends to increase through early adult life into middle age, suggesting a more
rapid decline in energy expenditure than in food intake during this time. In contrast,
body weight tends to decrease in older people, suggesting a faster decline in food
intake than in energy expenditure in later life.
Body Weight
Population studies show that on average people in westernised countries gain weight
until they are about 50–60 years old and after that tend to lose weight [10]. Although
some of the decline in mean body weight after age 50–60 years detected in cross-sec-
tional studies is due to the premature death of obese people, the decline in body
weight among older people has also been demonstrated in longitudinal studies. For
example, in one 2-year prospective study, community-dwelling American men over
65 years lost on average 0.5% of their body weight per year and 13.1% of the group
had weight loss of 4% per annum or more [11]. As a result of this weight loss in older
people, and the premature death of obese people at younger ages, the prevalence of
overweight and obesity, as defined by standard BMI criteria (BMI ⬎25 and ⱖ30,
respectively) peaks around age 50–60 years. It then remains fairly stable until about
age 70–75 years, before decreasing.
A substantial minority of older people have quite marked weight changes over
time. In one study [12], 17% of home-dwelling people in the USA over 65 years lost
98 McPhee Chapman
5% or more of their initial body weight over 3 years, while 13% gained 5% or more.
Other studies provide similar numbers [11]. There is evidence for interactive effects
on health of body weight category and change in body weight, particularly of adverse
effects in already underweight people who lose weight and in already overweight peo-
ple who gain weight [13].
Body Composition
As indicated above, people do not usually gain weight in old age. From this, it follows
that the high and increasing proportion of overweight and obesity in older people is
mainly due to the high and increasing proportion of adults who reach old age already
overweight. The causes of obesity in the elderly are therefore largely those of obesity
in younger adults. These have been addressed elsewhere in this book.
Older people are often not weight stable and a substantial minority of older people
have quite marked weight changes over time [11]. There are therefore some people
who only become overweight or whose weight increases dramatically in later life.
This can be due to medical illnesses, which restrict mobility and hence alter the food
intake-energy expenditure balance in favour of the former. Most common among
these is musculoskeletal disease, such as osteoarthritis of the lower limbs or back,
while cardiac failure and depression are also important causes, as is increased food
intake due to corticosteroid use.
As in younger adults, obesity in older adults is associated with absolute and rela-
tive increases in both mortality and morbidity. Obesity-related co-morbidities are
summarised in table 1. Most of these conditions become more common and severe
with increasing age even in those who are not overweight. These effects of age are
exacerbated by excess weight.
Mortality
The relative increase in the risk of death associated with being obese compared to being
of normal weight is not as great in older as in young adults. An assessment of thirteen
observational, prospective studies in which non-hospitalised people over 65 years were
followed for at least 3 years [22] found an association between mortality and increased
BMI in only a few, and then only above a BMI of 27–28.5, with little or no increase in
mortality at any BMI for people over 75 years. Where an optimum BMI could be identi-
fied, it was usually in the range 27–30. Consistent with this, a combined analysis of the
American NHANES I–III (1974–2000) study results found no significant increase in
mortality with any degree of overweight in people over 70 years, and an increased death
rate only in the ‘morbidly’ obese (BMI ⱖ35) among those 60–69 years [23]. Nevertheless,
the relative risk of mortality is still increased at high BMIs until about the age of 75 years,
and because of the greater background death rate in older people the absolute increase in
death rate attributable to obesity is substantial; 25% of the excess deaths attributed to
obesity in that NHANES analysis occurred in people over 70 years of age [5].
The causes of increased mortality are essentially the same as those in younger adults:
diabetes, hypertension, sleep apnoea, cardiovascular disease and an increased risk with
obesity of developing certain cancers, including breast, uterus, colon and prostate.
Obesity in older people is associated with increased rates of cataracts, mechanical uri-
nary and bladder problems, sleep apnoea and other respiratory problems [reviewed in
10]. After the age of 65 years, over 60% of people have symptomatic osteoarthritis [24]
commonly affecting the hip and knee, and this is a major cause of disability. Excess
weight hastens the development of knee osteoarthritis [25], exacerbates the symptoms
of lower limb osteoarthritis, and makes surgical treatment more hazardous.
Functional capacity and mobility are significantly reduced in obese compared to
lean older adults [26]. Obesity at any age is associated with a reduced quality of life
[27], but this is particularly so in older adults, in whom it is associated with reduced
muscle mass and strength and hence with physical frailty [20]. Compared to their
non-obese counterparts, obese older people are less likely to be pain free, have greater
limitations of physical function and are more likely to be homebound. Obesity is pre-
dictive of a greater rate of future disability, declines in functional status [10], particu-
larly when associated with loss of skeletal mass [21], and an increased admission rate
to nursing homes. Among women age 65 years or more in the Nurses Health Study, a
weight gain of 20 lb (⬇9 kg) or more over 4 years was associated with a reduction in
reported physical functioning [13]. Increased fat mass appears to be the factor specif-
ically responsible for obesity-related disability [26].
A beneficial effect of obesity in older people is that it is associated with increased
bone density in both weight bearing and non-weight bearing bones. This combines with
cushioning of falls provided by the extra fat stores, particularly around the hips
(‘endogenous’ hip protectors) to reduce fracture rates in older people [28]. Most studies
show that when overweight adults, young or older, intentionally lose weight they also
lose bone [reviewed in 10, 29]. Substantial unintentional weight loss in older people is
associated with an increased risk of hip fracture [30]. This is probably due to a combina-
tion of reduced bone mass and the effects of whatever illness caused the weight loss, for
example by increasing the likelihood of falls. Although it is reasonable to assume that
there is at least some increase in fracture risk due to the weight loss-associated bone loss,
in overweight older people who intentionally lose weight this has not been established.
Treatment options available to achieve weight loss in older people are the same as
those in younger adults; changes in diet and exercise, medications and surgery [10].
There is limited information about the effectiveness and safety of weight loss treat-
ments, particularly medications and surgery, in older people.
Lifestyle Measures
There is little reported experience with weight loss drugs such as sibutramine, which
reduces appetite and food intake, and orlistat, which inhibits lipase activity and thus
fat absorption, in older people. Such drugs should be used with caution in older peo-
ple, because of limited efficacy data, the possibility of interactions with other (multi-
ple) medications, and potentially worse side effects. Orlistat appears to be as effective
in older as in young adults [10], with a weight loss over 1 year approximately 3 kg
more than with placebo, but can cause diarrhoea and other gastrointestinal side
effects such as faecal incontinence. Side effects of sibutramine can include insomnia,
constipation and increased blood pressure.
Only a few studies so far have examined the outcomes of weight loss surgery in
overweight people over 60 years. The results are encouraging. In three studies com-
paring the effects of gastric bypass surgery in 115 people over 60 years to those in
3,470 people less than 60 years [42–44], there was one peri-operative death in the
older group (0.86%) and a possible but not definite slight increase in operative com-
plication rate compared to the younger patients. The surgery resulted in a marked
mean weight loss of 39–43 kg at 1 year in the older subjects and a significant reduc-
tion in the number of obesity-related morbidities and number of medications
needed to treat these. Both reductions were slightly, but significantly less in older
than young adults. The mean age of the ‘older’ subjects in these studies was quite
young, however, at 63–65 years, and the results of such surgery in even older people and
also of other forms of obesity surgery in these age groups remain to be determined.
Conclusions
Many older people are overweight or obese and this is associated with increased mor-
bidity and mortality. Weight loss is achievable and is beneficial in these people, par-
ticularly when the excess weight is associated with functional impairments. Weight
loss should probably not be recommended to older people on the basis of their body
weight alone, rather than co-existent morbidity, as the relative disadvantages of excess
weight per se are less in older than young adults.
References
1 Li F, Fisher KJ, Harmer P: Prevalence of overweight 10 Villareal DT, Apovian CM, Kushner RF, Klein S:
and obesity in older U.S. adults: estimates from the Obesity in older adults: technical review and position
2003 Behavioral Risk Factor Surveillance System sur- statement of the American Society for Nutrition and
vey. J Am Geriatr Soc 2005;53:737–739. NAASO, The Obesity Society. Am J Clin Nutr
2 Ogden CL, et al: Prevalence of overweight and obe- 2005;82:923–934.
sity in the United States, 1999–2004. JAMA 2006;295: 11 Wallace JI, Schwartz RS, LaCroix AZ, Uhlmann RF,
1549–1555. Pearlman RA: Involuntary weight loss in older out-
3 Banks J, Marmot M, Oldfield Z, Smith JP: Disease patients: incidence and clinical significance. J Am
and disadvantage in the United States and in Geriatr Soc 1995;43:329–337.
England. JAMA 2006;295:2037–2045. 12 Newman AB, Arnold AM, Burke GL, O’Leary DH,
4 Cameron AJ, et al: Overweight and obesity in Manolio TA: Cardiovascular disease and mortality in
Australia: the 1999–2000 Australian Diabetes, older adults with small abdominal aortic aneurysms
Obesity and Lifestyle Study (AusDiab). Med J Aust detected by ultrasonography: the cardiovascular
2003;178:427–432. health study. Ann Intern Med 2001;134: 182–190.
5 Flegal KM, Carroll MD, Ogden CL, Johnson CL: 13 Fine JT, et al: A prospective study of weight change
Prevalence and trends in obesity among US adults, and health-related quality of life in women. JAMA
1999–2000. JAMA 2002;288:1723–1727. 1999;282:2136–2142.
6 Mokdad AH, et al: The continuing epidemics of obe- 14 Prentice AM, Jebb SA: Beyond body mass index.
sity and diabetes in the United States. JAMA 2001; Obes Rev 2001;2:141–147.
286:1195–1200. 15 Beaufrere B, Morio B: Fat and protein redistribution
7 Wurtman JJ, Lieberman H, Tsay R, Nader T, Chew B: with aging: metabolic considerations. Eur J Clin Nutr
Calorie and nutrient intakes of elderly and young 2000;54(suppl 3):S48–S53.
subjects measured under identical conditions. J 16 Cree MG, et al: Intramuscular and liver triglycerides
Gerontol 1988;43:B174–B180. are increased in the elderly. J Clin Endocrinol Metab
8 Morley JE: Anorexia of aging: physiologic and patho- 2004;89:3864–3871.
logic. Am J Clin Nutr 1997;66:760–773. 17 Melton LJ, 3rd, Khosla S, Riggs BL: Epidemiology
9 Chapman IM: Endocrinology of anorexia of ageing. of sarcopenia. Mayo Clin Proc 2000;75(suppl):
Best Pract Res Clin Endocrinol Metab 2004;18: S10–S12; discussion S12–S13.
437–452. 18 Janssen I, Baumgartner RN, Ross R, Rosenberg IH,
Roubenoff R: Skeletal muscle cutpoints associated
with elevated physical disability risk in older men
and women. Am J Epidemiol 2004;159:413–421.
Models of ‘Obesity’ in
Large Animals and Birds
Iain J. Clarke
Department of Physiology, Monash University, Melbourne, Australia
Abstract
Most laboratory-based research on obesity is carried out in rodents, but there are a number of other
interesting models in the animal kingdom that are instructive. This includes domesticated animal species
such as pigs and sheep, as well as wild, migrating and hibernating species. Larger animals allow particu-
lar experimental manipulations that are not possible in smaller animals and especially useful models
have been developed to address issues such as manipulation of fetal development. Although some of
the most well-studied models are ruminants, with metabolic control that differs from monogastrics, the
general principles of metabolic regulation still pertain. It is possible to obtain much more accurate
endocrine profiles in larger animals and this has provided important data in relation to leptin and ghrelin
physiology. Genetic models have been created in domesticated animals through selection and these
complement those of the laboratory rodent. This short review highlights particular areas of research in
domesticated and wild species that expand our knowledge of systems that are important for our under-
standing of obesity and metabolism. Copyright © 2008 S. Karger AG, Basel
Obesity is clearly a major issue for humans and there is a very substantial research
effort directed towards a better understanding of the causes and consequences of the
problem. Basic and clinical researchers seek ways to prevent or treat obesity, and most
laboratory-based research is carried out in rodents such as rats and mice, since these
species have obvious advantages. Nevertheless, a range of other species offer special
attributes that allow lines of investigation different to those achievable with small lab-
oratory rodents. In particular, domesticated animal species allow some experimental
procedures that are not possible in smaller animals. A good example of this is the
methodology for the sampling of hypothalamic secretions into the hypophysial portal
blood in a sentient state; this was established as a technique 25 years ago and is clearly
the method of choice if one wishes to measure the output of hypothalamic hormones.
Advantages in using livestock and poultry for studies of metabolic function and
body composition are as follows:
– Farmers have been selecting for various traits related to millennia which has
segregated genes in relation to breed. Genomic analysis is well advanced for a
number of breeds and gene chips are available for species such as cattle, pigs and
poultry.
– Most farm animals are relatively easy to keep and are low cost, without the require-
ment for special housing. Their ‘natural’ habitat and social structure is also some-
what different to that of laboratory rodents in cages.
– Larger animals allow serial sampling of blood, other body fluids and tissues. Serial
sampling of cerebrospinal fluid can provide an index of neuronal activity. Multiple
sampling of fat and muscle tissues allows longitudinal analysis of metabolism and
different compartments of fat can be studied by serial sampling.
– Domesticated animals are not nocturnal, as are rodents, so that important diur-
nal patterns of behaviour and physiology of the former are more similar to
humans. Environmental effects on food intake, energy expenditure and meta-
bolic function may be studied and the role of factors such as photoperiod, are
well described.
Disadvantages in using livestock and poultry are that the technology for genetic
manipulation is not as well developed as in small laboratory species such as mice.
Nevertheless, cloning is clearly possible in sheep (e.g. Dolly) and is now a relatively
straightforward procedure in cattle. The creation of transgenic animals is also some-
what more difficult, although this has been achieved; it is only a matter of time before
this becomes more straightforward. Mice are obviously the species of choice for gene
knock-out and knock-in studies. Genetic selection studies can be done in domesti-
cated animals, but the longer generation times, compared to those of smaller labora-
tory species, make this somewhat arduous. Nevertheless, important models of obesity
can be generated and should provide important insights into the genetics of obesity
and related physiology.
This chapter will discuss how non-rodent models of metabolic function, appetite
and energy expenditure contribute to our understanding of the mechanisms involved
in the development of obesity. Predisposing genetic and epigenetic factors will be
examined and endocrine factors will be discussed. Natural models of appetite drive
and energy expenditure will be explored, including those that are found in migrating
species and those in animals adapted to extreme environments.
In the broadest sense, the model of metabolic control that pertains to laboratory rats
and mice as well as humans also relates to larger animals and birds. There are, how-
ever, some differences that are mentioned below. In animals subjected to ‘natural’
environment, especially in non-domesticated species, seasonal patterns of energy
intake and expenditure are evident and can be quite profound. Some widely
108 Clarke
studied species, such as sheep and cattle, are ruminants and this has significant
implications in regard to metabolic control. Ruminants have 4 stomachs, which are
specialised for the digestion of cellulose, and the predominant source of energy from
nutrient is absorbed as volatile fatty acids. Accordingly, the livers of ruminants have
exceptional capacity for gluconeogenesis and low hexokinase activity, favouring glu-
cose rather than glucose-6-phosphate which is a substrate for complex carbohy-
drates [1]. This means that the animal is able to ‘buffer’ blood glucose levels very
effectively with minimal changes over extreme body conditions. Thus, in sheep, pro-
found alteration of body weight and adiposity alters insulin levels but does not
change plasma glucose levels and a 72-hour fast has minimal effect on plasma glu-
cose levels [2]. On the other hand, short-term fasting elevates plasma non-esterified
fatty acid levels, indicating mobilisation of fat reserves to provide substrate for
gluconeogenesis. The study of ruminants is valuable because it allows us to compare
and contrast control of energy balance in animals that have different metabolic
substrates.
In terms of appetite control by central mechanisms, there is also widespread con-
formity across the species, although some species differences are apparent.
Accordingly, the most important appetite regulators such as neuropeptide Y (NPY)
and agouti-related peptide (AgRP) are found in all species, including avians. In sheep,
as in small laboratory rodents, leptin reduces food intake when administered intrac-
erebroventricularly or peripherally [3, 4], even though these animals are ruminants
and do not experience short-term emptying of their gastrointestinal tract. The acute
response to leptin treatment in the ruminant, in terms of food intake, is similar to that
observed in rats [5] such that, in both, the anorectic effect is not sustained with con-
tinued treatment. In rats given continuous treatment, body weight is reduced and
remains low, even though food intake returns to normal with continued leptin treat-
ment; this seems most likely due to an effect of leptin on energy expenditure.
Intracerebroventricular infusion of leptin into sheep for 3 months did not cause pro-
found loss of body weight [6], even though other effects (such as on bone morphol-
ogy and function) were manifest. In birds (quail), leptin treatment caused a transient
reduction in food intake and body weight, but this lasted for only 2 days [7]. Thus,
sustained loss of body weight occurs in rats treated with leptin, but this does not
occur in sheep or birds. Extensive studies in cattle and in pigs have led investigators in
these fields to conclude that the predominant role of leptin is to act as a ‘barometer’ of
energy reserve, such that low levels of circulating leptin indicate energy deficiency. In
states of negative energy balance, the neuroendocrine system of the brain responds in
a predicable manner.
Another ‘appetite regulator’ that has received widespread attention is ghrelin,
produced by the stomach. In laboratory species and in humans, this hormone stim-
ulates food intake [reviewed in 8], but knockout mice show minimal phenotype [9],
raising the question as to the extent that ghrelin is involved in normal metabolic
regulation. Given the precedent that NPY knockout mice also have only minor
110 Clarke
70
Body
35
weight (kg)
0
40
Percent
20
body fat
0
GUR 16
(µmol/kg/min)
8
at 2 mU/kg/min
insulin 0
GUR 30
(µmol/kg/min)
15
at 6 mU/kg/min
insulin 0
Thin Normal Fat
In sheep [10] and in cattle [1], insulin levels rise after a meal and glucose levels fall, in
spite of the fact that these animals absorb most of their energy from substrates other
than sugar. The rise in insulin levels leads to a fall in plasma glucose levels.
It is well known that the obese state leads to a state of insulin resistance in humans
and in laboratory animals. Even though plasma glucose levels remain relatively stable
over a wide range of body weight in sheep (vide supra), insulin resistance develops
with increasing adiposity. This is exemplified in figure 1. In spite of this, obese rumi-
nants do not progress to fully developed type 1 diabetes, but the reason for this is not
known.
The notion that leptin informs the brain of metabolic status has been expounded by
various writers in the field and this is well demonstrated in a recent human study
There is a wide variety of genetic models of obesity in rodents, due to the ease with
which transgenics can be created. In domesticated animals, genetic models have
relied on selection or identification of naturally occurring mutations. A flock of sheep
in New Zealand was screened for backfat thickness and then the extreme phenotypes
were backcrossed through a number of generations [24]. The selection process cre-
ated flocks of fat, lean and normal (unselected) groups. Interestingly, the three groups
have similar body weights and food intake, but differ profoundly in the level of adi-
posity. Predictable differences are seen in circulating GH concentrations, with lean
animals having high levels and fat animals having low levels. No single mutation has
been identified in these animals and the genetic selection was probably polygenic.
This is consistent with observations made in humans, lending support to the notion
that body conformation has a high heritability. Studies are in progress to ascertain the
levels of gene expression for appetite-regulating peptides in the hypothalami of these
animals, but analysis of a range of body tissues may be required in order to under-
stand the genetic and/or phenotypic character of these animals.
Polymorphisms have been identified in the leptin gene of the pig and more exten-
sively in dairy cattle [reviewed in 25]. Various polymorphisms in the bovine gene
appear to be important for production (such as milk yield) and fertility. In some
cases, polymorphisms have been linked to food intake and relative adiposity, but
further studies are probably required before this can be regarded as conclusive.
Perhaps the most interesting allelic variant is the C to T transition in exon 2 of the
leptin gene, which causes an Arg25Cys transition in the mature protein [26]. The T
allele is associated with fatter carcasses. It has been speculated that the extra Cys in
the leptin sequence destabilises the molecule, which normally has a single Cys-Cys
disulphide bridge. Another explanation offered by these researchers is that the tran-
sition affects the A helix of the leptin molecule which affects binding to the leptin
receptor.
112 Clarke
90
Fig. 2. Natural variation in the propensity to gain weight with high-energy diet. Mature ovariec-
tomised ewes were run on pasture with supplementation of pasture hay. Over the period indicated
by the bars, the animals were given a dietary supplement of 500 g lupin grain per week as a high-
energy source. The animals were managed so that there was no competition for the supplement and
were weighed at regular intervals. As can be seen from the heavy lines representing two animals of
similar starting weight, particular animals have a propensity to gain much more weight than others
in a repeatable manner [Clarke, unpubl. data].
Predisposition to Obesity
The rodent model of diet-induced obesity that is unmasked by feeding a high fat diet
is one which has particular pertinence to the human condition of obesity and has
been extensively studied by Levin and colleagues. This model is predicated on the
propensity of a subset of animals in a population becoming obese (diet-induced
obese) when fed a high-fat diet as opposed to another subset that do not develop obesity
(diet-resistant) on the same diet. The model unmasks genetic variants in the popula-
tion. Such a model can also be applied to sheep even though these animals are rumi-
nants and do not eat high-fat foodstuffs. Figure 2 shows data from a group of animals
that were given a high-energy dietary supplement (lupin grain) for various periods
over a 5-year period. Although the animals were of similar body weight at the start of
the study, it is clear that some animals gained more weight on the high-energy sup-
plement than others. The data from this relatively small sample of animals suggest
that the relative propensity to become obese with high-energy intake is genetically
determined in sheep as in other species.
114 Clarke
Understanding the particular dietary preferences and the factors that drive appetite
and appetite preference during migration would be most instructive.
The study of hibernating animals could also prove most instructive with respect to
the mechanisms controlling food intake and energy expenditure. Whereas much is
known about the phenomenon and the switch from carbohydrate to fatty acid metab-
olism, very little is known about the central control mechanisms [32]. Understanding
the profound circannual cycles of appetite drive, fat deposition and altered metabo-
lism during torpor could provide key indicators of points at which fat accumulation
and/or utilisation could be manipulated.
Large animals with relatively long gestation periods are ideal for the study of fetal
development because the fetus can be accessed surgically for the purpose of sampling
or manipulation. Studies on effects of maternal overfeeding, underfeeding, disease,
environmental stress (such as heat) and toxins have been carried out in ruminants,
pigs and horses. Intra-uterine growth retardation of the fetus in the sheep has been
widely used as a model of low birth weight. This can be achieved by surgical removal
of some of the placental caruncles, restriction of the uterine arterial flow, prenatal
testosterone treatment (female fetuses only) or altered nutrition (over-nutrient or
under-nutrition) of the mother. The result is reduced birth weight, followed by catch-
up growth with significant sequelae in adult life, such as disturbed metabolic ‘set point’,
insulin resistance, hypertension and relative obesity amongst other things [reviewed in
33] and with focus on large animals [34]. In general, intra-uterine growth retardation
results in reduced muscle fibre and increased adipose mass in the offspring.
Some work on possible epigenetic effects of environmental factors has been car-
ried out in sheep. For example, treatment of ewes with bisphenol between days 30–90
of gestation led to reduced birth weights and minor reproductive disruptions, but
metabolic effects were not reported [35]. It would be of considerable interest to ascer-
tain whether domesticated and/or wild animals show metabolic perturbations and/or
develop obesity following endocrine disruption, since exposure to environmental dis-
ruptors such as pesticides is likely in particular habitats.
Conclusion
This brief review has sought to highlight how work on species other than laboratory
rodents has contributed to our understanding of mechanisms that control adiposity.
The examples that have been discussed exemplify the utility of work in these species,
in an attempt to promulgate the notion that examination of a range of animal models
will enhance our knowledge of the fundamental mechanisms of homeostasis.
116 Clarke
23 Chan JL, Matarese G, Shetty GK, Raciti P, Kelesidis I, 29 Clarke IJ, Rao A, Chilliard Y, Delavaud C, Lincoln
Aufiero D, De Rosa V, Perna F, Fontana S, Mantzoros CS: GA: Photoperiod effects on gene expression for
Differential regulation of metabolic, neuroendocrine, hypothalamic appetite-regulating peptides and food
and immune function by leptin in humans. Proc Natl intake in the ram. Am J Physiol Regul Integr Comp
Acad Sci USA 2006;103:8481–8486. Physiol 2003;284:R101–R115.
24 Francis SM, Venters SJ, Duxson MJ, Suttie JM: 30 Maillet D, Weber JM: Performance-enhancing role of
Differences in pituitary cell number but not cell type dietary fatty acids in a long-distance migrant shore-
between genetically lean and fat Coopworth sheep. bird: the semipalmated sandpiper. J Exp Biol 2006;
Domest Anim Endocrinol 2000;18:229–239. 209:2686–2695.
25 Liefers SC, Veerkamp RF, Te Pas MF, Chilliard Y, van 31 Kochan Z, Karbowska J, Meissner W: Leptin is syn-
der Lende T: Genetics and physiology of leptin in thesized in the liver and adipose tissue of the dunlin
periparturient dairy cows. Domest Anim Endocrinol (Calidris alpina). Gen Comp Endocrinol 2006;148:
2005;29:227–238. 336–339.
26 Buchanan FC, Fitzsimmons CJ, Van Kessel AG, Thue 32 Carey HV, Andrews MT, Martin SL: Mammalian
TD, Winkelman-Sim DC, Schmutz SM: Association hibernation: cellular and molecular responses to
of a missense mutation in the bovine leptin gene with depressed metabolism and low temperature. Physiol
carcass fat content and leptin mRNA levels. Genet Sel Rev 2003;83:1153–1181.
Evol 2002;34:105–116. 33 Gluckman PD, Hanson MA, Pinal C: The develop-
27 Mesteig K, Tyler NJ, Blix AS: Seasonal changes in mental origins of adult disease. Matern Child Nutr
heart rate and food intake in reindeer (Rangifer 2005;1:130–141.
tarandus tarandus). Acta Physiol Scand 2000;170: 34 Wu G, Bazer FW, Wallace JM, Spencer TE: Board-
145–151. invited review: intrauterine growth retardation:
28 Anukulkitch C, Rao A, Dunshea FR, Blache D, implications for the animal sciences. J Anim Sci
Lincoln GA, Clarke IJ: Influence of photoperiod and 2006;84:2316–2337.
gonadal status on food intake, adiposity and gene 35 Savabieasfahani M, Kannan K, Astapova O, Evans NP,
expression of hypothalamic appetite regulators in a Padmanabhan V: Developmental Programming:
seasonal mammal. Am J Physiol Regul Integr Comp Differential Effects Of Prenatal Exposure To
Physiol 2006;292:R242–R252. Bisphenol-A Or Methoxychlor On Reproductive
Function. Endocrinology 2006.
Abstract
The current worldwide epidemic of obesity and metabolic diseases has energised the search for new
approaches to treat these conditions. Type 2 diabetes appears to involve an interplay between suscepti-
ble genetic backgrounds and environmental factors including highly calorific westernised diets. The lat-
ter may generate ‘glucolipotoxic’ conditions which affect both the pancreatic β -cell and insulin-sensitive
tissues. Here we focus on efforts to better understand the basic signalling mechanisms through which
the β -cell senses changes in glucose concentration and how this process may become defective in type
2 diabetes. The recent demonstrations, through whole genome association studies, of important roles
for genes involved in the control of cell cycle, as well as intracellular ion homeostasis, further highlight
the central role of the β -cell in both the pathogenesis of the disease and as a therapeutic target.
Copyright © 2008 S. Karger AG, Basel
Diabetes mellitus is emerging as an ‘epidemic’ of the 21st century, currently afflicting 150
million individuals and projected to affect 200 million by 2025 [1]. Likewise, obesity is a
growing health care problem with ⬃20% of UK. (http://www.statistics.gov.uk/StatBase/)
and 30% of the U.S. (http://www.cdc.gov/nccdphp/dnpa/obesity/) adult population now
clinically obese. In this review we will discuss potential mechanisms whereby -cell
function and/or mass are altered both in each condition.
Insulin secretion is usually tightly regulated and ensures near-constant blood glucose
levels in the face of changes in supply and utilisation.
At the level of the individual -cell, glucose is generally believed to prompt the first
phase of insulin secretion via metabolic coupling mechanisms [2–6]. Uptake of
glucose involves a facilitative glucose transporter (Glut2 in rodents; Glut2 or Glut1 in
man) [7], followed by glucokinase-mediated glucose phosphorylation [8–10] and the
glycolytic generation of pyruvate. Oxidation of pyruvate by mitochondria is strongly
favoured over its conversion to lactate by remarkably low levels of lactate dehydroge-
nase [11–14] and plasma membrane lactate/monocarboxylate (MCT) transport activ-
ity [15, 16] in -cells: disposal of glycolytic NADH is facilitated by correspondingly
elevated levels of mitochondrial glycerolphosphate dehydrogenase activity [11, 17] and
ensures that at least 85% of glucose carbon is fully oxidised to CO2 and H2O [11, 13].
Enhanced mitochondrial oxidative metabolism [18] then leads to increases in intracel-
lular ATP/(ADPAMP) ratio [19, 20] and the closure of ATP-sensitive K (KATP)
channels [21–23] followed by cell depolarisation. The consequent opening of L-type
voltage-gated Ca2 channels and Ca2 influx [24] prompt the exocytosis of preformed
large dense core insulin-containing vesicles at the plasma membrane [25] via still
incompletely defined vesicle-associated Ca2 sensors such as synaptotagmin V [26],
Ca2-dependent activator protein for secretion [27] or cysteine string protein [28, 29].
Increases in intracellular citrate, malonyl-CoA [30] or glutamate [31] have each
been proposed to play a role in amplification, although both the ‘malonyl-CoA’ [32]
and ‘glutamate’ [33] hypotheses have been challenged.
A further important player in the control of insulin secretion are nutrient-sensitive
protein kinases including AMP-activated protein kinase (AMPK) [34–38] and per-
arnt-sim kinase [39]. Whereas the former seems likely to be regulated through
changes in the concentrations of adenine nucleotides [34] and to affect recruitment of
dense core vesicles to the cell surface [40], the means by which per-arnt-sim kinase is
regulated remain obscure. These mechanisms are illustrated in figure 1.
ATP
Lactate
X Ca2
Glut2 Pyruvate AMPK
Glucose
GK
Glucose G-6-P
PKC
IP3 Docking Priming Kiss and run
PLC DAG Ca 2
Ca2
Insulin release
Fig. 1. Molecular mechanisms of GSIS in the healthy -cell. Uptake and phosphorylation (via glu-
cokinase, GK) of glucose ultimately leads to increases in intracellular ATP concentrations, closure of
KATP, plasma membrane depolarisation and Ca2 influx. Fusion of docked vesicles at the plasma
membrane via a transient ‘kiss and run’ or ‘cavity recapture’ mechanisms leads to release of cargo
insulin. Increases in ATP/AMP ratio also cause inhibition of AMPK, leading to the mobilisation of vesi-
cles from a reserve to a readily releasable pool and the ‘second phase’ of insulin release. Note that
non-nutrient secretagogues such as acetyl-choline may potentiate secretion by activating phospho-
lipase C, generating Ca2 increase and the activation of protein kinase C (PKC) as well as other pro-
teins. See Rutter [6] for other details and the text for discussion of changes in the T2D -cell.
Genetic Contribution
As discussed above, both insulin resistance [49] and changes in -cell function [50,
51] appear to underlie the progression of T2D (fig. 2). Strongly suggestive of an inher-
ited component, concordance rates for T2D in monozygotic twins have been reported
to be as high as 90% [52], and the dynamics of insulin release (i.e. the size of the first
and second phases) are remarkably similar between monozygotic twins [53]. Clear
linkage has been demonstrated to individual genes in monogenic forms of diabetes
(‘maturity-onset diabetes of the young’, MODY) [54], and has been shown to involve
IGT
-Cell failure
Type 2 diabetes
Fig. 2. Contribution of -cell failure to the development of T2D. Increases in blood glucose concen-
tration during the development of diabetes are indicated by the pyramid on the left, showing the
change from normal (NGT) to impaired (IGT) glucose tolerance, before the onset of frank diabetes.
mutations in the transcription factors hepatocyte nuclear factor (HNF) 4 and HNF1
(MODY 1 and MODY3), as well as PDX1/IPF (MODY4) glucokinase (MODY2), all
of which affect -cell development or function. Recent studies have also revealed
mutations in subunits of the ATP-sensitive potassium channel, kir6.2 (encoded by
KCNJ11) [55] and SUR1 (ABCC8) [56] as contributing to certain cases of permanent
neonatal diabetes.
On the other hand, genetic associations with more common forms of T2D have
been more elusive [57]. Genome scanning in several different ethnic groups has iden-
tified chromosome regions harbouring susceptibility to more common forms of T2D
such as calpain 10 (CAPN10) [58], E23K in KCNJ11 [59] and P12A in PPARG [60],
where the ‘odds ratio’ (i.e. the increased likelihood of affected individuals having the
disease) is small (⬃1.2). Most recently, and dramatically, a highly significant associa-
tion with T2D has been demonstrated for polymorphisms in an intronic region of the
transcription factor TCF7L2 (formerly known as TCF-4) gene [61–75] involved in
wnt/-catenin signalling. The odds ratio for inheritance of the at-risk allele (i.e. the
increased risk of diabetes compared to the common allele) is ~1.5. These findings have
been confirmed in recent ‘whole genome scans’ [76–78], and it has been estimated
that as much as 20% of disease cases may be attributable to variants in this gene [74].
The functional significance of the effects of these intronic changes remains unclear,
Environmental Factors
As discussed above, the precise molecular basis of the changes in glucose sensing in
the T2D -cell remains as yet unclear. There is nonetheless good evidence for
decreased glucose-induced ATP rises, and oxygen consumption appears to be
decreased in diseased islets from T2D subjects [124] and in rodents, e.g. GK rat [44].
Changes in the expression of multiple genes were identified in islets from ZDF rats
[46], including in the pre-diabetic stage, where glucose-induced insulin secretion is
abnormal [47]. Of note, the latter study revealed decreases in the expression of the
glucose transporter, Glut2. Very recent studies have shown that Glut2 is subject to
post-translational modification (N-glycosylation), a process which may become
defective in T2D models [125].
A recent microarray-based study of human T2D islets [126] revealed changes in
the expression of a wide range of mRNAs, including the glycolytic genes glucose 6-
phosphate isomerise and phosphofructokinase (note that GK and Glut2 mRNAs dis-
played a non-significant tendency towards lower levels in T2D islets), and most
strikingly in the expression of a PAS domain containing transcription factor termed
ARNT/HIF1 (hydrocarbon nuclear receptor translocator/hypoxia-inducible factor
1, decreased 90%). Moreover, knock-out of ARNT selectively in the -cells of
mice or silencing in MIN6 cells led to defective glucose-induced insulin secretion.
Both in rodent and human studies, decreases in the expression of glycolytic genes
were observed, and seem likely to underlie the defective glucose-induced increases in
metabolic signalling. Interestingly, increases in the levels of lactate dehydrogenase,
which are vanishingly low in -cells [11] (see above and fig. 1) are also observed at
least in some models of -cell dysfunction including partial resection [127] and in
MicroRNAs
Another exciting area is that of microRNAs. Recent studies by Poy et al. [147] and by
Plaisance et al. [148] have revealed the important role played by this new class of
short RNA species, termed microRNAs [147], in regulating -cell function. Thus,
miR-375 regulates late events in vesicle movement and fusion [147] whereas miR-9
appears to be involved in the control of multiple genes including granuphilin/slp4
[148] which in turn regulate the secretory process. Similarly, our own data [149] indi-
cate that miR-124, an miR whose expression is massively induced during pancreatic
development, controls the expression of both key proteins involved in determining -
cell excitability and Ca2 influx (including KATP channel subunits) but also regulating
the sensitivity of the secretory machinery to these Ca2 changes.
Acknowledgements
G.A.R. is supported by grants from the Wellcome Trust and the Medical Research Council UK.
L.E.P. thanks the American Heart Association for a Fellowship.
References
1 Zimmet P, Alberti KG, Shaw J: Global and societal 3 Meglasson MD, Matschinsky FM: Pancreatic islet
implications of the diabetes epidemic. Nature 2001; glucose metabolism and regulation of insulin secre-
414:782–787. tion. Diabetes Metab Rev 1986;2:163–214.
2 Ashcroft FM, Ashcroft SJM: Mechanisms of insulin 4 Ishihara H, Wollheim CB: What couples glycolysis to
secretion; in Insulin, Molecular Biology to Pathology mitochondrial signal generation in glucose-stimu-
Ashcroft FM and Ashcroft SJH (eds.) Oxford lated insulin secretion? Iubmb Life 2000;49: 391–395.
University Press, Oxford, New York, Toronto, 1992, 5 Rutter GA: Nutrient-secretion coupling in the pan-
pp 97–150. creatic islet -cell: recent advances. Mol Aspects Med
2001;22:247–284.
Abstract
The endogenous cannabinoid system (ECS) is a neuromodulatory system recently recognized to have a
role in the regulation of various aspects of eating behavior and energy balance through central and
peripheral mechanisms. In the central nervous system, cannabinoid type 1 receptors and their endoge-
nous ligands, the endocannabinoids, are involved in modulating food intake and motivation to consume
palatable food. Moreover, the ECS is present in peripheral organs, such as liver, white adipose tissue, mus-
cle, and pancreas, where it seems to be involved in the regulation of lipid and glucose homeostasis.
Dysregulation of the ECS has been associated with the development of obesity and its sequelae, such as
dyslipidemia and diabetes. Conversely, recent clinical trials have shown that cannabinoid type 1 receptor
blockade may ameliorate these metabolic abnormalities. Although further investigation is needed to
better define the actual mechanisms of action, pharmacologic approaches targeting the ECS may pro-
vide a novel, effective option for the management of obesity, type 2 diabetes and cardiovascular disease.
Copyright © 2008 S. Karger AG, Basel
The appetite-stimulating and antiemetic properties of Cannabis sativa have been known
for centuries. However, although synthetic and plant-derived cannabinoids (CBs) such
as ⌬9-tetrahydrocannabinol (⌬9-THC) have long been recognized to influence food
intake, only recently CB receptors, endogenous CBs (endocannabinoids), and their
biosynthetic and degradative pathways have been identified. Collectively, these discover-
ies have led to the characterization of an endogenous signaling system now known as the
endocannabinoid system (ECS). In mammals, the ECS has been recently recognized to
have a role in the modulation of several physiological processes, including neuroprotec-
tion, regulation of hormone secretion, locomotion, and energy homeostasis, among oth-
ers. The purpose of this chapter is to outline the recent advances in our understanding of
the role of the ECS in energy balance regulation and to discuss the clinical relevance of
this work with respect to obesity and associated metabolic risk factors.
Cannabinoid agonists
Ca2⫹
Extracellular
CB1
G
Intracellular K⫹
AC
MAPK
Fig. 1. Cellular effects of CB1 receptor activation in the brain. Activation of CB1 through the action
of CB receptor agonists, such as ECs, induces stimulation of Gi/o heterotrimetic proteins (G). CB1 stim-
ulation of Gi/o proteins is directly coupled to stimulation of inwardly rectifying K⫹ channels and inhi-
bition of voltage-activated Ca2⫹ channels, effects that in neurons lead to inhibition of
neurotransmitter release. Activation of CB1 leads also to inhibition of adenylate cyclase (AC) or to
stimulation of mitogen-activated protein kinase (MAPK). Modulation of these intracellular pathways
may affect gene expression. See also references [1–3, 7].
CB Receptors
The first CB receptor was characterized using reverse pharmacology approach in
1988. This receptor was named CB1 after the second CB receptor, named CB2, was
cloned in 1993. CB1 is found in the central nervous system as well as in peripheral
organs [1]. CB2 is mainly localized in the immune system, and shares a 48% homol-
ogy with CB1 [2, 3]. Both are 7-transmembrane-spanning Gi/o receptors that inhibit
adenylyl cyclase and activate mitogen-activated protein kinase [1, 2]. In addition,
the CB1 receptor is reported to affect both potassium and calcium channels (fig. 1).
CB1 receptors are among the most abundant G-protein-coupled receptors in the
brain, having similar densities as receptors for ␥-aminobutyric acid (GABA) and
glutamate-gated ion channels [2]. In the central nervous system, the distribution of
CB1 receptors is heterogeneous, with higher densities in the basal ganglia, hip-
pocampus and cerebellum [2]. CB1 expression in the hypothalamus, a key integra-
tive area in the regulation of energy homeostasis, is relatively low; however,
activation of hypothalamic CB1 is highly efficient [2]. Activation of CB1 receptors,
136 Cota
usually located presynaptically, modulates the release of several neurotransmitters,
such as GABA, dopamine (DA), noradrenaline, glutamate and serotonin. CB1
receptors are expressed by astrocytes as well as by neurons [4]. Interestingly, activa-
tion of CB1 on astrocytes increases available energy to local neuronal circuits, and
the administration of CB agonists increases overall energy metabolism in the brain
[4]. CB1 are also located on nerve terminals innervating the gastrointestinal tract as
well as found in other organs involved in energy balance regulation, such as white
adipose tissue, liver, pancreas and skeletal muscle [reviewed in 5]. Although CB2 are
located mainly in tissues of the immune system, they have recently also been identi-
fied in neurons in several regions of the brain [6]. For an in-depth description of the
characteristics and functions of CB receptors, the reader should refer to recent
reviews [1–3, 7].
Endocannabinoids
All endocannabinoids (ECs) identified so far are amides, esters or ethers of arachi-
donic acid, a long-chain polyunsaturated fatty acid that is a constituent of cell mem-
branes [1]. Anandamide (the ethanolamide of arachidonic acid) was the first
described EC, and 2-AG was later identified. These two ECs are the best character-
ized and their properties have been largely studied in recent years. Their biosynthetic
and degradative pathways have been recently reviewed [1, 2]. Other putative ECs
have been discovered [1]; however, neither the mechanisms of synthesis and deacti-
vation, nor the specific functions have been defined for these other compounds. The
concentrations of anandamide and 2-AG within the brain vary considerably by area.
In most brain areas, 2-AG is around two orders of magnitude more abundant than
anandamide, with the highest levels of both occurring in the brainstem, hippocampus
and striatum [1, 2, 7]. Both anandamide and 2-AG are produced in several peripheral
tissues, including skin, gut, liver, adipose tissue and testis [1, 2, 7]. Moreover,
although its physiological function is unknown, anandamide is present in the blood-
stream where it is bound to albumin [reviewed in 5]. Anandamide and 2-AG each
bind to both CB1 and CB2 receptors. However, 2-AG is a full agonist at CB1, whereas
anandamide is a partial CB1 agonist [1]. Unlike the case for classical neurotransmit-
ters which are synthesized and stored in vesicles until needed, ECs are made de novo
during neuronal hyperactivity and immediately released into the extracellular space
of synapses [7]. As they bind to CB1, a sequence of intracellular events is induced,
and the ECs are then rapidly eliminated via reuptake and degradation by neurons and
possibly glia cells [1, 2]. This ‘on demand’ process is mainly utilized by neurons dur-
ing periods of high-frequency membrane stimulation [7]. For instance, many CNS
CB1 receptors are localized presynaptically on GABAergic interneurons. ECs released
from postsynaptic membranes therefore must diffuse or be transported retrogradely
in order to interact with CB1 and thereby decrease the release of presynaptic neuro-
transmitter. The generally accepted model is that when electrical activity in the form
The ECS influences energy balance at many sites in the brain and throughout the
body, with the net and highly coordinated effect being anabolic, i.e. increased EC
activity enhances food intake and supports fat deposition. Within the brain, CB1 and
ECs are present in both hypothalamic and extrahypothalamic circuits. Both networks
contribute to excessive eating when ECS activity is high. Within peripheral organs
including adipose tissue, liver and skeletal muscle, locally produced ECs are also ana-
bolic, contributing to lipogenesis and fat storage and reduced energy expenditure. In
the gastrointestinal system, ECs are thought to increase fuel absorption and decrease
satiety. Thus, the ECS biases behavioral and metabolic processes to promote obesity.
Consistent with this, genetic polymorphisms of components of the ECS have been
associated with overweight and obesity in humans [10].
138 Cota
The ECS and the Hypothalamus
The hypothalamus is an important integrative center for the control of energy
homeostasis. Neurons in the ARC are sensitive to the circulating hormones leptin
and insulin, and the ARC in turn projects to many other sites in the hypothalamus
and elsewhere in the brain [11]. Two categories of ARC neurons are known to
participate in energy balance regulation. ARC pro-opiomelanocortin (POMC)
neurons synthesize the neurotransmitter ␣-melanocyte-stimulating hormone (␣MSH),
which acts at melanocortin receptors to induce reduction in food intake and body
weight [11]. Increased plasma levels of leptin and/or insulin, both signals of
increased body fat, stimulate POMC neurons, initiating the catabolic response.
POMC cells in the ARC release ECs [9], and leptin reduces hypothalamic endo-
cannabinoid levels, inhibits ECs’ action in several hypothalamic areas, thus facili-
tating its anorexigenic effect [12]. Other ARC neurons synthesize the anabolic
neurotransmitters neuropeptide Y (NPY) and agouti-related protein [11]. NPY
acts on Y receptors throughout the hypothalamus to stimulate food intake,
whereas agouti-related protein antagonizes ␣MSH, thus inhibiting the catabolic
action of POMC cells [11]. Recent findings have demonstrated that CB agonists
increase the secretion of NPY in the hypothalamus, consistent with ECs increasing
food intake. Moreover, hypothalamic 2-AG levels increase during fasting, decline
as animals are re-fed and return to normal values when animals eat to satiation
[13]. CB1 mRNA has been colocalized with many hypothalamic neuropeptides
involved in energy balance regulation, including corticotropin-releasing hormone,
cocaine-amphetamine-regulated transcript, pre-pro-orexin and melanin-concen-
trating hormone [14]. Moreover, as discussed above, CB1 are also expressed by
GABAergic neurons entering the ARC [9]. Intrahypothalamic administration of
⌬9-THC increases food intake in laboratory rats. Anandamide not only increases
food intake when administered systemically, but does so also when injected into
the hypothalamus; and pretreatment with the selective CB1 antagonist SR 141716
(rimonabant) attenuates anandamide-induced hyperphagia [reviewed in 5].
Interestingly, animals lacking CB1 (CB1⫺/⫺ mice) have reduced food intake,
decreased body weight and a lean phenotype [14]. CB1⫺/⫺ mice have increased lev-
els of hypothalamic anorexigenic neuropeptides [14], and they do not increase
their food intake when treated with orexigenic compounds, such as NPY.
Moreover, the hyperphagia elicited by ghrelin and by orexin A can be attenuated by
administration of SR 141716; and CB1 are coexpressed with orexin 1 receptors
[reviewed in 5]. Thus, the actions of several orexigenic peptides within the hypo-
thalamus, including NPY, orexin and ghrelin, seem to be facilitated by ECs.
Furthermore, administration of SR 141716 attenuates the orexigenic effect of
melanocortin antagonists, but the administration of melanocortin agonists does
not interfere with the action of CB agonists, suggesting that endocannabinoids may
act downstream of melanocortins [reviewed in 5].
140 Cota
greater number of CB1, and CB1 expression is higher in adipocytes of obese as com-
pared to lean animals [18]. Studies in humans also indicate that mature adipocytes
have increased CB1 expression as compared to preadypocytes [19]. ECs are also pro-
duced in the liver [20] where they help regulate hepatic blood flow. Interestingly, acti-
vation of CB1 in hepatic tissue induces lipogenic enzymes, thus stimulating de novo
fatty acid synthesis [20]. Diet-induced obese mice have increased activity of this
hepatic lipogenic pathway as well as elevated expression of CB1 and increased anan-
damide levels in the liver [20]. Interestingly, recent evidence has shown that the activ-
ity of AMPK (an enzyme that may mediate the action of antidiabetic drugs, such as
metformin) is inhibited by CB agonists in both liver and adipose tissue [21]. Most
meal-generated satiety signals from the gastrointestinal tract reach the hindbrain via
vagal afferent nerves, and the signals are then conveyed to the hypothalamus and else-
where in brain areas controlling food intake. CB1 have been localized in vagal afferent
nerves, in neurons in the dorsal vagal complex and in vagal efferent nerves. The obser-
vation that cholecystokinin (CCK) inhibits the expression of CB1 on vagal afferents
[22] is consistent with CCK’s role as a satiety signal that reduces food intake. Moreover,
anandamide levels are greatly increased in the small intestine during fasting, and
decreased during refeeding, suggesting a role of ECs in influencing food intake via
modulation of gastrointestinal signals [reviewed in 5]. In fact, ECs modulate many
aspects of gastrointestinal function including gastric emptying and intestinal peristal-
sis. Although its function is less defined, the ECS is also present in the muscle (with
increased levels of CB1 expression during exposure to high fat diet) and in the pan-
creas. Furthermore, few studies suggest that ECs might affect energy expenditure. The
administration of the CB1 antagonist SR 141716 reportedly increases glucose uptake
and basal oxygen consumption of skeletal muscle from genetically obese mice, and
chronic administration of SR 141716 elicits increased expression of genes for thermo-
genesis in brown adipose tissue [reviewed in 5].
There is mounting evidence that the ECS, normally transiently activated, might become
tonically overactivated in both animal models of genetic and diet-induced obesity and
in human obesity. For instance, the levels of hypothalamic ECs are increased in geneti-
cally obese rodents with defective leptin signaling, and treatment of these genetically
obese mice with a CB1 antagonist attenuates their hyperphagia and reduces their weight
gain, implying that overactivation of the hypothalamic ECS system may be a contribut-
ing factor in the development of obesity [12]. Interestingly, CB1⫺/⫺ mice do not become
obese on a high-fat diet and have increased sensitivity to the catabolic action of leptin
[reviewed in 5]. In animal models of diet-induced obesity, there is evidence of increased
CB1 receptor expression and ECs levels in both adipose tissue and liver. In obese
women, plasma anandamide and 2-AG levels are significantly elevated compared with
When rodents are fed a high-fat diet, they overeat and become obese. In one series of
experiments, mice maintained on high-fat diet and chronically administered the CB1
antagonist SR 141716 ate less food [reviewed in 5]. Significantly, the anorectic action
of the antagonists lessened and then disappeared altogether over a week or two in
spite of continued dosing, indicating the development of tolerance. Nevertheless,
there was a sustained reduction of body weight and body fat mass over the next sev-
eral weeks, suggesting that possibly other actions of the CB1 antagonists were contin-
uing to exert metabolic effects. At the end of the experiments, mice receiving the CB1
antagonists chronically had reduced body weight, body fat, plasma leptin and plasma
insulin, and an improved lipid profile, perhaps due to the ability of the drug to
increase adiponectin expression and secretion from adipose tissue [reviewed in 5].
Adiponectin, differently from other adipokines, improves insulin sensitivity and
induces fatty acid oxidation in muscle and liver. Somehow related to these findings,
experiments using CB1⫺/⫺ mice revealed that the lean phenotype is predominantly
caused by decreased caloric intake when the mice are young, whereas peripheral
metabolic factors are the major cause of maintaining the lean phenotype in the adults
[14]. As discussed above, CB1⫺/⫺ mice are resistant to diet-induced obesity. They
maintain their lean phenotype and they do not develop hyperglycemia or insulin
resistance. Accordingly, chronic treatment with the CB1 antagonist in mice exposed
to high-fat diet altered gene expression profiles in both white and brown adipose tis-
sue to support fat oxidation and increased thermogenesis [reviewed in 5]. Therefore,
142 Cota
chronic treatment with a selective CB1 antagonist reduces body fat as well as
improves many alterations commonly associated with the metabolic syndrome.
Hence, it seems that the early transient reduction of food intake when animals are
started upon the antagonist cannot account for the sustained improvement in meta-
bolic parameters. Recent data from clinical studies seem also to support this hypoth-
esis. Clinical trials conducted on overweight and obese patients with BMI ⱖ27 who
were on a hypocaloric diet, have shown that oral CB1 receptor antagonist (rimona-
bant 20 mg/day) administration for 1 year resulted in weight loss of up to 8.6 kg com-
pared with 2.3 kg in placebo-treated individuals [23, 24]. The reduction in body
weight occurred during the first 36–40 weeks of the study period. Long-term treat-
ment (2 years) with CB1 receptor antagonist was able to stabilize body weight and
prevent weight regain [25]. CB1 receptor antagonist (rimonabant 20 mg/day) treat-
ment in overweight and obese individuals also improved lipid profiles by increasing
high-density lipoprotein cholesterol (HDL) and decreasing triglycerides [23, 24].
Levels of HDL increased continuously throughout 2-year treatment with the CB1
receptor antagonist, whereas body weight stabilized [25]. Significantly, only a portion
(approximately 50%) of the effect associated with CB1 receptor antagonist treatment
on HDL and triglycerides was attributed to weight loss. Hence, the additional
improvement in lipid profile appears to involve a CB1 receptor-mediated effect that is
independent of weight loss [25]. Moreover, fasting glucose and fasting insulin in
overweight and obese individuals are decreased following CB1 receptor antagonist
therapy [24]. The amelioration of metabolic indices related to insulin sensitivity after
treatment with CB1 receptor antagonists are, at least in part, a consequence of the
decrease in body weight, and specifically in adipose tissue. However, a direct effect of
CB1 receptor antagonists on peripheral metabolism, specifically glucose homeostasis,
cannot be currently excluded. A possible explanation that has been given for the
improved glucose metabolism resides in the increased mRNA and plasma levels of
adiponectin observed after treatment with CB1 receptor antagonists.
Conclusions
References
1 Piomelli D: The molecular logic of endocannabinoid 10 Sipe JC, Waalen J, Gerber A, Beutler E: Overweight
signalling. Nat Rev Neurosci 2003;4:873–884. and obesity associated with a missense polymor-
2 Freund TF, Katona I, Piomelli D: Role of endogenous phism in fatty acid amide hydrolase (FAAH). Int J
cannabinoids in synaptic signaling. Physiol Rev Obes (Lond) 2005;29:755–759.
2003;83:1017–1066. 11 Schwartz MW, Woods SC, Porte D Jr, Seeley RJ,
3 Howlett AC, Barth F, Bonner TI, Cabral G, Casellas P, Baskin DG: Central nervous system control of food
Devane WA, Felder CC, Herkenham M, Mackie K, intake. Nature 2000;404:661–671.
Martin BR, Mechoulam R, Pertwee RG: International 12 Di Marzo V, Goparaju SK, Wang L, Liu J, Batkai S,
Union of Pharmacology. XXVII. Classification Jarai Z, Fezza F, Miura GI, Palmiter RD, Sugiura T,
of cannabinoid receptors. Pharmacol Rev 2002;54: Kunos G: Leptin-regulated endocannabinoids are
161–202. involved in maintaining food intake. Nature 2001;
4 Stella N: Cannabinoid signaling in glial cells. Glia 410:822–825.
2004;48:267–277. 13 Kirkham TC, Williams CM, Fezza F, Di Marzo V:
5 Cota D, Woods SC: The role of the endocannabinoid Endocannabinoid levels in rat limbic forebrain and
system in the regulation of energy homeostasis. hypothalamus in relation to fasting, feeding and sati-
6 van Sickle MD, Duncan M, Kingsley PJ, Mouihate A, ation: stimulation of eating by 2-arachidonoyl glyc-
Urbani P, Mackie K, Stella N, Makriyannis A, Piomelli erol. Br J Pharmacol 2002;136:550–557.
D, Davison JS, Marnett LJ, Di Marzo V, Pittman QJ, 14 Cota D, Marsicano G, Tschop M, Grubler Y,
Patel KD, Sharkey KA: Identification and functional Flachskamm C, Schubert M, Auer D, Yassouridis A,
characterization of brainstem cannabinoid CB2 recep- Thone-Reineke C, Ortmann S, Tomassoni F, Cervino
tors. Science 2005;310:329–332. C, Nisoli E, Linthorst AC, Pasquali R, Lutz B, Stalla
7 Di Marzo V, Matias I: Endocannabinoid control of GK, Pagotto U: The endogenous cannabinoid system
food intake and energy balance. Nat Neurosci affects energy balance via central orexigenic drive
2005;8:585–589. and peripheral lipogenesis. J Clin Invest 2003;112:
8 Losonczy A, Biro AA, Nusser Z: Persistently active 423–431.
cannabinoid receptors mute a subpopulation of hip- 15 Saper CB, Chou TC, Elmquist JK: The need to feed:
pocampal interneurons. Proc Natl Acad Sci USA homeostatic and hedonic control of eating. Neuron
2004;101:1362–1367. 2002;36:199–211.
9 Hentges ST, Low MJ, Williams JT: Differential regu- 16 Cota D, Tschop MH, Horvath TL, Levine AS:
lation of synaptic inputs by constitutively released Cannabinoids, opioids and eating behavior: the mol-
endocannabinoids and exogenous cannabinoids. J ecular face of hedonism? Brain Res Brain Res Rev
Neurosci 2005;25:9746–9751. 2006;51:85–107.
144 Cota
17 Kirkham TC, Williams CM: Endogenous cannabi- 22 Burdyga G, Lal S, Varro A, Dimaline R, Thompson
noids and appetite. Nutr Res Rev 2001;14:65–86. DG, Dockray GJ: Expression of cannabinoid CB1
18 Bensaid M, Gary-Bobo M, Esclangon A, Maffrand JP, receptors by vagal afferent neurons is inhibited by
Le Fur G, Oury-Donat F, Soubrie P: The cannabinoid cholecystokinin. J Neurosci 2004;24:2708–2715.
CB1 receptor antagonist SR141716 increases Acrp30 23 Despres JP, Golay A, Sjostrom L; Rimonabant in
mRNA expression in adipose tissue of obese fa/fa rats Obesity-Lipids Study Group: Effects of rimonabant
and in cultured adipocyte cells. Mol Pharmacol on metabolic risk factors in overweight patients with
2003;63:908–914. dyslipidemia. N Engl J Med 2005;353: 2121–2134.
19 Engeli S, Bohnke J, Feldpausch M, Gorzelniak K, 24 van Gaal LF, Rissanen AM, Scheen AJ, Ziegler O,
Janke J, Batkai S, Pacher P, Harvey-White J, Luft FC, Rossner S; RIO-Europe Study Group: Effects of the
Sharma AM, Jordan J: Activation of the peripheral cannabinoid-1 receptor blocker rimonabant on
endocannabinoid system in human obesity. Diabetes weight reduction and cardiovascular risk factors in
2005;54:2838–2843. overweight patients: 1-year experience from the
20 Osei-Hyiaman D, DePetrillo M, Pacher P, Liu J, RIO-Europe study. Lancet 2005;365:1389–1397.
Radaeva S, Batkai S, Harvey-White J, Mackie K, 25 Pi-Sunyer FX, Aronne LJ, Heshmati HM, Devin J,
Offertaler L, Wang L, Kunos G: Endocannabinoid Rosenstock J; RIO-North America Study Group:
activation at hepatic CB1 receptors stimulates fatty Effect of rimonabant, a cannabinoid-1 receptor
acid synthesis and contributes to diet-induced obe- blocker, on weight and cardiometabolic risk factors
sity. J Clin Invest 2005;115:1298–1305. in overweight or obese patients: RIO-North America:
21 Kola B, Hubina E, Tucci SA, Kirkham TC, Garcia EA, a randomized controlled trial. JAMA 2006;295:
Mitchell SE, Williams LM, Hawley SA, Hardie DG, 761–775.
Grossman AB, Korbonits M: Cannabinoids and
ghrelin have both central and peripheral metabolic
and cardiac effects via AMP-activated protein kinase.
J Biol Chem 2005;280:25196–25201.
Daniela Cota, MD
Department of Psychiatry, University of Cincinnati
2170 East Galbraith Road
Cincinnati, OH 45237 (USA)
Tel. ⫹1 513 558 5866, Fax ⫹1 513 558 8990, E-Mail daniela.cota@uc.edu
11-Hydroxysteroid Dehydrogenase
Type 1 and Obesity
Nicholas M. Morton ⭈ Jonathan R. Seckl
Endocrinology Unit, Centre for Cardiovascular Sciences, Queens Medical Research Institute,
Edinburgh University, Edinburgh, UK
Abstract
The metabolic syndrome consists of a constellation of co-associated metabolic abnormalities such as
insulin resistance, type 2 diabetes, dyslipidaemia, hypertension and visceral obesity. For many years
endocrinologists have noted the striking resemblance between this disease state and that associated
with Cushing’s syndrome. However, in the metabolic syndrome plasma cortisol levels tend to be normal
or lower than in normal individuals. Nevertheless there is strong evidence that glucocorticoid action
underlies metabolic disease, largely from rodent obesity models where removing glucocorticoids
reverses obesity and its metabolic abnormalities. The apparent paradox of similar metabolic defects –
despite the opposing plasma glucocorticoid profiles of Cushing’s and idiopathic metabolic syndrome –
remained intriguing until the discovery that intracellular glucocorticoid reactivation was elevated in adi-
pose tissue of obese rodents and humans. The enzyme that mediates this activation, conversion of corti-
sone (11-dehydrocorticosterone in rodents) to cortisol (corticosterone in rodents), locally within tissues is
11 β -hydroxysteroid dehydrogenase type 1 (11 β -HSD1). In order to determine whether elevated tissue
11 β -HSD1 contributed to obesity and metabolic disease, transgenic mice overexpressing 11 β -HSD1 in
adipose tissue or liver were made. Adipose-selective 11 β -HSD1 transgenic mice exhibited elevated
intra-adipose and portal, but not systemic corticosterone levels, abdominal obesity, hyperglycaemia,
insulin resistance, dyslipidaemia and hypertension. In contrast, transgenic overexpression of 11 β -HSD1
in liver yielded an attenuated metabolic syndrome with mild insulin resistance, dyslipidaemia, hyperten-
sion and fatty liver, but not obesity or glucose intolerance. Together with early data using non-selective
11 β -HSD1 inhibitors to insulin sensitise humans, this corroborated the notion that the enzyme may be a
good therapeutic target in the treatment of the metabolic syndrome. Further, a transgenic model of
therapeutic 11 β -HSD1 inhibition, 11 β -HSD1 gene knock-out (11 β -HSD1⫺/⫺) mice, exhibited improved
glucose tolerance, a ‘cardioprotective’ lipid profile, reduced weight gain and visceral fat accumulation
with chronic high-fat feeding. Recent evidence further suggests that high fat-mediated downregulation
of adipose 11 β -HSD1 may be an endogenous pathway that underpins adaptive disease resistance in
genetically predisposed mouse strains. This mechanism could feasibly make up a genetic component of
innate obesity resistance in humans. The efficacy of 11 β -HSD1 inhibitors has recently been extended to
include increased energy expenditure and reduction of arteriosclerosis, and therefore may be of signifi-
cant therapeutic value in the metabolic syndrome, with complementary effects upon liver adipose tis-
sue, muscle, pancreas and plaque-prone vessels. Copyright © 2008 S. Karger AG, Basel
The incidence of obesity and its associated metabolic syndrome has grown to ‘epidemic’
proportions globally. At least 300 million are estimated to be clinically obese (World
Health Organisation 2003) and more than three times this number are overweight.
Within the next 5 years a quarter of adults in the UK are expected to be clinically obese
and obesity of childhood, a predictor of morbidity and mortality in adulthood, is accel-
erating [1]. Whilst obesity associates with a substantially increased risk of serious
adverse effects ranging from type 2 diabetes, hypertension and cardiovascular disease to
cancers and arthritis, and thus to increased mortality, not every obese subject will get
these complications. The distribution of fat is important in determining the risks of
such disorders; excess upper body, centripetal or visceral adipose excess has a dire rep-
utation, whereas peripheral lower body or peripheral fat has little adverse effect.
There are strong morphological and metabolic resemblances between rare
Cushing’s syndrome of glucocorticoid excess and the extremely common insulin resis-
tance or metabolic syndrome [2]. Excess cortisol production and release in Cushing’s
syndrome causes predominantly visceral (abdominal) fat deposition, while peripheral
fat is reduced. This may result from opposing effects of glucocorticoids on these dis-
tinct adipose depots and their differential sensitivity both to catecholamine-mediated
fat mobilisation and insulin-mediated lipogenic effects. Whilst the association between
visceral adiposity and metabolic and cardiovascular disease risk is clear [3], the molec-
ular mechanisms linking the two are unknown. Current theories include evidence that
visceral fat is relatively more sensitive to lipolytic stimuli from the adrenergic receptor
system, though subcutaneous adipocytes are bigger and make a larger overall contribu-
tion to whole body lipolysis [4]. On the other hand other factors such as adipokines
and cortisol may be released from the visceral fat and have a pronounced and more
immediate detrimental impact on hepatic function because these fat depots drain
directly into the liver. This is the ‘portal hypothesis’ of insulin resistance [5]. In terms of
glucocorticoid action, glucocorticoid receptors are more highly expressed in visceral
fat [6, 7] and therefore glucocorticoids are expected to have a greater impact on both
metabolic responses and adipokine expression and release into the portal blood (fig. 1).
Thus, glucocorticoids may increase lipolysis and downregulate lipoprotein lipase, liber-
ating free fatty acids from peripheral fat, but also stimulate pre-adipocyte differentia-
tion and lipoprotein lipase expression, enhance glyceroneogenesis and triglyceride
synthesis in visceral adipose tissue [6–9]. However, despite the phenotypic similarities
with Cushing’s syndrome, in the metabolic syndrome plasma cortisol levels are mod-
estly if at all elevated and are usually reduced in ‘simple’ obesity [2, 9–10].
Ffa
adipokines Insulin Adrenal
glucagon Cortisol
Visceral fat Liver ?
Kidney
Ffa Excretion
adipokines clearance
cortisol
All Portal blood
Cortisone
tissues
Fig. 1. 11-HSD1 is a local tissue amplifier of glucocorticoid action. Glucocorticoid action is modulated
by 11-HSD1 (re-activation of inert cortisone in humans, 11-dehydrocorticosterone in rodents) and is
represented by the rightward-pointing green arrow boxes. The number of boxes represents an approxi-
mation of the relative expression levels of the enzyme within the tissues. Substrate for 11-HSD1 is cre-
ated principally by the inactivation of adrenal gland-derived circulating cortisol (corticosterone in
rodents) by the enzyme 11-HSD2 (leftward-pointing red arrow boxes) in the kidney. Circulating cortisol
(corticosterone) is generated largely by the adrenal gland through the action of the HPA axis of the brain
by release of ACTH from the anterior pituitary and by clearance through the kidney and liver. Tissue-spe-
cific changes in the expression level of 11-HSD1 may contribute to altered glucocorticoid action in
obesity and metabolic syndrome with increased adipose 11-HSD1 and reduced hepatic 11-HSD1.
11-HSD, discovered over 50 years ago, catalyses the interconversion of active gluco-
corticoids (cortisol, corticosterone) and their inert 11-keto forms (cortisone, 11-dehy-
drocorticosterone) that have very low affinity for nuclear receptors. 11-HSD activity
was subsequently described in a broad range of cells and tissues [26]. In the 1980s
Monder and White [27] purified an 11-HSD activity from rat liver. Homogenates,
microsomal preparations and purified enzyme catalysed both 11-dehydrogenation of
cortisol to inert cortisone and the 11-reduction of cortisone to active cortisol. As
with A-ring reductases, 11-HSD was originally thought to represent one of several
routes for clearing glucocorticoids. Subsequently, two 11-HSD isozymes, the prod-
ucts of distinct genes, were characterised and their cDNAs cloned [28, 29].
11-HSD type 2 is a high affinity, NAD-dependent 11-dehydrogenase which
catalyses the rapid conversion of active cortisol to inert cortisone [28, 30]. In adults
11-HSD2 is expressed principally in tissues where aldosterone induces its effects on
sodium excretion, including distal nephron, sweat glands, salivary glands and colonic
mucosa as well as the endothelium [30]. The 11-HSD2 enzyme excludes active glu-
cocorticoids from intrinsically non-selective MR in vivo [31]. Kidney 11-HSD2
serves also as a high throughput clearance route for circulating cortisol and indeed as
the principal generator of substrate for the 11-HSD1 enzyme and therefore regula-
tion of its activity may have profound effects on circulating and tissue glucocorticoid
levels where 11-HSD1 is expressed [32].
11-HSD1 has a much lower affinity for cortisol and corticosterone and is an
NADP(H)-dependent enzyme. 11-HSD1 is widely expressed in many tissues [33],
Liver
Studies using relatively non-selective liquorice-based inhibitors [58, 59] and, more
recently with selective 11-HSD1 inhibitors [60, 61] show that reduced activity of
11-HSD1 in liver is associated with features of reduced glucocorticoid action,
reduced glucocorticoid levels and increased insulin sensitivity in hepatocytes. 11-
HSD1⫺/⫺ mice develop normally and are viable, fertile and grossly normotensive
[56]. This model demonstrated that 11-HSD1 is the sole major 11-reductase, at
least in mice, since adrenalectomised 11-HSD1 knockout mice cannot convert
exogenous 11-dehydrocorticosterone to corticosterone. Plasma corticosterone levels
are modestly elevated at the diurnal nadir, presumably due to somewhat deficient
feedback upon the HPA axis (11-HSD1 is expressed in hippocampus, paraventricu-
lar nucleus and anterior pituitary) [62]. This is in contrast to the H6PDH null mice,
which have low circulating corticosterone because the 11-HSD1 lacks co-factor to
drive reduction of 11-dehydrocorticosterone and thus acts as a dehydrogenase [45].
Despite slightly elevated basal plasma corticosterone levels, 11-HSD1⫺/⫺ mice have
a phenotype compatible with impaired intracellular glucocorticoid regeneration and
increased insulin sensitivity. They show impaired induction of the key glucocorti-
coid-inducible hepatic enzymes PEPCK and glucose-6-phosphatase with fasting and
attenuated hyperglycaemic responses to novel environment stress or chronic high-fat
feeding [56, 63]. Importantly, the mice have more than adequate stress-induced HPA
axis responses [62] and do not exhibit hypoglycaemia with prolonged fasting [56].
11-HSD1⫺/⫺ mice also have lower plasma triglyceride and elevated ‘cardioprotec-
tive’ HDL cholesterol levels [63], including higher circulating levels of the HDL
apolipoprotein AI, whereas serum apolipoprotein CIII, which interferes with efficient
transfer of triglycerides into the liver, is reduced. There is evidence for increased
hepatic insulin sensitivity when the mice are re-fed after fasting, and hepatic expres-
sion of the co-ordinate transcription factor that drives fatty acid -oxidation, PPAR␣,
is elevated in 11-HSD1⫺/⫺ mice. Additionally, preliminary data suggest favourable
effects upon haemostasis as in liver the glucocorticoid-inducible A␣-fibrinogen tran-
script levels are reduced [63].
11b-HSD1 in Obesity
Studies in leptin-resistant obese Zucker rats and leptin deficient ob/ob mice revealed
that obesity was associated with decreased 11-HSD1 expression and activity in liver,
Blood from visceral adipose tissue drains via the hepatic portal vein to the liver.
Unsurprisingly therefore, aP2-11-HSD1 transgenic mice have elevated levels of
Deficiency of 11-HSD1
To confirm the key importance of reduced adipose tissue glucocorticoid levels in the
beneficial phenotype of 11-HSD1 null mice, a new strain ectopically expressing the
glucocorticoid inactivating isozyme, 11-HSD2, selectively in adipose tissue were
generated, again exploiting the aP2 promoter [100]. Expression levels were chosen to
11-HSD1 as a Drug Target for Therapy in Obesity and the Metabolic Syndrome
Acknowledgments
Work described from the authors’ laboratory was generously funded by a Wellcome Trust
Programme Grant (J.R.S.) and a Wellcome Trust Intermediate Research Fellowship (N.M.M.).
N.M.M. is a Wellcome Trust Career Development Fellow.
Abstract
Following the discovery of secretin in 1902, a host of further peptide hormones that are synthesised
and released from the gastrointestinal tract have been identified. While their roles in the regulation
of gastrointestinal function have been known for some time, it is now evident that many of these
hormones also physiologically regulate energy balance. Our understanding of how gut hormones
signal to the brain has advanced significantly in recent years. Several hormones, including peptide
YY, pancreatic polypeptide, oxyntomodulin, glucagon-like peptide 1 and cholecystokinin function as
satiety signals. In contrast, only ghrelin, produced by the stomach, has emerged as a putative hunger
signal, appearing to act both as a meal initiator and a long-term body weight regulator. Recent
research suggests that gut hormones can be manipulated to regulate energy balance in man
and that obese subjects retain sensitivity to the actions of gut hormones. The worldwide obesity
pandemic continues unabated, despite public health initiatives and current best therapy. Future
gut hormone-based therapies may provide an effective and well-tolerated treatment for obesity.
Copyright © 2008 S. Karger AG, Basel
The concept of the gut as an endocrine organ is hardly a new. The gut peptide secretin
was the first substance to be termed a hormone whilst the appetite inhibitory actions
of cholecystokinin (CCK) were first reported over 30 years ago [1, 2]. However, in
recent years, further scientific endeavour in this field has been motivated by the need
to develop new strategies to tackle the global pandemic of obesity.
The prevalence of obesity in adults has increased by over 75% worldwide since
1980. Given that obesity is causally associated with cardiovascular disease, type 2 dia-
betes, hypertension, stroke, obstructive sleep apnoea and certain cancers, this has
translated into healthcare costs of over half a billion pounds every year in the UK
alone. Obesity is not only a problem in the developed world, but is set to overtake
infectious diseases as the most significant contributor to ill-health worldwide and has
been classified as an epidemic by the World Health Organization [3].
Public health initiatives have failed to reverse the rising incidence of obesity.
Medical and behavioural interventions, with the exception of bariatric surgery, have
limited success, as discussed in the treatment section of this volume. This chapter will
focus on the peptide hormone signals from the gut that communicate the status of
body energy stores to the brain and the brain centres on which they act. These regu-
latory systems are not only of academic interest, but are likely to underpin any future
strategy to tackle obesity, by providing drug targets for the holy grail of safe sustain-
able weight loss.
Peripheral signals that regulate energy balance are often categorised as long and short
acting. Long-acting signals, typified by the adipocyte hormone leptin, characteristi-
cally reflect the levels of energy stores and regulate body weight and the amount of
energy stored as fat [see the chapter by Ahima and Osei, this vol., pp. 182–197].
Short-acting gastrointestinal signals are typified by gut hormones such as CCK and
mechanical factors, such as gastric distension, which characteristically relay a sense of
‘fullness’ resulting in post-prandial satiation and meal termination. Other gut hor-
mones, including peptide YY (PYY) and ghrelin appear to blur the boundaries
between long- and short-term signals, having features of both. Current evidence sug-
gests that they not only regulate appetite on a meal-by-meal basis but also participate
in longer term energy balance.
Neuro-hormonal signals from the gut and adipose tissue converge on the hypothala-
mus where they are integrated and in turn regulate energy intake and energy expen-
diture. The hypothalamic neurocircuits regulating energy balance have been
reviewed extensively [4, 5]. In brief, the hypothalamic arcuate nucleus (Arc), a cir-
cumventricular organ which is accessible to circulating factors via a relatively defi-
cient blood brain barrier, may be viewed as a conduit for peripheral signals. Two key
neuronal populations have been identified within the Arc with opposing effects on
energy balance. A group of neurones in the medial Arc co-express neuropeptide Y
(NPY) and agouti-related peptide (AgRP) and act to stimulate food intake and weight
gain. In contrast pro-opiomelanocortin (POMC) and cocaine- and amphetamine-
regulated transcript co-expressing neurones in the lateral Arc inhibit feeding and
promote weight loss. The balance of activity between these two groups of neurones is
critical to body weight regulation. The Arc neurones have multiple connections with
other central nervous system (CNS) sites of appetite regulation, including the brain-
stem and other hypothalamic nuclei.
166 Wren
Adiponectin
Resistin
Visfatin
Leptin
Adipose
tissue
Insulin
PP
Pancreas
Vagal
afferents
OXM
PYY 3-36
CCK + GLP-1
Ghrelin Stretch
Intestine Nutrients
Stomach
Fig. 1. Overview of peripheral factors regulating energy balance and their routes of signalling to the
brain; from Wren and Bloom [29].
The nucleus of the solitary tract (NTS) and the area postrema (AP), components
of the dorsal vagal complex are also key central integrators of peripheral signals. They
receive inputs from vagal afferents and circulating factors and are reciprocally con-
nected with hypothalamic nuclei controlling energy balance. These brainstem centres
can also respond independently to peripheral signals when communication with
higher brain centres are surgically interrupted [6]. In addition, inputs from the cortex
(emotional, social and behaviour cues) and the mesolimbic dopaminergic reward cir-
cuits influence energy balance and communicate with the hypothalamus.
Peripheral feedback to the hypothalamus is complex, as illustrated in figure 1. Many
circulating signals, including gut hormones, have access to the Arc and leptin, for exam-
ple, is thought to act primarily via a direct action here [5]. In contrast, other peripheral
Ghrelin is the only known circulating orexigen. In contrast, all the other peripheral
factors that regulate energy balance act to restrain eating and weight gain. Ghrelin
was discovered as an endogenous ligand for the growth hormone (GH) secretagogue
receptor (GHS-R1a) [7]. However, early work on this peptide demonstrated a GH-
independent action to powerfully increase food intake and body weight. The pre-
dominant focus of subsequent research has shifted onto the role of ghrelin in energy
balance [8–11].
Ghrelin is a 28-amino acid peptide, cleaved from a precursor, preproghrelin [7]. It
is principally synthesised in endocrine cells of the stomach, termed X/A-like or ghre-
lin cells, and particularly found in the gastric fundus. About two thirds to three quar-
ters of circulating ghrelin is of gastric origin. Lesser concentrations of ghrelin are
found throughout the small intestine, with the duodenum producing approximately
ten times less than the stomach and progressively lower concentrations found more
distally [12]. Ghrelin undergoes post-translational modification with attachment of a
medium-chain fatty acid, typically octanoic acid, to the serine-3 residue. This acyla-
tion is entirely unique among biologically active peptides and is required for ghrelin
to bind to and activate its classical receptor, the GHS-R1a [7]. The GHS-R1a is widely
expressed. In the CNS, it is found in areas involved in regulation of appetite and
energy balance including hypothalamic nuclei, the dorsal vagal complex and the
mesolimbic dopaminergic system. Peripherally, it is expressed in the pituitary, and
pharmacologically ghrelin acts at both pituitary and hypothalamic levels to power-
fully stimulate GH secretion [7, 9, 12, 13]. The physiological relevance of ghrelin in
GH regulation is debated. Ghrelin is not essential for GH secretion as ghrelin and
GHS-R1a null mice are not growth restricted, but it may play a role in augmentation
of GHRH-stimulated GH pulses. GHS-R1a receptor expression has also been
described in diverse peripheral sites including the myocardium, stomach, small intes-
tine, pancreas, colon, adipose tissue, liver, kidney, placenta, and T cells. An equally
diverse series of biological actions of exogenous ghrelin have been documented,
including effects on glucose homeostasis, gut motility, pancreatic exocrine secretion,
cardiovascular function, immunity and inflammation. The physiological relevance of
these actions remains unclear. There is also evidence for a number of pharmacologi-
cal actions of des-acyl (or unacylated) ghrelin, which must be mediated via receptors
other than the GHS-R1a. The physiological significance of these actions is contentious,
168 Wren
as reviewed elsewhere. However, experiments in GHS-R1a knockout mice have
definitively established that this receptor is required for the orexigenic and GH-stim-
ulating effects of acylated ghrelin [14–16].
When administered into the CNS, ghrelin stimulates food intake as potently as NPY,
previously the most powerful known orexigen, and more powerfully than any other
substance examined [8–10]. Ghrelin also stimulates appetite and food intake when
administered systemically in rodents [8, 11] and humans [17]. This property is unique
to ghrelin and not shared by any known neuropeptide or circulating hormone. The
duration of feeding stimulation in response to central or peripheral ghrelin administra-
tion is short, similar to that observed for central NPY. Indeed several lines of evidence
suggest that ghrelin acts via arcuate NPY/AgRP neurones, almost all of which express
the GHS-R1a. Ghrelin stimulates feeding most potently when injected directly into the
Arc and also stimulates release of NPY from hypothalamic explants in vitro [11, 13].
Arcuate NPY/AgRP neurones are activated by ghrelin, as demonstrated by enhanced c-
fos, NPY and AgRP expression following ghrelin administration and by electrophysio-
logical studies. Further, the orexigenic actions of ghrelin are abolished in NPY/AgRP
dual knockout mice and in mice with post-embryonic ablation of NPY/AgRP neurones.
Whilst this neuronal population is the most well-characterised ghrelin target, there is
also evidence for an indirect action on these neurons via the vagus nerve. Some authors
have found that surgical vagotomy or vagal deafferentation using capsaicin abolished
the feeding and Arc c-fos response to peripheral ghrelin. However, more recently it has
been reported that the orexigenic response to peripheral ghrelin is intact in rats follow-
ing subdiaphragmatic vagal deafferentation, suggesting that the acute orexigenic
actions of ghrelin do not require vagal afferent signalling. Other ghrelin targets include
several other hypothalamic nuclei, the dorsal vagal complex of the brainstem and com-
ponents of the mesolimbic dopaminergic system [14–16].
170 Wren
to physiological levels results in weight regain. However, mice with global deletions of
ghrelin or the GHS-R1a were initially reported to have minimal disruption of body
weight homeostasis. As always with such mouse models, one must consider con-
founding by developmental adaptation, and indeed the phenotype of conditional
knockout models is more robust. Further studies on mice lacking ghrelin or the GHS-
R1a have demonstrated resistance to diet-induced obesity in mature mice. The phe-
notype of ghrelin null mice might be further complicated by the observation that the
gene that codes for ghrelin has been found to code for another peptide, named
obestatin. Obestatin was originally reported to reduce food intake when administered
peripherally or ICV, and to reduce body weight gain when administered peripherally,
via the orphan G protein-coupled receptor, GPR39. However, subsequent reports
have not supported the initial findings and suggest that obestatin may not signal via
GPR39 or play a role in the regulation of food intake. Mice devoid of ghrelin sig-
nalling certainly lack the extreme phenotypes associated with mice lacking leptin sig-
nalling. However, taken together, data from knockout models are compatible with a
role for ghrelin in long-term energy homeostasis [15].
Given that circulating ghrelin is already low in obese subjects, one might question
how much therapeutic benefit could be obtained from further ghrelin suppression.
However, it has been shown that obese subjects may be more sensitive to appetite
stimulation by exogenous ghrelin [19]. Thus inhibition of ghrelin may have therapeu-
tic potential, particularly in enhancing further weight loss and preventing weight
regain following diet-induced weight loss, when ghrelin levels become elevated.
Several major pharmaceutical companies have pursued programmes investigating
ghrelin inhibition. Interestingly, the GHS-R1a exhibits constitutive activity, suggest-
ing that an inverse agonist may be more therapeutically useful than an antagonist.
Another strategy is to design compounds which bind to ghrelin itself and prevent
interaction with its receptor. A novel group of molecules called RNA spiegelmers,
oligonucleotides containing L-ribose, have been designed which block interaction of
ghrelin with the GHS-R1a, inhibit ghrelin-induced GH secretion and reduce food
intake and adiposity in mice fed a high-fat diet. As expected, these molecules have no
effect in ghrelin knockout mice, attesting to the specificity of their actions. To date, no
ghrelin-blocking products have progressed as far as phase I trials [14–16].
Theoretically, there may be safety concerns about ghrelin blocking agents in view
of the possible role of ghrelin in regulation of the growth axis, as well as reported ben-
eficial cardiovascular and anti-inflammatory effects of ghrelin. Regulation of
octanoylation of ghrelin may provide an alternative drug target which is, as yet, rela-
tively unexplored. A more direct therapeutic application of ghrelin is in the treatment
of anorexia and cachexia. To this end, proof of principle studies have demonstrated
Satiety Signals
After a meal, nutrients pass into the stomach and intestine, and a number of gastroin-
testinal signals are released. These peptides and other signals act to optimise the diges-
tive process. Some also function as short-term satiety signals and possibly long-term
regulators of body weight. CCK is the prototypical satiety hormone. It is now over 30
years since CCK was first shown to inhibit food intake and it remains one of the most
intensively studied of the gut hormones. CCK will not be discussed in depth and the
reader is directed to an excellent recent review [26]. However, the use of CCK as a
potential novel obesity therapy is in some doubt as, in animals, repeated pre-prandial
administration of CCK reduces food intake but also increases meal frequency, with no
net effect on body weight. Furthermore, continuous CCK administration is ineffective
after the first 24 h.
PYY, pancreatic polypeptide (PP) and NPY are members of the PP-fold peptide family,
i.e. they share a common tertiary structure, called the PP-fold structural motif. In addi-
tion there is significant homology between peptide sequences within the family. They all
have 36 amino acids, contain several tyrosine residues and require C-terminal amidation
for biological activity. PYY and PP are putative satiety signals. The PP-fold family exert
their effects via the Y family of G protein-coupled receptors. Four receptor subtypes have
been identified, Y1, Y2, Y4 and Y5, all of which are expressed in the hypothalamus. Y1 and
Y5 have both been put forward as the putative receptors via which NPY exerts its orexi-
genic action. The Y2 receptor is thought to function as an autoinhibitory pre-synaptic
172 Wren
receptor, expressed on NPY neurones, and to mediate the anorectic actions of PYY,
whilst the Y4 receptor appears to mediate the anorectic actions of PP [27–29]
Peptide YY
PYY occurs in two forms, PYY1–36 and PYY3–36. PYY3–36, the major circulating form, is
a truncated 34-amino acid form created by cleavage of the N terminal Tyr-Pro residues
by dipeptidyl peptidase IV (DPPIV). Whilst full length PYY binds with similar affinity
to all Y receptors, PYY3–36 shows selectivity for the Y2 receptor, for which it has high
affinity. PYY is secreted from entero-endocrine L-cells. These PYY immunoreactive
cells are found throughout the entire gastrointestinal tract, but particularly in the dis-
tal portion. PYY immunoreactivity is almost absent in the stomach, sparse in the duo-
denum and jejunum, common in the ileum and colon, and at very high levels in the
rectum (the opposite distribution pattern to that observed for ghrelin). The pattern of
secretion is also a mirror image of ghrelin, i.e. PYY is released into the circulation fol-
lowing meals and suppressed by fasting [27–29]. PYY has long been known to exert
numerous effects on the gastrointestinal tract. Administration of PYY increases the
absorption of fluids and electrolytes from the ileum after a meal and inhibits pancre-
atic and gastric secretions, gallbladder contraction and gastric emptying. Peripheral
administration of PYY, like ghrelin, also exerts effects on numerous other body sys-
tems. For example, it reduces cardiac output, causes vasoconstriction and reduction in
glomerular filtration rate, plasma renin and aldosterone activity. The physiological sig-
nificance of these numerous actions has not been established [27–29].
PYY levels rise to a plateau at 1–2 h post-prandially, with these peak levels influenced
by both the number of calories and the composition of the food consumed. The onset
of PYY release occurs before nutrients have reached the predominant sites of PYY
production in the distal gastrointestinal tract. This implies that peptide release may
occur via a neural reflex, possibly through the vagus nerve. Systemic administration
of PYY3–36 inhibits food intake in rodents and man. Initially these findings were con-
tentious, with several authors unable to reproduce feeding inhibition in rodents. A
probable explanation is that the effects of anorectic agents in rodents are easily
masked by stress, causing a reduction of food intake in the control group. Thus sig-
nificant inhibition of feeding by PYY3–36 cannot be detected in rodents that are not
fully acclimatised to experimental procedures or following transfer to a novel envi-
ronment [27–29]. In man intravenous infusion of PYY3–36 reduces appetite and food
intake at a subsequent meal by about one third. Whilst initial reports detected inhibi-
tion of food intake at plasma PYY concentrations in the physiological range, others
Chronic systemic administration of PYY3–36 results in weight loss in some but not all
rodent models. Data from knockout mice also provide evidence for a long-term role
for PYY in regulation of energy balance. Three separate knockout models have been
generated, two of which develop obesity. One model did not become obese. This
model also had disruption of expression of the PP gene. One would not predict that
loss of a second putative anorectic signal would attenuate obesity. However, subtle
differences in technique and background strain have frequently been reported to
result in differing phenotypes in other knockout mouse models. In one of the obese
PYY null models administration of PYY corrected the phenotype, suggesting that
obesity was being driven by PYY deficiency [28, 29]. In contrast to leptin, circulating
levels of PYY are not elevated in obese individuals who also retain full sensitivity to
the anorectic actions of PYY3–36. It has been reported that obese subjects have lower
fasting and post-prandial circulating PYY than lean subjects. In order to produce an
equivalent stimulation of PYY and equivalent satiety, obese individuals needed to
consume a much greater caloric load than their lean counterparts. However, not all
studies have detected a difference in fasting PYY concentrations between lean and
obese groups. Current data suggest that impaired post-prandial PYY release may, at
least, impair satiety and help to maintain obesity, if not act as a primary driver of ini-
tial development of obesity. Whether or not reduced PYY signalling is a primary
cause of obesity, it is certainly true that retained PYY sensitivity in the obese make it
an attractive therapeutic target [27–29].
The exact mechanism whereby PYY3–36 inhibits appetite and food intake is unclear.
Interestingly, in contrast to peripheral administration, intracerebroventricular admin-
istration of PYY stimulates food intake. This is thought to be via an action on Y1 and
Y5 receptors in the paraventricular nucleus (PVN), the second order neurones tar-
geted by orexigenic Arc NPY neurones. Several lines of investigation suggest a direct
anorectic action of circulating PYY3–36 on the Arc. C-fos is observed in the Arc in
response to peripheral administration of PYY3–36 and direct micro-injection into the
Arc inhibits feeding. This action is thought to be via auto-inhibitory Y2 receptors on
the orexigenic NPY neurones. In support of this hypothesis, a highly specific Y2
174 Wren
agonist inhibits feeding following intra-arcuate injection, whilst the anorectic actions
of PYY3–36 are absent in the Y2 knockout mouse and blocked by a Y2 receptor antago-
nist. Furthermore, PYY3–36 reduces expression of NPY in the Arc and release of NPY
from hypothalamic explants. Electrophysiological studies suggest that PYY3–36 directly
inhibits activity of arcuate NPY neurones, secondarily disinhibiting anorectic POMC
neurones. However, POMC does not appear to be essential for the anorectic action as
PYY3–36 is effective in mice with lacking melanocortin signalling. However, the picture
appears to be more complex than a simple action on the Arc. There is also evidence
that PYY3–36 acts at the level of the vagus and the dorsal vagal complex. The relative
contribution of these various putative sites of action to physiological appetite regula-
tion remains unclear. The ascending vagal-brainstem-hypothalamic pathways have,
however, been implicated in mediating sensations of nausea. In keeping with this
PYY3–36 has, at high doses, been reported to cause conditioned taste avoidance in
rodents and nausea in humans. However, lower doses inhibit appetite and food intake
in rodents and humans without aversive effects or nausea. Coupled with the observa-
tion that PYY null mice become obese, this suggests a role for PYY in appetite regula-
tion independent of aversive effects [27–29]. Drug companies developing analogues of
PYY for treatment of obesity will need to be mindful of the potentially narrow thera-
peutic window in order to design successful agents.
Pancreatic Polypeptide
PP is produced largely in the endocrine pancreas, but also in the exocrine pancreas,
colon, and rectum. Like PYY, PP is released in response to a meal and inhibits
appetite. PP binds with greatest affinity to Y4 receptors (with greater affinity than
PYY) and Y5 receptors. The role of PP in appetite regulation has been investigated
for over 30 years. It was initially noted that ob/ob mice lacked pancreatic PP cells,
and peripheral administration of PP reduced their food intake and body weight.
Peripheral administration of PP resulting in physiological plasma levels has
been shown to reduce food intake in normal mice, with associated reduction in
gastric emptying and gastric ghrelin expression and increased vagal tone. PP also
increased oxygen consumption and stimulated sympathetic activity, leading to the
suggestion that PP may also increase energy expenditure. In normal-weight human
volunteers infusion of PP reduces food intake without altering gastric emptying.
Subjects with PWS are reported to have suppressed basal and post-prandial PP lev-
els, whilst PP administration to PWS subjects reduces food intake. It is, therefore,
possible that PP deficiency contributes to the hyperphagia in this obesity syndrome
[28, 29].
Apart from its acute effects on appetite and food intake, PP may also modulate
long-term energy balance. Transgenic mice that overexpress PP have a lean pheno-
type with reduced food intake [30]. Repeated administration of PP to ob/ob mice
176 Wren
Oxyntomodulin
SP1
GRPP Glucagon
Gut
Brain
SP1
GRPP Glucagon
GLP-1 GLP-2
Glicentin
Preproglucagon
Signal
SP1
SP2
peptide
GRPP Glucagon GLP-1 GLP-2
Pancreas
MPGF
GRPP Glucagon GLP-1 SP2 GLP-2
Fig. 2. Overview of differential preproglucagon processing in pancreas versus brain and gut.
SP ⫽ Spacer peptide.
Gila monster saliva marketed as exenatide (Byetta), as well as orally active DPPIV
inhibitors. Central administration of GLP-1, both intracerebroventricularly and into
the PVN, reduces food intake in rodents, whilst the GLP-1 receptor antagonist
exendin 9–39 increases food intake. Chronic administration of GLP-1 into the CNS
attenuates weight gain and peripheral GLP-1 injection inhibits food intake in rodents
and man. Evidence suggests GLP-1 secretion is reduced in obese subjects, and weight
loss normalises the levels. Reduced GLP-1 secretion could, therefore, contribute to
obesity, and replacement may restore satiety. Obese subjects receiving subcutaneous
GLP-1 for 5 days, just before each meal, reduced their calorie intake by 15% and lost
0.5 kg in weight. However, in view of the powerful incretin action of GLP-1 there is a
risk of hypoglycaemia in non-diabetic subjects [28, 29].
Oxyntomodulin
It has been suggested that a cause of the current obesity epidemic may be that mod-
ern processed foods bypass our natural satiety mechanisms. Low fat diets are the
most well-established means of dietary weight loss. It has been reported that weight
loss in response to a low-fat diet does not produce the expected elevation in plasma
ghrelin. This may be due to an increase in the proportion of calories consumed as
carbohydrate that more potently suppresses ghrelin per calorie consumed than
does fat. High-protein diets have also become popular in recent years as a means to
178 Wren
promote satiety and weight loss. Diets high in protein have recently been reported
to elevate circulating PYY and enhance satiety more effectively that other macronu-
trients; however, previous data suggested that, at a single meal, higher plasma con-
centrations of PYY were stimulated by high fat isocaloric meals, compared with
protein or carbohydrate. It is an intriguing possibility that designer diets may help
promote the most favourable gut hormone profile to allow sustained weight loss
[15, 28, 29].
An Obesity Poly-Pill?
The only treatment to date associated with dramatic and sustained weight loss in the
morbidly obese is gastric bypass surgery. However, its cost and associated morbidity
and mortality make it an impractical treatment for the majority of obese patients
and it is generally reserved for the morbidly obese. Gastric bypass results in signifi-
cant increases in plasma PYY, GLP-1 and oxyntomodulin whilst ghrelin either falls
or fails to rise, despite significant weight loss [15, 28, 29]. Interestingly, bypass
patients report dramatically reduced hunger long before substantial weight loss
occurs. Furthermore, in rodent models many of the beneficial effects of bypass can
be mimicked by gut hormone administration [34]. It is notable that the changes in
the four gut hormones above all favour weight loss following gastric bypass. This co-
ordinated action, mimicking natural satiety, may be a key to effective anti-obesity
therapy. As noted above, individual gut hormones administered at high concentra-
tions, have been associated with aversive behaviours in rodents and nausea in
humans. We have reported that low doses of PYY3–36 and GLP-1 inhibit food intake
additively [35]. Analogous to current treatment for hypertension where several
agents are commonly used, a smart cocktail of gut hormone-based drugs may prove
a more effective anti-obesity treatment than targeting a single hormone. This
approach could potentially provide the sustained weight loss offered by gastric
bypass surgery, without the associated morbidity and mortality. The major thera-
peutic disadvantage of gut hormones is their short duration of action and the
requirement for subcutaneous or intravenous administration. In the GLP-1 system,
degradation-resistant analogues and drugs that inhibit enzymes that degrade the
endogenous hormone have already been brought to market for the treatment of type
2 diabetes. Similar approaches may be successful for oxyntomodulin, PYY and PP,
whilst intra-nasal delivery systems or development of orally active small molecule
mimetics could avoid the need for administration by injection.
The obesity epidemic is advancing relentlessly and current treatments, bar
bariatric surgery, are insufficiently effective. Recent data suggest that gut hormones
regulate when and how much we eat for every meal and offer a logical drug target.
Mimicking natural satiety mechanisms by delivering combinations of gut hormones
may replace bariatric surgery as the only truly effective anti-obesity treatment.
180 Wren
30 Ueno N, Inui A, Iwamoto M, Kaga T, Asakawa A, 34 Le Roux CW, Aylwin SJ, Batterham RL, Borg CM,
Okita M, Fujimiya M, Nakajima Y, Ohmoto Y, Ohnaka Coyle F, Prasad V, Shurey S, Ghatei MA, Patel AG,
M, Nakaya Y, Miyazaki JI, Kasuga M: Decreased food Bloom SR: Gut hormone profiles following bariatric
intake and body weight in pancreatic polypeptide- surgery favor an anorectic state, facilitate weight loss,
overexpressing mice. Gastroenterology 1999;117: and improve metabolic parameters. Ann Surg 2006;
1427–1432. 243:108–114.
31 Holst JJ: On the physiology of GIP and GLP-1. Horm 35 Neary NM, Small CJ, Druce MR, Park AJ, Ellis SM,
Metab Res 2004;36:747–754. Semjonous NM, Dakin CL, Filipsson K, Wang F,
32 Drucker DJ: The biology of incretin hormones. Cell Kent AS, Frost GS, Ghatei MA, Bloom SR:
Metab 2006;3:153–165. Peptide YY3–36 and glucagon-like peptide-17–36
33 Abbott CR, Monteiro M, Small CJ, Sajedi A, Smith inhibit food intake additively. Endocrinol 2005;146:
KL, Parkinson JR, Ghatei MA, Bloom SR: The 5120–5127.
inhibitory effects of peripheral administration of
peptide YY(3–36) and glucagon-like peptide-1 on
food intake are attenuated by ablation of the vagal-
brainstem-hypothalamic pathway. Brain Res 2005;
1044:127–131.
Adipokines in Obesity
Rexford S. Ahima ⭈ Suzette Y. Osei
Division of Endocrinology, Diabetes and Metabolism,
Department of Medicine, University of Pennsylvania School of Medicine,
Philadelphia, Pa., USA
Abstract
Adipose tissue is the source of soluble mediators (adipokines), secreted mainly by adipocytes. Leptin acts
on the brain and peripheral organs to regulate energy homeostasis and the neuroendocrine axis.
Adiponectin regulates glucose and lipid metabolism by targeting the liver and skeletal muscle. Adipose-
derived proinflammatory cytokines, vasoactive peptides, coagulation and complement factors, visfatin,
vaspin and retinol-binding protein signal through paracrine and hormonal mechanisms. Understanding
the biology of adipose tissue and the rapidly growing list of adipokines provides new insights into nor-
mal physiological regulation, as well as the pathogenesis and treatment of obesity, diabetes and disor-
ders of lipid metabolism and cardiovascular system. Copyright © 2008 S. Karger AG, Basel
The obesity epidemic has focused attention on the biology of adipose tissue.
Adipocytes provide a large energy storage capacity mainly in the form of triglyceride
[1]. The levels of insulin, glucose and nutrients increase during feeding and stimulate
energy storage in the liver and adipocytes [1]. Conversely, fasting activates the sym-
pathetic nervous system and increases the levels of glucagon, epinephrine and gluco-
corticoids, leading to enhancement of glycogenolysis and gluconeogenesis, and
maintenance of glucose supply to the brain and vital organs [1]. Prolonged fasting
also stimulates lipolysis, generating fatty acids to be used by muscle, liver and periph-
eral organs, in addition to providing ketones for the brain [1].
Interactions between adipose tissue and the vascular and immune systems are now
increasingly recognized [2, 3]. As the number and size of adipocytes expand in obe-
sity, the vascular supply increases under stimulation of angiogenic factors secreted by
adipocytes [2]. Conversely, antiangiogenic treatment reduces adipose vascularity and
prevents obesity in rodents [4, 5]. Obesity is associated with histological and bio-
chemical changes characteristic of inflammation [3]. Adipose tissue from obese indi-
viduals accumulates a large number of activated macrophages which form giant cells
and secrete cytokines, e.g. tumor necrosis factor-␣ (TNF-␣) and interleukin-6 (IL-6)
[3] (table 1). C-reactive protein is also increased in obesity, and the levels of intracel-
lular cell-adhesion molecule-1 and platelet/endothelial cell adhesion molecule-1
increase in adipose endothelial cells, inducing the migration and adhesion of mono-
cytes [3]. Monocyte chemoattractant protein-1 and various chemokines recruit
monocytes to adipose tissue [6, 7]. Collectively, these changes induce insulin resis-
tance in adipose and liver, and interact with adipose-derived coagulation factors and
vasoactive peptides, leading to cardiovascular disease in obesity [3].
This chapter will discuss the current understanding of leptin as a prototypic
adipokine related to energy balance and neuroendocrine regulation. Next, the role of
adiponectin in glucose and lipid metabolism and vascular biology will be reviewed.
Finally, putative roles of retinol-binding protein 4 (RBP4), resistin and visfatin and
various adipokines will be reviewed (table 1).
Leptin
The discovery of leptin more than a decade ago was a major turning point in our under-
standing of adipokines [8]. Mice and humans homozygous for the leptin gene mutation
develop hyperphagia, severe early-onset obesity, insulin resistance preceding obesity,
excess lipid accumulation outside adipose tissue (steatosis), and neuroendocrine abnor-
malities, notably, hypothalamic hypogonadism and tertiary hypothyroidism [9–11].
Moreover, there is evidence for immunosuppression in congenital leptin deficiency [10,
11]. Leptin is expressed and secreted mainly by adipocytes, but low levels are present in
the gastric fundus, mammary gland, placenta, pituitary and skeletal muscle [9].
Leptin has a relative molecular mass of 16 kDa and circulates in free or bound
forms. The latter represents leptin bound mainly to its soluble receptor and is thought
to be inactive. The concentration of leptin is higher in obese than lean individuals [9].
Leptin falls rapidly during fasting and increases gradually during feeding. Studies in
rodents and human indicate a link between these changes in leptin and insulin [9].
Higher leptin level in women is explained partly by increased production in subcuta-
neous adipose tissue and stimulation by estrogens. On the other hand, leptin is sup-
pressed by androgens in males [9]. Chronic glucocorticoid exposure, TNF-␣ and IL-6
increase leptin, while adrenergic stimulation decreases leptin [9]. Leptin exhibits a
diurnal rhythm, peaking at night in humans and in the morning in rodents [12, 13].
A pulsatile leptin rhythm has also been recognized in humans, although the underly-
ing mechanisms and functional significance are unknown [12, 14].
Five leptin receptor isoforms, LRa–LRe, are derived from alternate splicing of lepr
mRNA [15, 16]. LRa is the predominant ‘short leptin receptor’ which lacks the key
cytoplasmic domain required for signaling through the JAK/STAT (signal transduc-
tion and activators of transcription) pathway. LRa is abundantly present in brain cap-
illary endothelium and peripheral tissues, and is thought to mediate leptin transport.
LRe is shed into the circulation as a ‘soluble receptor’ and binds leptin. The ‘long lep-
tin receptor’ LRb mediates effects of leptin in the brain [15, 16]. Leptin crosses the
blood-brain-barrier via a saturable transport system, and acts directly on neurons in
the arcuate nucleus that express neuropeptide Y (NPY), agouti-related peptide
(AGRP), pro-opiomelanocortin (POMC) and cocaine-and amphetamine-regulated
transcript [9]. Binding of leptin to LRb leads to association with JAK2, autophospho-
rylation of JAK2, phosphorylation of Tyr985 and Tyr1138 on LRb and phosphoryla-
tion and activation of signal transducer of transcription-3 (STAT3), which acts as a
transcription factor to regulate neuropeptides and various leptin target genes [16].
LRb-phosphorylated Tyr985 recruits the tyrosine phosphatase SHP-2 (Src-homology
protein tyrosine phosphatase-2), and terminates leptin signaling via induction of sup-
pressor of cytokine signaling-3 (SOCS3) [16]. Rising leptin level suppresses NPY and
AGRP and increases ␣-melanocyte-stimulating hormone (␣-MSH), derived from
POMC. ␣-MSH inhibits feeding and induces thermogenesis through melanocortin 4
receptors in the paraventricular nucleus and other areas of the hypothalamus [14].
Normally, AGRP serves as an antagonist of ␣-MSH and stimulates feeding in concert
with NPY. Thus, by suppressing expression of AGRP and NPY, leptin has a net effect
to reduce appetite and decrease weight [9].
As predicted, deletion of LRb or STAT3 in the hypothalamus, or specifically in POMC
neurons, resulted in obesity [17–20]. In contrast, SOCS3 deficiency prevented obesity by
improving sensitivity to leptin [21–23]. Studies have also revealed a cross-talk between
Adiponectin
Resistin
Resistin has a relative mass of 12 kDa and belongs to a family of cysteine-rich C-termi-
nal domain proteins called resistin-like molecules [74, 75]. Initial studies demonstrated
that resistin was suppressed by TZDs and induced insulin resistance when administered
in rodents [74]. Multimeric complexes of resistin have been identified [76]. Each
resistin protomer consists of a C-terminal disulphide-rich -sandwich head and an N-
terminal ␣-helical tail. The latter associates with itself forming three-stranded coils.
Interchain disulphide linkages form tail-to-tail hexamers. Thus, resistin exists as hexa-
mers and trimers in mouse serum [76]. As predicted, the lack of resistin decreased glu-
cose and enhanced insulin sensitivity in mice [77]. Conversely, transgenic overexpression
Additional Adipokines
Conclusions
This review highlights the switch from the notion of adipose tissue as a dormant tissue
to one where adipose tissue actively regulates energy homeostasis and diverse systems
via adipokines. Key areas under investigation include the depot-specific functions of
adipose tissue, and how these relate to normal physiology and disease. Apart from
their classic roles as hormones, leptin and various adipokines have paracrine and
autocrine actions which may serve to modulate adipogenesis, nutrient fluxes and
metabolic changes locally and in adjacent organs. Knowledge of specific signaling
pathways will benefit the treatment of obesity and associated metabolic diseases. There
is no doubt that genetic technology in rodents has contributed immensely to the cur-
rent understanding of adipokines and their targets in the brain and peripheral organs.
In some cases, e.g. leptin and adiponectin, the biology of adipokines is similar between
mice and humans (table 1). However, there are numerous examples where putative
adipokines in rodents are derived from different sources and act differently in humans
(table 1). Thus, future research on adipokines demands a combination of cellular and
molecular approaches, rodent physiology and importantly human studies.
Acknowledgements
This work was supported by grant RO1-DK62348 and PO1-DK49210 from the National Institutes
of Health.
Abstract
AMP-activated protein kinase (AMPK) is a major regulator of energy metabolism at both the cell and at
the whole body level. Numerous genetic and obesity models as well as human studies have suggested a
role for AMPK in the physiological regulation of fatty acid and glucose metabolism, and in the regulation
of appetite. Changes in AMPK activity have been reported in obesity, type 2 diabetes, the metabolic syn-
drome and cardiovascular disease, which jointly represent a major health and economical problem
worldwide. Whether AMPK changes are one of the causes or the consequence of these pathological
conditions remains a matter of debate, but AMPK clearly represents a major potential pharmacological
target in the treatment of these conditions. Copyright © 2008 S. Karger AG, Basel
Obesity is a major health and economic problem in both Western and developing
societies. Its continuing rise in prevalence, 20% in England and 30% in USA [1, 2]
seems to be unstoppable despite multiple efforts to attempt to halt this trend. Obesity
is characterised by multiple metabolic changes such as insulin resistance, dyslipi-
daemia and hypertension. The diseases arising as a consequence of obesity such as
type 2 diabetes (T2D), cardiovascular disease and certain cancers, are increasingly
important causes of morbidity and mortality. In the last decades, a huge amount of
research has been dedicated to the study of the complex pathophysiology of obesity
and to the research for new medical therapies.
AMP-activated protein kinase (AMPK) has emerged in the last years as a major regu-
lator of cell and whole body metabolism. Numerous papers have reported evidence for
its role in the regulation of appetite, of body weight and of metabolism [3–5]. Therefore,
it is natural to consider AMPK as a major player in the development of obesity. The
AMPK complex is an evolutionally conserved serine/threonine heterotrimer kinase
complex consisting of ␣-, - and ␥-subunits [for detailed reviews see 5, 6]. AMPK is
activated by cellular stress, which depletes cellular ATP leading to a concomitant rise in
AMP. AMP activates AMPK by three distinct mechanisms: (a) allosteric activation, (b)
stimulation of phosphorylation of the ␣-subunit on Thr172 by upstream kinase(s)
[LKB1 and calmodulin kinase kinase-␣ or - and recently a new possible AMPK kinase
candidate, the transforming growth factor--activated kinase (TAK1), which phospho-
rylates AMPK on Thr-172 in HeLa cells [7], has been reported], and (c) inhibition of
dephosphorylation by protein phosphatases [5, 8–10]. Cellular stresses that cause a rise
in the AMP/ATP ratio include metabolic poisons (arsenite, oligomycin), oxidative
stresses, hypoxia, low glucose, muscle contraction and nutrient deprivation. Osmotic
stress also activates AMPK even without a change in the AMP/ATP ratio. Once acti-
vated, AMPK switches off anabolic pathways such as gluconeogenesis, glycogen, fatty
acid, triglyceride, cholesterol and protein synthesis (mTOR-p70SK-E2 pathway), and
switches on catabolic pathways such as glycolysis, glucose uptake, and fatty acid oxida-
tion. It also leads to mitochondrial biogenesis, which improves the ATP synthesis capac-
ity of the cell [11]. Metabolic changes induced by AMPK are both acute changes due to
phosphorylation of key enzymes and longer-term effects on the expression of genes
involved in metabolic regulation. AMPK, through several mediators, plays a role in vari-
ous physiological and pathological processes in different tissues (fig. 1). Therefore, it was
logical to hypothesise that abnormal AMPK activity would be present in conditions of
deregulated energy balance, such as obesity and T2D.
ACC
Skeletal muscle
PEPCK, G6Pase, GPAT, L-PK,
ChREBP, TORC2, HNF␣,
GLUT4
Fig. 1. Metabolic targets of AMPK in muscle, liver and adipose tissues. AMPK regulates the expres-
sion and phosphorylation of enzymes and genes involved in glucose and lipid metabolism.
GLUT4 ⫽ Glucose transporter 4; MEF-2 ⫽ myocyte enhancer factor-2; GEF ⫽ GLUT4 enhancer factor;
AS-160 ⫽ Akt-substrate-of-160 kDa; ACC ⫽ acetyl-coenzyme A carboxylase; PEPCK ⫽ phospho-
enolpyruvate carboxykinase; G6Pase ⫽ glucose-6-phosphatase; GPAT ⫽ glycerol-3-phosphate acyl-
transferase; L-PK ⫽ L-pyruvate kinase; ChREBP ⫽ carbohydrate response element-binding protein;
TORC2 ⫽ transducer of regulated CREB activity 2; HNF␣ ⫽ hepatic nuclear factor ␣; FAS ⫽ fatty
acid synthase; SREBP-1c ⫽ sterol regulatory element binding protein-1; HMGR ⫽ 3-hydroxy-3-
methylglutaryl-coenzyme A reductase; HSL ⫽ hormone-sensitive lipase.
only endurance exercise and not resistance exercise can induce AMPK activation
[20, 22]. AMPK activation in endurance exercise could also explain the lack of mus-
cle hypertrophy in distance running in contrast to weight lifting. This is possibly
due to the effect of AMPK on the mTOR pathway [20]. The mTOR pathway stimu-
lates protein synthesis and hence cell growth and hypertrophy in response to growth
factors and amino acids. Therefore, AMPK inhibition of this pathway would result in
inhibition of protein synthesis and lack of muscle hypertrophy. AMPK also stimulates
fatty acid oxidation in muscle. This results in lower lipid deposition and increases the
ability of the muscle to meet energy needs by increasing glucose uptake and fatty acid
oxidation as well. Studies with transgenic animals (AMPK ␣1 and ␣2 knockout mice,
muscle-specific over-expression of dominant negative AMPK ␣2, AMPK ␥3 knock-
out, muscle-specific over-expression of AMPK ␥3 and muscle-specific over-expression
of AMPK ␥3 R225Q overactive mutant and skeletal muscle-specific LKB1 knockout
[for detailed descriptions, see 20, 23]), have provided further evidence for AMPK
being the main mediator, although not the only one, of the adaptations (i.e. increased
PKB1 ACC
mTOR CPT1
Fig. 2. Regulation of hypothalamic AMPK and possible downstream pathways. Black lines show path-
ways established in the hypothalamus, grey lines show pathways that have been described in rat mus-
cle, rat liver, myotubes, hepatocytes, fibroblasts and lung carcinoma cells [97, 98] but not directly in the
hypothalamus.
Animal models of obesity and diabetes have provided evidence for implication of
AMPK in the pathogenesis of these conditions and also provided evidence for a pos-
sible role of AMPK modulators in their treatment (table 2).
High glucose ⇓[39, 60, 61] ⇓[36] ⇓[62] ⇓[36, 63, 64]
Insulin ⇓[39] ⇒[36] ⇒[50] ⇒[36] ⇓[30] ⇒[36] ⇓[65]
Leptin ⇓[39, 42, 66] ⇑[67, 68] ⇑[46, 50, 69, 70] ⇑[71] ⇒[63] ⇒[72, 73]
Ghrelin ⇑[66, 74] ⇒[74] ⇓[74, 75] ⇓[74] ⇑[74]
Adiponectin ⇑[76–78] ⇑[50, 76, 77] ⇑[33] ⇑[79] ⇑[80, 81]
Resistin ⇓[82] ⇓[83, 84]
Glucagon ⇑[85]
Cannabinoids ⇑[74] ⇒[74] ⇓[74] ⇓[74] ⇑[74]
Metformin ⇓[60] ⇑[86] ⇑[62, 86] ⇑[87] ⇑[37, 63] ⇑[88]
Rosiglitazone ⇑[54, 89, 90]
Martin et al. [43] showed that diet-induced obesity (DIO) in mice alters the effect
of leptin on AMPK activity both in skeletal muscle and in the hypothalamus. Leptin
increases AMPK activity in the skeletal muscle of chow-fed mice and decreases it in
the hypothalamus of the same animals but does not have an effect in the DIO mice.
While, most interestingly, a ciliary neurotrophic factor analogue (CNTFAx15) given
intracerebroventricularly not only reduces food intake in high-fat diet (HFD) mice
but also suppresses hypothalamic AMPK activity, bypassing therefore diet-induced
leptin resistance [44]. Rats on an HFD for 5 months exhibited decreased AMPK
phosphorylation and expression in skeletal muscle associated with decreased levels
of ACC and GLUT4 as well. Metformin treatment restored insulin sensitivity and
increased AMPK activity [45].
In Zucker rats who do not respond to leptin treatment because of defects in the
leptin receptor, administration of the AMPK activator AICAR results in leptino-
mimetic effects, leading to the prevention of ectopic lipid deposition and diabetes
[46]. Transgenic mice over-expressing leptin in liver are lean on a chow diet but
despite the high pre-existing leptin levels become obese and insulin resistant on an
HFD [47]. HFD for 15 weeks abolishes the increase in muscle AMPK activity
observed in the same animals on a chow diet.
Short hepatic over-expression of a constitutively active form of AMPK decreased
blood glucose levels in normal mouse, abolished hyperglycaemia in streptozotocin-
induced and in ob/ob mice and also reduced gluconeogenic enzyme expression. The
resulting low glucose levels led to a switch from glucose utilisation to fatty acid utili-
sation, associated with a decrease in white adipose tissue mass and development of
fatty liver [48].
Obese Zucker fa/fa rat AICAR increased muscle glucose transport and suppresses [91, 92]
endogenous glucose production and lipolysis
Reduced AMPK and ACC phosphorylation LKB1 activity [90]
and PGC-1 content
Rosiglitazone restores AMPK ␣2 activity in skeletal [46, 93]
muscle
Chronic AICAR/exercise training prevented
hyperglycaemia and increased whole-body
insulin sensitivity
ob/ob and db/db mice AICAR and short hepatic over-expression of a constitutively [48, 94]
active form of AMPK decreased blood glucose levels
HFD in rats Reduction of AMPK activity, ACC and GLUT4 levels in [45]
skeletal muscle. Metformin increases AMPK activity
Rosiglitazone enhanced AICAR-stimulated glucose [95]
uptake in muscle and adipose tissue. Total AMPK and
AMPK ␣2 activity increased in muscle
DIO mouse AICAR administration blocked weight gain, reduced total [96]
content epididymal fat and lipid accumulation in
adipocytes, restored adiponectin levels, improved glucose
tolerance and insulin sensitivity
DIO mice compared to chow-fed mice ate less, had lower [43]
respiratory exchange rate and lower ACC activity in
muscle. Leptin did not improve either of these
parameters or the AMPK ␣2 activity in muscle and
hypothalamus of the DIO
Ciliary neurotrophic factor analogue reduced food intake [44]
and AMPK hypothalamic activity, bypassing therefore
diet-induced leptin resistance
The majority of the research studies published have been performed on animals,
and it is important to establish that their conclusions can be extrapolated to human
physiology and pathology as the number of studies about AMPK activity in human
diseases is much more limited. Skeletal muscle AMPK activity has been analysed in
a limited number of obese vs. lean subjects, in obese diabetic versus obese non-
diabetic patients and in healthy subjects before and after exercise. Obesity in
humans is associated with leptin and insulin resistance and lipid accumulation.
Adiponectin or AICAR activate muscle AMPK in obese rodents, which stimulates
fatty acid oxidation, and it is reasonable therefore to hypothesise that pharmacologi-
cal activation of AMPK might be of therapeutic benefit in human obesity. However,
AMPK is not down-regulated in human skeletal muscle of obese females [52] and
AMPK activity and specific isoform expression are similar in muscle of obese sub-
jects with and without T2D [53]. These data suggest that impaired insulin action on
glycogen synthesis and lipid oxidation in skeletal muscle of these patients is
unlikely to involve changes in AMPK expression and activity. However, AICAR
treatment of muscle biopsies stimulated AMPK ␣2 activity and fatty acid oxidation,
suggesting that AMPK activation above basal levels may still be a valid therapeutic
approach [52]. In contrast to the previous studies, Bandyopadhyay et al. [54] showed
that there is a decrease in AMPK activity and an increase in ACC activity in insulin-
resistant muscle from obese and from T2D patients that results in elevated intracel-
lular levels of malonyl-CoA. Because, for the most part, the defects appear to be
expressed equally in the obese subjects and in T2D subjects (who were also obese),
the authors conclude that these differences from lean control subjects are caused by
insulin resistance/obesity rather than hyperglycaemia/diabetes. Finally, when the
T2D subjects were treated for 3 months with rosiglitazone, the various defects in
fatty acid and mitochondrial metabolism reverted towards normal. The beneficial
effect of AMPK activation in muscle was demonstrated in a study which showed
that acute intensive exercise (3 h) increased AMPK and ACC phosphorylation
altogether with an increase in expression of adiponectin receptor in the skeletal
muscle of 5 healthy females [55]. Interestingly, Roepstorff et al. [56] showed that
AMPK activation in muscle is sex-dependent: 90 min of exercise activated AMPK
in skeletal muscle of healthy male volunteers but in contrast to the former study,
not in females. Further data are needed to study the role of oestradiol on skeletal
muscle AMPK activity. The effect of exercise on AMPK is probably due, at least in
part, to IL-6, which is synthesised and released from skeletal muscle in large
amounts during exercise [57], and in rodents, the resultant increase in IL-6 concen-
tration correlates with increases in AMPK activity in multiple tissues. There are
no direct data of the effect of IL-6 on AMPK activity in humans but IL-6 treatment
was recently shown to enhance insulin-stimulated glucose disposal in humans
in vivo [58].
In conclusion, AMPK has emerged as a key regulatory enzyme of cell and whole body
metabolism. It influences cell metabolism in a way that favours insulin sensitivity and
maintains a favourable body energy homeostasis. It is the mediator of the metabolic
effects of many of the known hormones, nutrients and drugs. Thus, not only are
changes in AMPK implicated in the pathogenesis of insulin-resistant states, but
AMPK might also constitute a target for new treatments of these conditions.
However, a note of caution is required as generalised AMP activation might result in
unwanted effects (i.e. an appetite-stimulating effect and -cell inhibition), and thus
there is a need for tissue-specific modulators.
References
1 Grundy SM, Hansen B, Smith SC Jr, Cleeman JI, clues to modern understanding of metabolism. Cell
Kahn RA: Clinical management of metabolic syn- Metab 2005;1:15–25.
drome: report of the American Heart Association/ 7 Momcilovic M, Hong SP, Carlson M: Mammalian
National Heart, Lung, and Blood Institute/ American TAK1 activates Snf1 protein kinase in yeast and
Diabetes Association conference on scientific issues phosphorylates AMP-activated protein kinase in
related to management. Circulation 2004;109: vitro. J Biol Chem 2006;281:25336–25343.
551–556. 8 Woods A, Johnstone SR, Dickerson K, Leiper FC,
2 Marx J: Cellular warriors at the battle of the bulge. Fryer LG, Neumann D, Schlattner U, Wallimann T,
Science 2003;299:846–849. Carlson M, Carling D: LKB1 is the upstream kinase
3 Xue B, Kahn BB: AMPK integrates nutrient and hor- in the AMP-activated protein kinase cascade. Curr
monal signals to regulate food intake and energy bal- Biol 2003;13:2004–2008.
ance through effects in the hypothalamus and 9 Hawley SA, Pan DA, Mustard KJ, Ross L, Bain J,
peripheral tissues. J Physiol (Lond) 2006;574:73–83. Edelman AM, Frenguelli BG, Hardie DG: Calmo-
4 Carling D: The AMP-activated protein kinase cas- dulin-dependent protein kinase kinase-beta is an
cade–a unifying system for energy control. Trends alternative upstream kinase for AMP-activated pro-
Biochem Sci 2004;29:18–24. tein kinase. Cell Metab 2005;2:9–19.
5 Hardie DG, Hawley SA, Scott JW: AMP-activated 10 Woods A, Dickerson K, Heath R, Hong SP,
protein kinase–development of the energy sensor Momcilovic M, Johnstone SR, Carlson M, Carling D:
concept. J Physiol (Lond) 2006;574:7–15. Ca2⫹/calmodulin-dependent protein kinase kinase-
6 Kahn BB, Alquier T, Carling D, Hardie DG: AMP- beta acts upstream of AMP-activated protein kinase in
activated protein kinase: ancient energy gauge provides mammalian cells. Cell Metab 2005;2:21–33.
Abstract
Obesity is associated with several endocrine diseases, including common ones such as hypothyroidism
and polycystic ovarian syndrome to rare ones such as Cushing’s syndrome, central hypothyroidism and
hypothalamic disorders. The mechanisms for the development of obesity vary in according to the
endocrine condition. Hypothyroidism is associated with accumulation of hyaluronic acid within various
tissues, additional fluid retention due to reduced cardiac output and reduced thermogenesis. The patho-
physiology of obesity associated with polycystic ovarian syndrome remains complex as obesity itself
may simultaneously be the cause and the effect of the syndrome. Net excess of androgen appears to be
pivotal in the development of central obesity. In Cushing’s syndrome, an interaction with thyroid and
growth hormones plays an important role in addition to an increased adipocyte differentiation and adi-
pogenesis. This review also describes remaining rare cases: hypothalamic obesity due to central hypothy-
roidism and combined hormone deficiencies. Copyright © 2008 S. Karger AG, Basel
The heterogeneity of physical appearance in obesity and the association with various
endocrine disorders lead some clinicians and almost every patient with obesity to
believe that there is an underlying hormonal imbalance to account for their problem.
This is the reason that endocrinologists are frequently consulted to advise on the pos-
sible underlying diagnosis for secondary obesity. The term ‘secondary’ means that
obesity accompanies another illness that is considered to be the primary disease state.
Secondary morbid obesity (body mass index, BMI, ⱖ40) due to endocrine causes
is quite rare and is usually associated with hypothalamic disorders. However, lesser
degrees of a weight problem (BMI 26–39) are associated with thyroid disorders, poly-
cystic ovarian syndrome (PCOS) or Cushing’s syndrome.
An appropriate evaluation is the first step in developing a treatment plan for
overweight patients. A medical history should evaluate the natural history of the
development of obesity and aetiological factors involved. The physical examination
and laboratory testing should extend this evaluation to exclude likely endocrine
causes. The result of those then provide a guide to selecting an appropriate treatment plan.
The conditions associated with overweight to morbid obesity are discussed below,
in order of prevalence.
Hypothyroidism
particularly recent symptoms, should be tested with serum thyroid function tests [8].
The same group showed that change in symptoms is more powerful in predicting the
disease than current symptoms and the higher number of symptoms reported (cur-
rent as well as changed), the greater the likelihood of predicting hypothyroidism.
It is interesting that weight problem has not been even listed among the common
symptoms by Canaris et al. [8] but featured highly among symptoms and signs
analysed by Zulewski et al. [1].
Thus, overt hypothyroidism is a common condition that may be difficult to diag-
nose on clinical grounds alone and requires a high degree of clinical suspicion as well
as confirmation with biochemical testing for serum TSH and thyroid hormone levels.
Reliance on blood testing to diagnose an easily treatable condition has led to a
spectrum of results listed below (table 1).
Untreated hypothyroidism is well known to lead to hypercholesterolaemia, which
may improve or completely normalise on statin treatment. In cardiovascular patients
who do not reduce their cholesterol levels adequately on statin treatment, testing for
hypothyroidism is essential. Cardiovascular disease in hypothyroidism may have the
potential to be fatal, although this has never been tested in a controlled fashion.
Thyroid hormones are essential for the regulation of a number of important processes
in the body including growth and development, neuromuscular activity, thermogenesis,
energy consumption and many metabolic reactions [9]. The thyroxine (T4) pro-hormone
214 Weaver
is synthesised in the thyroid gland together with a small amount of the active hor-
mone 3,5,3⬘-L-triiodothyronine (T3). However, the majority of circulating T3 is gen-
erated by pre-receptor ligand metabolism resulting from activity of the iodothyronine
deiodinase enzymes D1 and D2 which convert T4 to T3 by 5⬘ monodeiodination.
In contrast, the D3 enzyme inactivates T4 and T3 by inner ring or 5-deiodination.
The intracellular level of T3 is dependent on the relative activities of these three deio-
dinases [10].
T3 regulates nuclear gene expression by binding the thyroid hormone receptors TR␣
and TR (genomic action). Thyroid hormone receptors recognise specific thyroid
response elements in the promoters of T3-target genes and activate or repress tran-
scription in response to hormone. The effect of genomic action of T3 can vary as from
hours to days [9]. Some of the T3 effects occur rapidly and are unaffected by inhibitors
of transcription and protein synthesis. Those actions are defined as non-genomic [11].
Growth Hormone
Lack of thyroid hormones leads to reduced growth hormone secretion and thus gen-
eration of IGF-I [16].
Leptin
Leptin and thyroid hormones possess the ability to increase energy expenditure.
Hypothyroidism has been shown to be associated with raised leptin levels [17].
Furthermore, thyroxine treatment reduced leptin secretion independently of adiposity
and noradrenaline levels. It appears that thyroxine and leptin are closely regulated by a
negative feedback mechanism. Leptin has been shown to stimulate the hypothalamo-
pituitary-thyroid axis and also to modulate 5⬘-deiodinases in different tissues, depending
on the energy status of animals. Leptin promoted a rise in serum TSH, suggesting that
leptin acts at the hypothalamus in order to stimulate the thyrotropin-releasing hormone
(TRH)-TSH axis. Leptin may alter thus hypothalamic, pituitary and brown adipose tis-
sue functions by regulation of 5⬘-deiodinases and directing the local T3 production [18].
In animals, hypothyroidism is characterised by decreased insulin responsiveness,
partly mediated by an exaggerated glucose-fatty acid cycle that is partly alleviated by
intracerebroventricular leptin administration [19].
Catecholamines
The hypothyroid state is characterised by decreased adrenergic activity due to
reduced responsiveness of cAMP to adrenaline. This may be related to the effect of
thyroid hormones on cAMP generation [20]. In vivo, thyroxine regulates thermogen-
esis and the lipolytic activities of catecholamines within 30 min [21]. However, there
is also evidence for a direct effect of thyroid hormone on mitochondrial gene expres-
sion and oxidative phosphorylation [22]. Thyroid hormones accumulate in mito-
chondria [23] and are major regulators of mitochondrial biogenesis playing a role in
proliferation, differentiation and maturation [24]. The synergistic actions of cate-
cholamines and thyroid hormone together result in a threefold increase in mitochon-
drial UCP1 levels in brown adipose tissue [25]. Catecholamines also rapidly increase
D2 expression in brown adipose tissue, leading to tissue-specific hyperthyroidism
and a twofold increase in plasma T3. Uniquely in brown adipose tissue, D2 induction
is potentiated by T3, whereas in other tissues T3 suppresses D2 [10]. In addition,
216 Weaver
thyroid hormones have been shown to increase expression of the non-BAT UCP2,
UCP3 and the ADP/ATP carrier suggesting a regulatory role in mitochondria from
other tissues [26]. Thyroid hormones also regulate mitochondrial gene expression by
increasing steady state mitochondrial mRNA levels, respiration, enzyme activity and
protein synthesis [27], and by increasing mitochondrial metabolic activity [28, 29].
The majority of currently presenting cases of PCOS are characterised by obesity (BMI
⬎27) with a central fat distribution. As this type of overweight is particularly associ-
ated with increased risk of cardiovascular disease and type 2 diabetes, obesity is often
defined as a waist circumference of more than 80 cm for women in classifications that
serve the diagnosis of the metabolic syndrome. In a large case series studies from
Pittsburg, the waist/hip ratio was associated with PCOS independently of BMI [48].
The insulin resistance seen in PCOS is partly determined by the presence of central
obesity, as when PCOS women are compared with control women matched for
abdominal adiposity, the difference in insulin resistance between the two groups is
much less marked than if the two groups are matched for BMI [49].
218 Weaver
The possible causal relation for central fat deposition may exist between excessive
testosterone concentrations in early development. This relation exists in women
during adulthood as seen in non-obese female to male transsexuals [50] and obese
postmenopausal women [51]. In addition, hyperinsulinaemia, through its direct
effect on the adipocyte, has also been suggested to be a possible determinant of
android fat deposition in women with PCOS [52]. In women with PCOS, it is possi-
ble that android fat deposition per se contributes to hyperandrogenaemia through its
adverse effects on insulin sensitivity and consequent gonadotrophic effects of hyperin-
sulinaemia on the ovaries. Thus, in women with PCOS, android fat may be both a
cause and an effect of hyperandrogenaemia [53]. The vicious circle in which android
fat begets android fat, and further exacerbates the predisposition towards weight gain
in women with PCOS has been suggested. The cycle can be interrupted by dietary
intervention and/or use of insulin-sensitising drugs.
Simple obesity and obesity associated with PCOS is associated with increasing
androgenicity due to generation of: (1) testosterone from androstendione mediated
by 17-hydroxysteroid dehydrogenase [54]; (2) dihydrotestosterone from testos-
terone mediated by 5␣ reductase [55]; (3) reduced generation of oestradiol from
testosterone, oestrone from androstenedione and oestriol from 16␣-hydroxylated
dehydroepiandrosterone due to reduced aromatase activity [56].
Both simple obesity and PCOS-associated obesity are characterised by the mecha-
nisms of limiting exposure to cortisol by reducing cortisol generation from cortisone.
Two key enzymes involved in the metabolism of cortisol include 11-hydroxysteroid
dehydrogenase type 1 (11-HSD1), involved in the conversion of cortisone to corti-
sol, and 5␣-reductase, involved in the catabolism of cortisol in addition to the con-
version of testosterone to dihydrotestosterone. It has been demonstrated that in
normal men and women or hypopituitary patients on fixed hydrocortisone replace-
ment therapy, obesity is associated with impairment of 11-HSD1 reductase and
enhanced 5␣-reductase activities [57, 58]. It has also been shown that 11-HSD1
activity is strongly related to fat distribution [57, 58]. The changes in 11-HSD and
5␣-reductase activity are even more accentuated when studied in PCOS as compared
to BMI-matched controls [59].
Leptin plays an important role in the regulation of appetite, body weight and metab-
olism and reproductive capacity in women, buts its role in PCOS is controversial. It
has been proposed that abnormally high serum concentrations of serum leptin may
Cushing’s Syndrome
Obese patients may present an opportunity for an early diagnosis of Cushing’s syn-
drome. This requires a high level of clinical suspicion as patient may not display all the
typical symptoms or signs. Harvey Cushing first described in 1932 a constellation of
symptoms of obesity, hirsutism and amenorrhoea attributed to primary pituitary
abnormality associated with adrenal hyperplasia. Adrenal tumours causing the same
syndrome were later described, but Cushing’s syndrome due to ectopic adrenocorti-
cotrophin hormone (ACTH) was only recognised in 1962 [69]. The term Cushing’s
syndrome describes all causes of cortisol excess, whereas Cushing’s disease one type due
to pituitary-dependent disease. In Cushing’s syndrome, patients usually present with
insidious onset of weight gain with central fat deposition. In addition, patients develop
a ‘buffalo hump’ due to fat accumulation over the thoracocervical spine. Typical ‘moon
shape facies’ are produced by fat depots over cheeks and temporal regions.
The most discriminating symptoms and signs of Cushing’s syndrome are bruising,
muscle weakness, skin thinning, plethora and truncal obesity. The associated condi-
tions include diabetes mellitus, hypertension, hirsutism, osteoporosis and impaired
glucose tolerance test [70].
220 Weaver
Table 2. Causes of Cushing’s syndrome
ACTH dependent
Cushing’s disease (pituitary dependent)
Ectopic ACTH syndrome
Ectopic CRH syndrome
Macronodular adrenal hyperplasia
Iatrogenic treatment with ACTH 1-24
ACTH independent
Adrenal adenoma and carcinoma
Primary pigmented adrenal nodular hyperplasia and Carney’s syndrome
McCune-Albright syndrome
Aberrant receptor expression (gastric inhibitory polypeptide, interleukin-1)
Iatrogenic Cushing (pharmacological doses of steroids)
Pseudo-Cushing’s syndrome
Alcoholism
Depression
Pseudo-Cushing’s Syndrome
This is a condition characterised by some if not all clinical features of Cushing’s syn-
drome. There may be evidence of an intermittent rise of cortisol. Chronic alcoholism
is the commonest cause of pseudo-Cushing’s syndrome; however, the aetiology is not
very clear. Studies showed that in chronic liver disease there is an impairment of cor-
tisol metabolism. Furthermore, alcohol and/or vasopressin raised in patients with
alcohol excess were found to be a stimulant for cortisol secretion [71].
Many patients with Cushing’s syndrome are depressed [72]. On the other hand,
depression may display some of the features of Cushing’s syndrome which resolve on
222 Weaver
Table 3. Diagnostic tests used in investigations of Cushing’s syndrome
Application Sensitivity/specificity %
Sensitivity and specificity values calculated from data provided in Stewart [87]. F ⫽ Cortisol;
CT ⫽ computed tomography; MRI ⫽ magnetic resonance imagining; dex ⫽ dexamethasone.
1
Ninety percent suppression of urinary basal cortisol.
2
Hundred percent rise in ACTH and 50% rise in cortisol excludes ectopic ACTH.
Indirect Mechanisms
Effect on Thyroid Status. Glucocorticoids suppress thyroid axis through suppression
of TSH secretion and additionally inhibiting 5⬘deiodination mediating the conver-
sion of thyroxine to triiodothyronine.
Effect on Growth Hormone Status. Chronic hypercortisolism leads to a blunting
response of GH secretion to all stimuli [78].
Successful treatment of Cushing’s syndrome and a reduction of glucocorticoid
levels lead to remarkable weight loss, although often not back to the pre-morbid state.
Hypothalamic Obesity
Acquired obesity not present from infancy, coupled with headache, a growth disorder
and other growth dysfunction, requires investigations for hypothalamic/pituitary disease.
Selective pituitary failure can arise from the deficiency of one or more hypothala-
mic hormones, including TRH and growth hormone-releasing hormone. The reason
for obesity associated with TRH deficiency has been discussed above under the sec-
tion ‘Hypothyroidism’.
224 Weaver
involved in the pathogenesis of the disease. Finally, an association between central
hypothyroidism and severe obesity has been described in patients with leptin recep-
tor mutations. The clinical consequences of central hypothyroidism in adult life vary
greatly depending on the aetiology, the severity of the thyroid impairment, the extent
of the associated hormone deficiencies, and the age of the patient at the time of the
onset of the disease. In general, acquired central hypothyroidism is less severe than
the congenital form because of the constitutive activity of the wild-type TSH receptor.
Symptoms and signs of thyroid insufficiency are usually milder than those of primary
hypothyroidism, and goitre is always absent [86].
One can conclude that obesity due to endocrine causes is not as common as per-
ceived by patients and some physicians. However, since the effective treatment of less
common underlying endocrine causes is available, the clinical management of
patients with obesity should include appropriate screening for endocrine conditions
which may be amenable to treatment.
References
1 Zulewski H, Muller B, Exer P, Miserez AR, Staub JJ: 8 Canaris GJ, Steiner JF, Ridgway EC: Do traditional
Estimation of tissue hypothyroidism by a new clini- symptoms of hypothyroidism correlate with biochem-
cal score: evaluation of patients with various grades ical disease? J Gen Intern Med 1997;12: 544–550.
of hypothyroidism and controls. J Clin Endocrinol 9 Yen PM: Physiological and molecular basis of thyroid
Metab 1997;82:771–776. hormone action. Physiol Rev 2001;81:1097–1142.
2 Gavin LA: The diagnostic dilemmas of hyperthyrox- 10 Bianco AC, Salvatore D, Gereben B, Berry MJ, Larsen
inemia and hypothyroxinemia. Adv Intern Med PR: Biochemistry, cellular and molecular biology,
1988;33:185–203. and physiological roles of the iodothyronine selen-
3 Tachman ML, Guthrie GP Jr: Hypothyroidism: diver- odeiodinases. Endocr Rev 2002;23:38–89.
sity of presentation. Endocr Rev 1984;5:456–465. 11 Davis PJ, Davis FB, Lawrence WD: Thyroid hormone
4 Schectman JM, Kallenberg GA, Shumacher RJ, regulation of membrane Ca2(⫹)-ATPase activity.
Hirsch RP: Yield of hypothyroidism in symptomatic Endocr Res 1989;15:651–682.
primary care patients. Arch Intern Med 1989;149: 12 Smith TJ, Bahn RS, Gorman CA: Connective tissue,
861–864. glycosaminoglycans, and diseases of the thyroid.
5 Singer PA, Cooper DS, Levy EG, et al: Treatment guide- Endocr Rev 1989;10:366–391.
lines for patients with hyperthyroidism and hypothy- 13 Schmidt BM, Martin N, Georgens AC, et al: Non-
roidism. Standards of Care Committee, American genomic cardiovascular effects of triiodothyronine in
Thyroid Association. JAMA 1995;273: 808–812. euthyroid male volunteers. J Clin Endocrinol Metab
6 Helfand M: Screening for subclinical thyroid dys- 2002;87:1681–1686.
function in non pregnant adults: a summary of the 14 Davis PJ, Davis FB: Nongenomic actions of thyroid
evidence for the U.S. Preventive Services Task Force. hormone on the heart. Thyroid 2002;12:459–466.
Ann Intern Med 2004;140:128–141. 15 Wang YG, Dedkova EN, Fiening JP, Ojamaa K,
7 Vanderpump MP, Ahlquist JA, Franklyn JA, Clayton Blatter LA, Lipsius SL: Acute exposure to thyroid
RN: Consensus statement for good practice and audit hormone increases Na⫹ current and intracellular
measures in the management of hypothyroidism and Ca2⫹ in cat atrial myocytes. J Physiol 2003;546:
hyperthyroidism. The Research Unit of the Royal 491–499.
College of Physicians of London, the Endocrinology 16 Valcavi R, Jordan V, Dieguez C, et al: Growth hormone
and Diabetes Committee of the Royal College of responses to GRF 1–29 in patients with primary
Physicians of London, and the Society for Endocri- hypothyroidism before and during replacement therapy
nology. BMJ 1996;313:539–544. with thyroxine. Clin Endocrinol (Oxf) 1986;24:693–698.
226 Weaver
44 Balen AH, Conway GS, Kaltsas G, et al: Polycystic 57 Andrew R, Phillips DI, Walker BR: Obesity and gen-
ovary syndrome: the spectrum of the disorder in der influence cortisol secretion and metabolism in
1741 patients. Hum Reprod 1995;10:2107–2111. man. J Clin Endocrinol Metab 1998;83:1806–1809.
45 Taponen S, Martikainen H, Jarvelin MR, et al: 58 Weaver JU, Taylor NF, Monson JP, Wood PJ, Kelly
Hormonal profile of women with self-reported WF: Sexual dimorphism in 11 beta hydroxysteroid
symptoms of oligomenorrhea and/or hirsutism: dehydrogenase activity and its relation to fat distrib-
Northern Finland birth cohort 1966 study. J Clin ution and insulin sensitivity: A study in hypopitu-
Endocrinol Metab 2003;88:141–147. itary subjects. Clin Endocrinol (Oxf) 1998;49: 13–20.
46 Kiddy DS, Hamilton-Fairley D, Bush A, et al: 59 Rodin A, Thakkar H, Taylor N, Clayton R: Hyper-
Improvement in endocrine and ovarian function androgenism in polycystic ovary syndrome: evidence
during dietary treatment of obese women with poly- of dysregulation of 11 beta-hydroxysteroid dehydro-
cystic ovary syndrome. Clin Endocrinol (Oxf) genase. N Engl J Med 1994;330:460–465.
1992;36:105–111. 60 Brzechffa PR, Jakimiuk AJ, Agarwal SK, Weitsman
47 Dunaif A, Segal KR, Futterweit W, Dobrjansky A: SR, Buyalos RP, Magoffin DA: Serum immunoreac-
Profound peripheral insulin resistance, independent tive leptin concentrations in women with polycystic
of obesity, in polycystic ovary syndrome. Diabetes ovary syndrome. J Clin Endocrinol Metab 1996;81:
1989;38:1165–1174. 4166–4169.
48 Talbott E, Guzick D, Clerici A, et al: Coronary heart 61 Zachow RJ, Magoffin DA: Direct intra-ovarian
disease risk factors in women with polycystic ovary effects of leptin: impairment of the synergistic action
syndrome. Arterioscler Thromb Vasc Biol 1995;15: of insulin-like growth factor-I on follicle-stimulating
821–826. hormone-dependent estradiol-17 beta production by
49 Holte J, Bergh T, Berne C, Berglund L, Lithell H: rat ovarian granulosa cells. Endocrinology 1997;138:
Enhanced early insulin response to glucose in rela- 847–850.
tion to insulin resistance in women with polycystic 62 Maffei M, Halaas J, Ravussin E, et al: Leptin levels in
ovary syndrome and normal glucose tolerance. J Clin human and rodent: measurement of plasma leptin
Endocrinol Metab 1994;78:1052–1058. and ob RNA in obese and weight-reduced subjects.
50 Elbers JM, Asscheman H, Seidell JC, Megens JA, Nat Med 1995;1:1155–1161.
Gooren LJ: Long-term testosterone administration 63 Pirwany IR, Fleming R, Sattar N, Greer IA, Wallace
increases visceral fat in female to male transsexuals. J AM: Circulating leptin concentrations and ovarian
Clin Endocrinol Metab 1997;82:2044–2047. function in polycystic ovary syndrome. Eur J Endo-
51 Lovejoy JC, Bray GA, Bourgeois MO, et al: crinol 2001;145:289–294.
Exogenous androgens influence body composition 64 Wren AM, Seal LJ, Cohen MA, et al: Ghrelin
and regional body fat distribution in obese enhances appetite and increases food intake in
postmenopausal women – a clinical research cen- humans. J Clin Endocrinol Metab 2001;86:5992.
ter study. J Clin Endocrinol Metab 1996;81: 65 Pagotto U, Gambineri A, Vicennati V, Heiman ML,
2198–2203. Tschop M, Pasquali R: Plasma ghrelin, obesity, and
52 Pasquali R, Casimirri F, Balestra V, et al: The relative the polycystic ovary syndrome: correlation with
contribution of androgens and insulin in determin- insulin resistance and androgen levels. J Clin
ing abdominal body fat distribution in pre- Endocrinol Metab 2002;87:5625–5629.
menopausal women. J Endocrinol Invest 1991;14: 66 Schofl C, Horn R, Schill T, Schlosser HW, Muller MJ,
839–846. Brabant G: Circulating ghrelin levels in patients with
53 Barber TM, McCarthy MI, Wass JA, Franks S: polycystic ovary syndrome. J Clin Endocrinol Metab
Obesity and polycystic ovary syndrome. Clin 2002;87:4607–4610.
Endocrinol (Oxf) 2006;65:137–145. 67 Gambineri A, Pagotto U, Tschop M, et al: Anti-
54 Quinkler M, Sinha B, Tomlinson JW, Bujalska IJ, androgen treatment increases circulating ghrelin lev-
Stewart PM, Arlt W: Androgen generation in adipose els in obese women with polycystic ovary syndrome.
tissue in women with simple obesity–a site-specific J Endocrinol Invest 2003;26:629–634.
role for 17beta-hydroxysteroid dehydrogenase type 68 Moran LJ, Noakes M, Clifton PM, et al: Ghrelin and
5. J Endocrinol 2004;183:331–342. measures of satiety are altered in polycystic ovary
55 Stewart PM, Shackleton CH, Beastall GH, Edwards syndrome but not differentially affected by diet
CR: 5 alpha-reductase activity in polycystic ovary composition. J Clin Endocrinol Metab 2004;89:
syndrome. Lancet 1990;335:431–433. 3337–3344.
56 Meinhardt U, Mullis PE: The aromatase cytochrome 69 Meador CK, Liddle GW, Island DP, et al: Cause of
P-450 and it’s clinical impact. Horm Res 2002;57: Cushing’s syndrome in patients with tumors arising
145–152. from ‘nonendocrine’ tissue. J Clin Endocrinol Metab
1962;22:693–703.
228 Weaver
Korbonits M (ed): Obesity and Metabolism.
Front Horm Res. Basel, Karger, 2008, vol 36, pp 229–259
Abstract
The relentless rise in the prevalence of obesity predicts an exponential increase in the incidence of
obesity-related complications. Medical and surgical treatments are necessary to prevent and treat obese
co-morbidities, thereby avoiding disability and premature death. Interventions for obesity should be
evaluated not by weight loss alone but against the new incidence in obesity-related co-morbidities, their
remission or improvement. In combination with lifestyle measures, currently available pharmacological
therapies – rimonabant, orlistat and sibutramine – achieve 5–10% weight loss, although a return to base-
line is the norm after cessation of medication. All these agents demonstrate approximately 0.5% reduc-
tion in HbA1c in diabetic subjects; orlistat also reduces the new incidence of type 2 diabetes. Modest
improvement in lipid profiles and reduced calculated cardiovascular risk is observed, but data on
improvement of other co-morbidities are sparse. In contrast, surgical procedures that restrict food inges-
tion and/or curtail the absorptive surface area of the gut consistently achieve substantial weight loss,
typically 20–35%, effect resolution of co-morbid conditions and improve quality of life. Although mortal-
ity is low, complications and hospitalisation are not uncommon after bariatric surgery. Intriguingly, surgi-
cal patients experience a reduction in appetite and report changes in food preference. Accentuation of
the normal gastrointestinal hormonal response to food intake and possible changes in vagal afferent sig-
nalling are proposed to induce satiety. Increased understanding of body weight homeostasis and
appetite regulation has provided an impressive list of potential targets for drug development, with the
promise that single or combination therapy may ultimately challenge the supremacy of bariatric surgery.
Copyright © 2008 S. Karger AG, Basel
The rising prevalence of obesity has attracted the attention of the medical and scientific
communities, editors of highbrow and popular lay press, broadcasters, intergovern-
mental agencies and political bodies. Whilst headlines, opinions and initiatives abound,
solutions at a population level remain elusive. Far from reversing the trend, the number
of WHO-defined obese subjects with a body mass index (BMI) ⬎ 30 grows at an ever-
increasing rate. In 2001 20.9% of the US population were obese (BMI ⬎ 30) and 2.3%
had a BMI ⬎ 40. Concerning as these statistics may be, the enormity of the problem
may be underestimated since the most pronounced trends in body weight are seen in
children, adolescents and young adults [1]. Although the degree to which prepubertal
obesity tracks to adult disease remains controversial, it is likely that by the age of 11 obe-
sity is established [1]. Moreover, recent data from a large prospective study demonstrate
that weight gain during early adulthood confers a particularly high risk of developing
type 2 diabetes (T2D) [2]. The more costly obesity-related complications evolve slowly,
becoming more evident and demanding more resource with increasing age and there-
fore the health burden is set to rise even if the prevalence of obesity were to stabilise [3].
Management of Obesity
Management of obesity in its broadest sense can be summarised in four steps: Step 1:
Prevention; Step 2: Weight management; Step 3: Prevention of complications; Step 4:
Management of complications.
Step 1: Prevention
Prevention of obesity is the public health holy grail. It is self-evident that energy and
resource need to be dedicated to what is rapidly becoming the most pressing public
health issue worldwide. In a BMJ review, Anjali Jain articulated the problem facing pub-
lic health policy makers: ‘The high rates of obesity do not allow us to wait for treatments
to be proved effective by the standards of evidence-based medicine. In other words,
something must be done soon, but we don’t know what.’ [4]. Public health measures are
both aimed to reduce the new incidence of obesity, and to limit the pace of weight gain
amongst overweight and obese individuals. Strictly speaking, there are therefore both
preventative and weight management (step 2) aspects to a program devised for the gen-
eral population. Furthermore, the health benefits of both weight loss through lifestyle
intervention, and altered diet without weight loss are disproportionately beneficial, and
therefore public health measures may also be considered as a component of step 3 –
limiting the development of complications [5]. Success might be defined as by achiev-
ing a fitter, if not thinner population, and it remains to be seen whether the governmen-
tal rhetoric will be followed by a proliferation in public facilities for exercise, cycle lanes
and so forth. However, the evidence base for effective large-scale interventions is lim-
ited [6, 7], and this area is largely outside the scope of this review.
Behind the statistics, however, a growing number of individuals are experiencing the
medical complications, the reduced function and quality of life, and also the social and
Most of the medical cost of obesity is attributable to the management of the compli-
cations, and in this aspect the evidence base for specific complications is broad. It
should be emphasised that the ‘complications’ of obesity include the loss of individual
income, increase in social security benefits and also the secondary consequences if a
severely obese individual ceases to provide a carer role through their own disability.
Paradoxically, the management of the complications of obesity – particularly in rela-
tion to prevention of the complications of T2D – may in fact involve sacrificing the
primary goal of weight loss to address the greater evil of hyperglycaemia.
Analysis of preventative measures against obesity, behavioural and dietary inter-
ventions and the management of established complications are outside the scope of
this chapter [for review, see 10]. We will focus on the treatment of the obese adult
individual. A central theme will also be to evaluate the effectiveness of interventions
not merely in terms of weight but with equal emphasis on the improvement and
The Role of Medical and Surgical Intervention in the Management of Obese States
The medical complications attributable to total fat mass (category A) are principally
those that place medical strain on the musculoskeletal system and cause morbidity by
pain or disability. In addition, the psychosocial factors including body image dysphoria,
relationship difficulties, poor workplace achievement and economic handicap through
the requirement for assistance in activities of daily living appear to be related to the
degree of total obesity, although in one study, abdominal circumference also contributed
to disability [11]. Certain other conditions occur too infrequently (e.g. idiopathic
intracranial hypertension, IIH, obesity-related cardiomyopathy, glomerulopathy) to
clearly establish their aetiology in relation to total or visceral fat. The procoagulant ten-
dency that remains a risk factor for venous thromboembolism and arterial occlusion
may be related both to visceral and total adiposity. In general, the use of BMI and WHO
definitions stratifies the risk of these conditions effectively and they tend to occur at
higher degrees of obesity. BMI is therefore a reasonable surrogate marker of risk.
BMI in isolation is less convincing as the best marker for risks associated with visceral
obesity, and other co-morbidities in category B. Although the gold standard for the
Finally, in category C there are a number of conditions where obesity compounds the
morbidity of medical diseases, notably reducing the functional capacity of patients
with angina, congestive cardiac failure and asthma where modest weight loss may be
beneficial [20]. Intriguingly, there are circumstances where obesity confers a statistical
advantage: obese patients on dialysis programmes appear to have reduced mortality.
D Diabetes Normal
E Economic No effects
complications
IGT ⫽ impaired glucose tolerance; GDM ⫽ gestational diabetes mellitus; w.c. ⫽ waist circumference;
and obesity merely in terms of BMI, and evaluating intervention on the basis of
change in weight alone is inadequate in categorizing individual risk and treatment
response. In brief, although we may have an increased repertoire of agents to deploy,
it will become ever more important to define those where greatest impact can be
achieved.
Osteoarthritis T2D
Disability Dyslipidaemia
OSA
Poor economic status Hypertension
thrombosis
Body image dysphoria PCOS
Obstetric complications NASH
Fig. 1. Complications of obesity. Complications of obesity can be segregated into those that relate to
total fat mass (A), visceral fat (B), and those that represent an exacerbating effect of obesity on other
co-existing medical conditions (C).
Proof beyond doubt that severe obesity can arise from organic disease is evident in
those with structural hypothalamic disease, in whom hyperphagia and severe weight
gain can occur. The recognition of a series of monogenic disorders involving appetite
regulatory pathways serves to further emphasise that excess body weight is not
merely founded in a lack of conscious restraint [22]. The astonishing reduction in
obesity demonstrated with leptin treatment for the handful of individuals with
monogenic leptin deficiency provides powerful proof of principle that biological
processes underpinning obesity can be manipulated [23]. For the vast majority of obese
individuals, however, the pathogenesis of obesity reflects a multi-factorial individual
predisposition coming into collision with an increasingly unsuitable environment,
where genetic and epigenetic factors, intrauterine adaptation, early life feeding,
learned attitudes to eating and social customs define the at-risk individual [8]. Unlike
the correction of congenital leptin deficiency, and like a number of other disease
states, monotherapy is insufficient in a polygenic disorder, and the nihilism that sur-
rounds the medical management of obesity is in part due to the fact that with only a
few agents, effective combination therapy has not yet been achieved.
The first prerequisite for any drug is that it should not cause harm, the second is
that it should have some effect, and the third that it should confer some meaningful
Orlistat
Arguably the only widely used obesity therapy to have been specifically designed for
its purpose, orlistat is an intestinal lipase inhibitor that limits the amount of dietary
fat that can be converted into an absorbable form as a mono- or di-glyceride. Orlistat
has been shown to have a significant benefit over placebo and after 12 months,
patients treated with orlistat in addition to lifestyle intervention achieved a weight
reduction of 8.4% compared to 2.6% with placebo [25]. Later studies confirmed an
effect of orlistat superior to placebo, but only a third of patients achieved a 5% weight
loss at 12 months [26]. Once the benefits of intensive dietary support are removed in
a primary care setting, the use of orlistat with brief counselling effected a 4.8-kg
weight loss at 6 months, compared to 3.8 kg with drug alone and 1.7 kg with brief
counselling alone; although no significant differences were present at 12 months [27].
However, if we can ignore the statistical violation of continuation on the basis of early
response, results are more interesting, and patients randomised to either 500- or
1,000-kcal deficit diets, after initial success of 5% weight loss at 3 and 6 months,
achieved weight loss of over 10 kg at 1 year [28]. In a similar study, 3-month weight
loss of ⬎5% also predicted 11.9% loss at 2 years [29]. In the absence of any a priori
determinant of response, weight loss at 3 months appears to be of benefit in identify-
ing those patients who might achieve meaningful change in the longer term.
Sibutramine
0 2 4 6 8 10 12 14 16 18 20 22 24
a Month
Year 2
Year 1 Placebo/placebo
20 mg of rimonabant/placebo
Placebo 20 mg of rimonabant/20 mg
5 mg of rimonabant of rimonabant
20 mg of rimonabant
0 0
⫺2 ⫺2
⫺4 ⫺4
⫺6 ⫺6
⫺8 ⫺8
⫺10 ⫺10
0 12 24 36 52 52 60 68 76 84 92 104
b Weeks Weeks
Patients
Placebo/placebo 292 284 254 247 233 222 216
Patients 20 mg of rimonabant/ 323 315 301 277 261 240 225
Placebo 590 496 413 347 309 placebo
5 mg of rimonabant 1191 1004 833 711 619 20 mg of rimonabant/ 328 319 297 286 272 263 256
20 mg of rimonabant 1189 1017 863 750 672 20 mg of rimonabant
Fig. 2. Recidivism after pharmacological weight loss. Following weight loss achieved through
pharmacological intervention, cessation of therapy is followed by a return to baseline weight.
a Sibutramine: following induction of weight loss with hypocaloric diet and sibutramine, subjects
were randomised to sibutramine or placebo [30]. b Rimonabant: weight loss induced by rimonabant
(left panel) is superior to placebo. After 12 months, re-randomisation to active compound or placebo
(right panel) demonstrates the effect of rimonabant on weight loss maintenance [99].
Rimonabant
Metformin
Although never proposed as a ‘diet pill’, metformin has long been recognised as hav-
ing at least a neutral effect on weight in comparison to sulphonylureas in the manage-
ment of T2D. Amongst obese diabetic subjects in the United Kingdom Prospective
Diabetes Study, metformin was the only treatment that was not associated with
weight gain. Although a neutral effect on weight is hardly headline-grabbing, this is
undoubtedly beneficial when seen in comparison to sulphonylurea and insulin therapy.
Metformin was associated with a significant reduction in any diabetic-related event,
diabetic-related death and all cause mortality and 39% reduction in myocardial
infarction [35]. Metformin has also been proven of value in the prevention of T2D: in
the Diabetes Prevention Trial metformin reduced the new incidence of diabetes and
was associated with greater weight loss (mean 2.1 kg) compared to placebo. Although,
first and foremost a treatment for T2D, conceptually at least, metformin has a broad
Surgical Treatment
Defining Success
The surgical literature has a lexicon that is different from that used by the rest of the
medical community. Obesity is typically defined as excess weight in relation to the
actuarial concept of ideal body weight at a BMI of 22. Hence, an individual with a
BMI of 33 has 50% ‘excess weight’, and one of 44, 100%. Whereas pharmacological
treatments might consider the proportion of patients losing 5 or 10% total weight as a
useful outcome measure, surgical reviews consider a weight loss of less than 10% as
being failure and 20% as being poor.
Surgical procedures (fig. 3) for obesity aim either to restrict the amount of foods that
can be ingested, or to reduce the effective surface area of the gut available for absorption,
and are referred to as ‘restrictive’ and ‘malabsorptive’, respectively. A further category of
procedures combines both of these elements, and is sometimes referred to as ‘hybrid’.
Although the first recognised surgical procedure for obesity was the jejuno-ileal bypass
a b c
Restrictive procedures
d e f
Fig. 3. Bariatric surgical procedures. a–c Malabsorptive bariatric procedures. a JIB. b BPD. c DS.
d–f Restrictive bariatric procedures. d VBG. e LAGB. f RYGB. Reproduced with permission from the
Endocrine Society [100].
After the disappointment and bad press that followed the JIB (see the section below),
focus shifted to restriction of the gastric pouch rather than risking the long-term
effects of iatrogenic malabsorption. A variety of techniques emerged, most notably
the vertical banded gastroplasty (VBG), or ‘stomach stapling’ in the vernacular. This
The original prototype malabsorptive procedure, the JIB was first pioneered in the
1950s. In this operation, the jejunum is divided and the proximal limb anastomosed
directly to the last portion of the ileum (fig. 3). Pancreatic and biliary secretions remain
in the newly fashioned tract, but a long segment of ileum remained as a blind loop. The
JIB was undoubtedly effective, but whilst many patients benefited, the side effects even-
tually led to its becoming discredited and abandoned. In addition to protein-calorie mal-
nutrition and fat malabsorption with its attendant vitamin deficiencies, increased oxalate
levels led to a polyarthritis, crystal nephropathy and poorly understood cirrhosis.
Newer techniques combine an element of malabsorption and restriction. The
Scopinaro procedure, or bilio-pancreatic diversion (BPD) and duodenal switch (DS)
are the modern descendants of purely malabsorptive operations, although they may
be additionally combined with a reduction in gastric volume. In both procedures, the
alimentary limb remains short, but the pancreatic and biliary secretions drain into
the redundant ileal limb. Diversion of the secretions has two effects: firstly to reduce
the mixing of food with digestive juices, and secondly to avoid the consequences of a
blind loop of intestine. Whilst treading a fine line between adequate weight loss and
clinical malnutrition – the surgical equivalent of a narrow therapeutic window – per-
formed by experienced practitioners these procedures are remarkably effective,
achieving a 40% weight loss in a typical bariatric patient.
Currently, the most widely used hybrid procedure is the RYGB. In this technique,
the jejunum is divided and the distal limb anastomosed to the stomach remnant of
As with any surgical procedure, the risks of the operation include those of any general
anaesthetic, and early and late consequences of the specific operation in question.
Evidently, obese subjects pose certain generic anaesthetic problems with airway
management and a higher risk of thrombotic complications. LAGB involves the
introduction of a foreign body into the peritoneum, and as well as the potential for
infection (rare) the band may on occasion lead to erosion through the upper stomach.
The most feared early complication of other surgical procedures is anastomotic leakage,
which may be difficult to diagnose promptly due to the difficulty in clinical examination
of the obese abdomen. Later complications include excessive capacity for ingestion,
Type 2 Diabetes
In addition to their effects on weight loss, currently available weight loss treatments
are effective in improving glycaemic control amongst patients with T2D. The Xendos
trial demonstrated a reduction in the new incidence of T2D amongst orlistat-treated
obese patients treated with impaired or normal fasting glucose. The RIO-Diabetes
study demonstrated approximately 0.5% reduction in HbA1c in patients treated with
rimonabant. Similar improvements in glycaemic control have been observed with
sibutramine [47].
Metabolic Syndrome
Dyslipidaemia improves with both sibutramine and orlistat, whereas blood pressure
falls amongst patients treated with orlistat. Rimonabant has beneficial effects on
components of the MetS. Consistent with the total weight loss, visceral adipose fat
mass has been shown to be reduced after bariatric surgery [50]. Improvements in
components of the MetS have also been observed consistently following bariatric
surgery, although they are less spectacular than the effects on diabetes. In the SOS
study, patients with hypertriglyceridaemia improved after surgical treatment com-
pared to controls, although there were only trends towards improvement in hyperc-
holesterolaemia at 10 years and no significant improvement in hypertension. In a
randomised controlled clinical trial comparing the effect of treating mild to moderate
obesity with LAGB to intensive medical program for 24 months there was a signifi-
cant improvement of the MetS [44]. Although potentially beneficial, these differences
fall short of the effects of specific lipid-lowering medications.
Hepato-Biliary Disease
Obstructive sleep apnoea (OSA) is a common complication of obesity, and its inci-
dence may be underestimated. In a recent open-labelled, uncontrolled cohort study,
10% weight loss with sibutramine was associated with a reduction in the severity of
OSA, with improvement in the Respiratory Disturbance Index and Epworth score (a
measure of sleepiness) over a 6-month period [53]. Most of the available data show a
reduction in the severity of OSA after bariatric surgery; in a prospective study, Dixon
et al. [54] showed improvement in the apnoea/hypopnoea index, Epworth score and
the need for assisted ventilation after LAGB.
Arguably, the success of bariatric surgery owes more to serendipity than to science.
Drug discovery proceeds with an orderly and regulated progression through phase I
to phase V clinical trials. In contrast, surgical procedures evolve, building on the suc-
cess and adverse events of earlier procedures, and the mechanisms of weight loss are
somewhat obscure. Two observations have defined recent advances in this area: (1) a
reduction in appetite accompanies weight loss and (2) resolution of T2D appears to
occur prior to significant weight loss.
The reduction in appetite after bariatric surgery has been long recognised but not
extensively studied. In a prospective study design, our group demonstrated that fol-
lowing RYGB, the sensation of fullness and a reduction in hunger was evident imme-
diately after a test meal and was maintained for 180 min. This effect was observed at
1 month and persisted at least to 6 months, despite a reduction in circulating leptin and
insulin and an increase in ghrelin that would be expected to increase hunger [65]. We
further demonstrated that the reduction in appetite was accompanied by an increase
in the negative experience of eating, which can be interpreted as either a conscious
adaptation to a smaller stomach or evidence for a restoration of negative feedback [64].
As early as 1982, researchers began to examine a possible humoral effect of the then
popular JIB. In a rarely cited paper, Atkinson and Brent demonstrated in a rodent
model of the JIB that serum derived from rats undergoing JIB had anorectic properties
when injected into control animals [66]. With the advent of radioimmunoassay for gut
peptides, patients who had undergone JIB were found to have elevated levels of peptide
YY (PYY) and enteroglucagons – a composite of products of the pre-pro-glucagon
gene. These observations became more significant once interest was reignited with the
growing body of evidence that gut peptides might contribute to appetite regulation. In
brief, a number of peptide hormones: peptide YY, glucagon-like peptide 1 (GLP-1),
pancreatic polypeptide and oxyntomodulin (OXM) have been shown to have
inhibitory effects on food intake in human and rodent studies, whereas the acylated
peptide hormone ghrelin is unique in stimulating food intake and appetite [67].
Shortly after the isolation of ghrelin from the stomach and the demonstration of its
orexigenic effects on meal size in man and adiposity in rodents, Cummings et al. [68]
showed that patients that had undergone RYBG had completely suppressed ghrelin
levels. This study provided the provocative notion that alterations in the levels of a gut
hormone might mediate the effects of malabsorptive bariatric surgery. A flotilla of
similar studies followed that had conflicting results, with ghrelin levels being
unchanged, decreased or even increased after RYGB. Variability of assays, surgical
techniques and experimental protocols may all contribute to the variability in results,
20
10
0
⫺30 0 30 60 90 120 150 180
Time point (min)
Fig. 4. PYY increase after bariatric surgery. PYY response to a test meal amongst obese and lean
controls, and patients undergoing RYGB or LAGB, demonstrating the exaggerated PYY response
after RYGB [74].
but overall it is unlikely that changes in circulating ghrelin represent a major contrib-
utor to the appetite-reducing effects of bariatric surgery [69].
Greater impetus to the effect of gut hormones on appetite regulation arose when
Batterham et al. [70] showed that peptide YY3–36 (PYY) limited meal size in lean
human volunteers and obese subjects. PYY null mice develop central adiposity that
may be reversed with exogenous treatment. Of particular interest, obese subjects are
not resistant to the effects of PYY, but exhibit a relative deficit that requires a larger
meal size to achieve an equivalent PYY response. Our group demonstrated that fol-
lowing RYBG, formerly obese patients had a greatly amplified PYY response to a
small test meal (fig. 4) [71]. This effect was not observed in patients who had lost sim-
ilar weight after LAGB, demonstrating that the effect was not simply a consequence of
weight loss. We further demonstrated in a prospectively designed study, a progressive
rise in PYY after RYGB over 6 months [65], maintained at least until 24 months
[unpubl. data]. Other groups have confirmed these results in both cross-sectional and
prospective designs with remarkable consistency [72].
The gut hormone response to a meal is not restricted to PYY and ghrelin. GLP-1,
better known for its incretin effects (see below), also has an effect on satiety. GLP-1
and other members of the enteroglucagon family are released in excess after RYBG,
and we have shown that enteroglucagon and GLP-1 are released in increased amounts
after RYGB [65]. See Aylwin [69] for a comprehensive review of the effects of bariatric
surgery on gut hormones.
Further work from Le Roux and colleagues [unpubl. data] has shown that suppres-
sion of the post-prandial PYY response with a somatostatin analogue increases meal size
and reduces satiety. Whilst these data do not specifically implicate PYY in mediating the
Debate continues as to whether the improvement or even cure of T2D after bariatric
surgery arises as a consequence of weight loss alone or due to some ‘magic’ from the
procedure. In an animal model it was shown that the bypass of the duodenum and
jejunum could improve post-prandial glucose excursion in rats prone to diabetes,
independent of an effect on weight loss [73]. In human subjects undergoing BPD,
recovery of first-phase insulin release in patients with T2D became progressively
apparent over a 12-month period: impressive evidence that the pathological hallmark
of T2D could be reversed.
Potential mechanisms for these effects have been proposed to involve an increase
in the islet cell trophic incretin hormones, GLP-1 and gastric insulinotrophic peptide,
thought to be respectively either reduced or less effective in T2D. Both RYGB and
BPD increase GLP-1 secretion both during fasting and after meal, a consequence of
L-cell stimulation by early arrival of nutrients in the distal ileum [65, 74]. The secre-
tion of GLP-1 promotes insulin release, influences glucose metabolism by inhibiting
glucagon secretion, delaying gastric emptying and stimulating glycogenesis and
GLP-1 has also been shown to inhibit pancreatic islet cell apoptosis. The potential for
stimulation of insulin release after RYGB has been highlighted by the development of
post-prandial hypoglycaemia amongst formerly obese subjects who had previously
undergone RYGB and the development of nesidioblastosis has been observed [75].
The unwanted hypoglycaemia may be an unusual side effect of RYGB but it is strong
circumstantial evidence that RYGB somehow induces an exaggerated insulin
response that might be mediated by increased release of GLP-1 [74].
In many cases, agents have been developed for other indications, and an effect on
body weight observed as a side effect during phase II/III development. It is worth
noting that both sibutramine and rimonabant owe their current indications to a sim-
ilar provenance. The following are examples of each class and the available data on
their proposed effect in the management of obesity. For an exhaustive account of
agents under evaluation, see Foster-Schubert et al. [76].
The discovery of leptin initiated an avalanche of research into appetite regulation that
proceeded to define the interface of peripheral hormones and the hypothalamus, out-
lined in the chapter by Wren [this vol., pp. 165–181]. In brief, the system includes an
orexigenic (EAT!) pathway initiated in a low-insulin/low-leptin state, by activation of
NPY and AgRP expressing hypothalamic neurones, and an opposing anorexigenic
(DON’T EAT!) pathway via leptin’s stimulation of POMC expression acting at the
melanocortin receptor MC4R. Experimental mouse models demonstrate that both
sides of the system are required for normal functioning, and human subjects with
congenital mutations [22] testify to the importance of this system in extreme mono-
genic obesity. To what extent appetite regulatory mechanisms are relevant to more
common forms of human obesity remains to be established.
Leptin
To the far-from satiated pharmaceutical industry, the discovery of leptin – a hormone
produced by adipocytes that circulated in proportion to fat mass, crossed the blood
brain barrier and influenced satiety [83] – seemed almost too good to be true.
Although a handful of individuals have been found to have congenital leptin defi-
ciency, in the vast majority of obese subjects leptin levels are increased. The higher
level of leptin suggests either (1) a state of relative leptin resistance, where the anorex-
igenic effect of rising leptin concentration eventually overcomes the refractory cen-
tral sensor and reaches a steady state set at a high body weight; or (2) that leptin
is merely an anti-starvation hormone, relevant only in deficiency where low levels
promote an overwhelming drive to eat, but higher concentrations have no effect
beyond a threshold level.
In the first randomised double-blind dose-ranging controlled trial with recombi-
nant leptin, obese subjects lost up to 7.1 kg after 24 weeks at the highest dose
compared to 0.7 kg for placebo [84]. The effect, however, was non-specific, being
dependent on doses that achieved levels 20- to 40-fold higher than at baseline. In a
further study using a pegylated form of leptin in which levels of total serum leptin
were increased twofold, in 30 obese men over a 12 week period, there were no differ-
ences on weight loss, percent body fat or respiratory quotient between pegylated leptin
and placebo. Leptin treatment has, however, been shown to undermine the homeostatic
NPY/AgRP
The roles of NPY in the regulation of food intake and energy expenditure are highly
complex, due to the number of receptors (Y1–Y6) and their widespread expression. A
Y5R antagonist was demonstrated to suppress weight gain in rodents but in a large
double-blind study of 1,661 subjects, the Y5R antagonist MK-0557 had no clinically
relevant effect [87].
Whilst leptin, and to an extent insulin, provide information about the state of avail-
able energy stores, a parallel system exists to inform the organism about the nutri-
tional value of recently ingested meals, and a growing number of peptides have been
identified that are released from the gastrointestinal tract in response to food intake.
Hormones released in the periphery that have an influence on central appetite regu-
lation make obvious and attractive targets for the treatment of obesity.
Amylin (Pramlintide)
Amylin is a peptide secreted by the pancreas in response to nutrients and other
insulinogenic stimuli; pramlintide is a subcutaneously administered synthetic ana-
logue of amylin. Amongst severely obese subjects (BMI ⬎ 40), pramlintide resulted
in significant reduction in weight (⫺3.2 kg placebo corrected) at 26 weeks with
Cholecystokinin
Cholecystokinin (CCK) was the first gut peptide recognised to have anorexigenic
effects. CCK is released from the proximal small bowel into the circulation and its
effects on gall bladder contraction are well established. CCK also acts at as an agonist
at the CCKA receptors (alternately named CCK1R) present on peripheral vagal affer-
ents. Peripheral CCK administration has been shown to increase satiety and reduce
food intake acutely in man [89], although the effects of CCK are short-term and
chronic administration has not been shown to have a sustained effect [90].
Ghrelin
In contrast to other gastrointestinal hormones, ghrelin remains the only known circu-
lating orexigenic peptide. Ghrelin was initially identified as the ligand for the formerly
orphan G-protein-coupled growth hormone secretagogue GHSR1a receptor, but its
principal effects are to co-ordinate a protective response to energy depletion both
through peripheral effects on intermediary metabolism and a central effect on feeding,
augmenting calorie intake in man [91]. Ghrelin antagonism therefore represents an
obvious potential strategy in the management of obesity. Proof-of-principle for the
validity of ghrelin antagonism has been demonstrated with the use of polyethylene gly-
col stabilised RNA oligonucleotides (Spiegelmers) that have been demonstrated to
inhibit the orexigenic properties of ghrelin [92]. The oligonucleotide NOX-B11-2 had
caused reduced intake and weight loss in diet-induced obese mice with chronic treat-
ment [92]. Whether such agents can be deployed in human studies is far from clear,
but it is likely that other antagonists of ghrelin will be actively pursued.
Oxyntomodulin
OXM is a member of the enteroglucagon family of polypeptides, which includes
glucagon and glucagon-like peptides GLP-1 and GLP-2, products of the pre-
proglucagon gene. OXM is part of a co-ordinated ileal satiety response and its actions
are similar to PYY and GLP-1, and it is also released from the same cells [93]. OXM,
given pre-prandially by subcutaneous injection three times a day in a parallel blinded
study design, resulted in a mean weight loss of 2.3 kg over 4 weeks compared to 0.5 kg
in the placebo group [94]. The precise mechanism of action of OXM is not known,
but its effects can be inhibited by a central GLP-1 antagonist and abolished in a GLP-1
receptor knockout model, suggesting that it acts via GLP1 receptors. The development
GLP-1 Agonists
The development of the long-acting GLP-1 analogue exendin-4 (exenatide) represents
the first incursion of a gut hormone into the mainstream pharmacopoeia. Exenatide is
resistant to degradation by dipeptidyl peptidase IV, has a longer half-life than the
native peptide and is now established as a therapy for T2D. In addition to its effects on
glycaemic control, exenatide has the additional advantage of weight loss, mean 4.0 kg
at 1.5 years [95]. Future studies are awaited in normoglycaemic obese human subjects.
Peptide YY
PYY was originally identified as modulator of gastrointestinal motility, released from
the L-cells of the small bowel and may have a role in delaying gut transit in malab-
sorptive conditions. The principal active form of PYY, PYY3–36 acts at the inhibitory
Y2 (NPY) receptor expressed by NPY neurons in the arcuate nucleus and has been
shown to act as an inhibitor of food intake. Recent data with an intranasal formula-
tion, however, have failed to demonstrate weight loss over a 12-week period [96].
Perspective
Public health initiatives, research into the aetiology of obesity, drug development and
establishing an increased capacity for bariatric surgery all have a role to play in the
future handling of obesity, but probably not in that order. It is becoming obvious that
increasing energy supply is closely associated with the increase of overweight and
obesity in western countries [97], and simply praying for a plateau rather than expect-
ing a reversal of the underlying trend may be more realistic.
The explosive pace of research into the adipose-hypothalamic and entero-
hypothalamic axes that have emerged as the principal homeostatic co-ordinators of
body weight will continue to offer opportunities for drug discovery. In this arena,
combination treatments have been slow to emerge, although it should be noted that
the combination of fenfluramine with fluoxetine has been shown recently to be
almost as potent as the now discredited fenfluramine-phentermine combination [98].
Medical therapy for obesity is very much in its infancy in comparison to other
chronic disorders, but it would be a fair bet to predict that treating individuals with
severe obesity will more resemble multi-drug treatment of hypertension rather than
the one-size-fits-all approach to hyperlipidaemia.
Management of the obese individual must extend beyond the boundaries of weight
management, and interventions need to be judged by their effects on co-morbidities.
References
1 Wardle J, et al: Development of adiposity in adoles- 4 Jain A: Treating obesity in individuals and popula-
cence: five year longitudinal study of an ethnically tions. BMJ 2005;331:1387–1390.
and socioeconomically diverse sample of young peo- 5 McQuigg M, et al: Empowering primary care to
ple in Britain. BMJ 2006;332:1130–1135. tackle the obesity epidemic: the Counterweight
2 Schienkiewitz A, et al: Body mass index history and Programme. Eur J Clin Nutr 2005;59(suppl 1):
risk of type 2 diabetes: results from the European Pros- S93–S100; discussion S101.
pective Investigation into Cancer and Nutrition (EPIC)- 6 Sahota P, et al: Randomised controlled trial of pri-
Potsdam Study. Am J Clin Nutr 2006;84: 427–433. mary school based intervention to reduce risk factors
3 Gunderson EP, et al: Excess gains in weight and waist for obesity. BMJ 2001;323:1029.
circumference associated with childbearing: The 7 Moore H, et al: Improving management of obesity
Coronary Artery Risk Development in Young Adults in primary care: cluster randomised trial. BMJ 2003;
Study (CARDIA). Int J Obes Relat Metab Disord 327:1085.
2004;28:525–535.
Abstract
The current obesity epidemic is largely driven by environmental factors, including nutritional transition
towards refined and fatty foods with the growing production of energy-dense food at relatively low cost,
increased access to motor vehicles, mechanisation of work and sedentary lifestyles. These influences in
modern society are modified by individual characteristics. Ultimately, energy intake in excess of caloric
expenditure causes obesity, but why this occurs in some but not all individuals is not known. Obesity is
more prevalent in the lower socioeconomic classes but even so, there is a varying relation of socioeco-
nomic status with obesity between countries at different stages of development and, even in the
Western world, socioeconomic gradients with respect to obesity are both heterogeneous and in transi-
tion. Potential mechanisms for an effect of obesity on subsequent social status have been proposed, the
most obvious being related to the stigmatisation experienced by the obese. Obesity seems to be
causally related to mood disturbances, whereas there is no conclusive evidence that the reverse is true.
When considering psychological aspects of obesity, depressive symptoms are more likely to be conse-
quences, rather than causes of obesity. Copyright © 2008 S. Karger AG, Basel
While there is, by now, little doubt that the current obesity epidemic is largely driven by
environmental factors, it is equally clear that these influences in modern society are
modified by individual characteristics. There is a growing literature on sociological and
psychosocial factors in relation to obesity. In this chapter, we describe psychosocial cor-
relates of obesity from a societal, as well as from an individual, perspective.
Societal Factors
pandemic was first apparent in the US, subsequently spreading to the rest of the west-
ernised world. With the urban transition in the developing countries in the world,
obesity is now rapidly increasing worldwide. Many developing countries are currently
affected by high rates of overweight that often surpass underweight as a public health
nutrition problem. In the case of urban Africa, recent analyses of national data on
body mass index (BMI) from women showed that prevalence of BMI ⱖ25 exceeded
that of BMI ⬍18.5 in 17 of 19 countries [1]. By 2020, two thirds of the global burden
of disease is estimated to be attributable to chronic non-communicable diseases, most
of them strongly associated with overnutrition [2]. The causes of the obesity epidemic
are multifactorial, including nutritional transition towards refined and fatty foods
with increased production of energy-dense food at relatively low cost, access to
motorised transport, mechanisation of work, and sedentary lifestyles. This increas-
ingly ‘obesogenic’ environment, reinforced by the cultural changes associated with
globalisation and urbanisation, makes the adoption of healthy lifestyles difficult. At
the same time undernutrition, while an uncommon problem in high- and middle-
income countries, still prevails in large segments of the population in low-income
countries.
On a societal level, decreasing daily physical activity is one of the important dri-
ving forces behind the obesity epidemic (table 1). Separation of work, housing and
shopping, often over long distances, necessitates the use of motorised transporta-
tion. Contemporary city structures strive to accommodate cars, buses and trams, but
not walking or cycling. Within the work environment, increased automation has led
to a lower workload with less physical exhaustion and with lower caloric expendi-
ture. During leisure, substantial portions of time are taken up by watching television
and using the computer. All these separate influences add up in expending less
energy, with obvious consequences for the net energy balance. It has been estimated
that a positive daily energy imbalance of 100 kcal/day results in slightly less than 1 kg
of weight gain per year, which is close to the actual weight gain observed in some US
populations [3]. The steady decline in physical activity is probably a major contribu-
tor to the rising prevalence of obesity. Small changes in behaviour, such as 30 min
per day of brisk walking or eating less at each meal, might contribute to arresting the
obesity epidemic [4]. Using terminology from a classic epidemiological triad (host,
vector, environment), it has been proposed that the contributions of physical inac-
tivity to the obesity epidemic may be attributed to combined effects of machines that
enable us to reduce the energy cost of everyday activities (‘disease vectors’) and other
10
2
5
4
00
00
97
98
99
97
98
99
–2
–2
–1
–1
–1
–1
–1
–1
99
99
71
76
88
71
76
88
19
19
19
19
19
19
19
19
Fig. 1. Prevalence trends in the disparity of overweight in American adolescents 10–18 years
(1971–2002) in the low, medium, and high SES groups. Overweight was defined as a BMI ⱖ95th per-
centile. Adapted from [14].
affordable during the more recent part of the observation period, especially in high SES
groups. Accordingly, the energy intake and expenditure patterns of all adolescents
regardless of SES, particularly for white adolescent boys, became more similar, resulting
in smaller economic disparities of the proportion of overweight subjects. Similarly, data
from the US show that in the 70s there was as much as a 50% relative difference in obe-
sity prevalence among those with less than high school education, compared to people
with college education, but by 1999–2000, the difference had decreased to 14% [13].
Similar observations of a decreasing socioeconomic gradient in obesity were reported
in one Swedish study of young adults [15]. These findings underline that individual
characteristics are probably not the main cause of the current obesity epidemic. In addi-
tion, changing patterns of consumption and of physical activity directly affect socioeco-
nomic differences in a way that is not always predictable.
The other side of the obesity-SES association is whether obesity can be shown to be a
risk factor for subsequent changes in SES. Among the first studies indicating that obe-
sity might affect social mobility was based on a Swedish population-based sample of
women examined in the late 1960s showing that the shift toward higher socioeconomic
level since childhood was more common in normal-weight than in overweight women
[16]. However, this study did not establish which subjects were already overweight
as children. One US study, which classified adolescents and young adults as being
overweight or normal at baseline, found that the overweight group, 7 years later, were
less often married, had lower income and had completed fewer years of education.
Stigmatisation in Obesity
In studies of the general population, early studies showed few consistent patterns with
respect to psychosocial distress and obesity, partly due to small samples and varying
assessment tools [21]. The relationship between BMI, smoking status, and depressive
symptoms was studied in a large US national sample, using validated instruments.
The investigators found that the relationship between obesity and depression varied
by sex. Among women, but not men, greater BMI was weakly associated with ele-
vated reports of depressive symptoms. This relationship remained significant after
controlling for age, years of education, and smoking status, indicating that relative
body weight is weakly related to psychological distress among women but not men
[25]. Another US study sought to test the relationships between relative body weight
and clinical depression, suicidal thoughts and suicide attempts in an adult US general
population sample comprising over 40,000 people. Outcome measures were past year
major depression, suicidal thoughts and suicide attempts. Among women, increased
BMI was associated with both major depression and suicide ideation. Among men,
lower BMI was associated with major depression, suicide attempts and suicidal
thoughts. There were no racial differences [26].
Studies of clinical populations have used psychometric instruments for assessment
of mental health and psychological functioning in obese individuals and compared
The obesity epidemic is driven by societal changes, as detailed above, but on an indi-
vidual level there is little evidence that psychosocial factors are causally related to obe-
sity. The socioeconomic influences are complex and variable; even in some of the
developing countries obesity is now becoming increasingly more associated with low,
not high, SES, whereas in the US, socioeconomic gradients may be decreasing. Obesity
seems to be causally related to mood disturbances, whereas there is no conclusive evi-
dence that the reverse is true. Although this has not been systematically investigated, it
seems plausible that negative and stigmatising attitudes towards obesity play a role in
the obesity-depression link. The low self-esteem in which many overweight and obese
people hold themselves secondary to the prevalent attitudes in society is probably also
important. In this respect, there are key areas in which these negative consequences
References
1 Mendez MA, Monteiro CA, Popkin BM: Overweight 13 Zhang Q, Wang Y: Trends in the association between
exceeds underweight among women in most devel- obesity and socioeconomic status in U.S. adults: 1971
oping countries. Am J Clin Nutr 2005;81:714–721. to 2000. Obes Res 2004;12:1622–1632.
2 Chopra M, Galbraith S, Darnton-Hill I: A global res- 14 Wang Y, Zhang Q: Are American children and ado-
ponse to a global problem: the epidemic of overnut- lescents of low socioeconomic status at increased risk
rition. Bull World Health Organ 2002;80:952–958. of obesity? Changes in the association between over-
3 Hill JO, Wyatt HR, Reed GW, Peters JC: Obesity and weight and family income between 1971 and 2002.
the environment: where do we go from here? Science Am J Clin Nutr 2006;84:707–716.
2003;299:853–855. 15 Lissner L, Johansson SE, Qvist J, Rossner S, Wolk A:
4 Morabia A, Costanza MC: Does walking 15 minutes Social mapping of the obesity epidemic in Sweden.
per day keep the obesity epidemic away? Simulation Int J Obes Relat Metab Disord 2000;24:801–805.
of the efficacy of a population wide campaign. Am J 16 Hallstrom T, Noppa H: Obesity in women in relation
Public Health 2004;94:437–440. to mental illness, social factors and personality traits.
5 Swinburn B, Egger G: Preventive strategies against wei- J Psychosom Res 1981;25:75–82.
ght gain and obesity. Obesity Reviews 2002:3: 289–301. 17 Gortmaker SL, Must A, Perrin JM, Sobol AM, Dietz
6 Swinburn B, Egger G: The runaway weight gain train: WH: Social and economic consequences of over-
too many accelerators, not enough brakes. BMJ 2004; weight in adolescence and young adulthood. N Engl J
329:736–739. Med 1993;329:1008–1012.
7 King T, Kavanagh AM, Jolley D, Turrell G, Crawford D: 18 Novak M, Ahlgren C, Hammarström A: A life-course
Weight and place: a multilevel cross-sectional survey of approach in explaining social inequity in obesity
area-level social disadvantage and overweight/obesity among young adult men and women. Int J Obes
in Australia. Int J Obes 2006;30:281–287. 2006;30:191–200.
8 Stunkard AJ, Sorensen TIA: Obesity and socioeco- 19 Rand CS, Macgregor AM: Successful weight loss fol-
nomic status: a complex relation. N Engl J Med 1993; lowing obesity surgery and the perceived liability of
329:1036–1037. morbid obesity. Int J Obes 1991;15:577–579.
9 Sundquist J, Johansson SE: The influence of socio- 20 Wadden TA: Adverse psychosocial consequences of
economic status, ethnicity and lifestyle on body mass extreme obesity and the effects of surgically induced
index in a longitudinal study. Int J Epidemiol 1998; weight loss. Surg Obes Relat Diseases 2005;1: 56–58.
27:57–63. 21 Fabricatore AN, Wadden TA: Psychological aspects
10 Song YM: Commentary: varying relation of socio- of obesity. Clin Dermatol 2004;22:332–337.
economic status with obesity between countries at 22 Staffieri JR: A study of social stereotype of body image
different stages of development. Int J Epidemiol 2006; in children. J Pers Soc Psychol 1967;7:101–104.
35:112–113. 23 Rand CS, Macgregor AM: Morbidly obese patients’
11 Sobal J, Stunkard AJ: Socioeconomic status and obe- perceptions of social discrimination before and after
sity: a review of the literature. Psychol Bull 1989;105: surgery for obesity. South Med J 1990;83:1390–1395.
260–275. 24 Foster GD, Wadden TA, Kendall PC, Stunkard AJ,
12 Monteiro CA, Moura EC, Conde WL, Popkin BM: Vogt RA: Psychological effects of weight loss and
Socioeconomic status and obesity in adult popula- regain: a prospective evaluation. J Consult Clin Psychol
tions of developing countries: a review. Bull World 1996;64:752–757.
Health Organ 2004;82:940–946.
Abstract
This brief overview of obesity in art will look at how fatness has been depicted in Western art and its
antecedents from classical times to the present day; what, if anything, this can tell us about how preva-
lent obesity was in previous centuries, and how the meanings attached to being fat may have altered
over the years. Copyright © 2008 S. Karger AG, Basel
The earliest sculptural representations of the body all show it as female, large-but-
tocked, obese even, although the smooth contours of the Venuses of Willendorf (c.
30,000–22,000 BC; fig. 1), and Lespugue (c. 34,000–29,000 BC; fig. 2) contrast with
the lumpy obesity of the Venus of Laussel (c. 25,000–20,000 BC; fig. 3). Nigel Spivey,
writing about the emphasis on ‘fatness and fertility’ in primitive art, offers neurosci-
entist Vilanyur. S. Ramachandran’s theory that
‘in technical terms these [excessively fleshy] features amount to hypernormal stimuli that acti-
vate neuron responses in our brain . . . something that comes naturally to us because our brains
are hard-wired to concentrate perceptive focus upon objects with pleasing associations, or those
parts of objects that matter most. For palaeolithic people, the female parts that mattered most were
those required for successful reproduction: the breasts and pelvic girdle. The circuit of the palae-
olithic brain, therefore, isolated these parts and amplified them’ [1].
Spivey argues that the tendency to distort images of the body recurs across many
cultures and periods of history. In other words:
‘The drift of all popular art is towards the lowest common denominator, and there are more
women who look like a potato than the Cnidian Venus. The shape to which the female body tends
to return is one which emphasises its biological functions . . .’ [2].
Other theorists have denied the element of exaggeration in prehistoric art, pointing
instead to the ‘relative linearity of warm-dwelling peoples, and the relative globularity of
cold-dwelling ones’ at least as far back as the Palaeolithic era, so that, even allowing for
some artistic licence, the figurines probably bear some credible relation to the models
1 2 3
Fig. 1. Venus of Willendorf (c. 30,000–22,000 BC). Naturhistorisches Museum, Vienna. © Naturhistori-
sches Museum, Vienna. Photo: Alice Schumacher. With kind permission.
Fig. 2. Venus of Lespugue (c. 34,000–29,000 BC). Musée de l’Homme, Paris. Reproduced with kind
permission from the Musée de l’Homme, Paris.
Fig. 3. Venus of Laussel (c. 25,000–20,000 BC). Musée d’Aquitaine, Bordeaux. © City of Bordeaux.
Photo J.M. Arnaud. With kind permission.
who posed for them, and the archaeological evidence can therefore be assumed to
establish ‘an empirical record as well as a merely aesthetic one’. This is particularly likely
in view of the fact that the most globular of the figures were found at sites which must
have been the coldest at the time they were sculpted, while the more linear figures have
tended to be uncovered at more southerly or warmer sites [3].
Ancient Greek and Roman sculpture and pottery art tended to portray the human
body in an idealised though still naturalistic way, a corollary of the ancient concept of
anthropomorphism – the belief that deities took human shape – but also perhaps as
further evidence of the influence of climate on body shape. However, in the interest
of aesthetics it was quite usual to attenuate the limbs, particularly the legs, and to
downplay or omit any deformity or sign of ageing or disease, even in portrait sculp-
ture. True obesity, as opposed to bulky muscularity, is notably absent in surviving
examples: excessive fat tended to produce a too solid ‘line’, rather than the more desir-
able qualities of fluidity and movement.
272 Woodhouse
Because Greek civilisation laid down that pleasure was to be moderated for the
civic good, overindulgence, whether in food or other sensual pleasures, was seen as
ugly and improper. Hippocrates, father of Greek medicine, viewed fat as a disease [4],
and Plato, believing that excessive eating led to illness, recommended what in mod-
ern times has become known as the ‘Mediterranean diet’ of mainly cereals, vegetables,
fruits and fish, with strictly limited meat, alcohol and sweet things [5]. Fat people,
perceived as flawed, were marginalised by Greek society; the obesus, or stock fat char-
acter in Greek comedies by Aristophanes and others, was a sponger, drunkard, glut-
ton or idler, far from the contemporary ideal; a figure of mockery with the additional
function of flattering the spectator’s sense of superiority.
If the Greeks sought to moderate pleasure in eating, Christians sought to extin-
guish it altogether. Food was seen by some theologians as a distraction from religious
duties, as ‘external’ and polluting; preoccupation with food was viewed as the gateway
to worse sins, sloth and lust [6] and so over centuries, the Church evolved a complex
set of rules controlling when certain foods and drink could be consumed;
Wednesdays, Fridays and the period of Lent became by custom meatless, fasting days,
and by the later mediaeval period, it is thought there were between 140 and 160 des-
ignated fast days per year [7].
The early mediaeval painters shared the same overall idea of linkage between body
shape and moral character that the Greeks did, but with a completely different con-
cept of the value of the body: whereas to the ancients its form had been shared with
the gods and was essentially admirable, to Christians the flesh had been seen as a
cause of shame and humiliation since the Fall of Man in the Garden of Eden – and
therefore as implicitly sinful and not to be flaunted. Individual portraiture as we
know it today was unknown; one legacy of the Jewish foundation of Christianity had
been the belief that images of the human form could be construed as the breach of the
2nd Commandment (to make no graven images). Rulers were often depicted on a
larger scale than ordinary people; saints and lay people all tended to be shown as slim
or even emaciated, perhaps reflecting the great emphasis in the Early Church on
asceticism, fasting and the overall denial of the flesh or maybe just mirroring a world
where consumption was restricted, for the mass of people, and hedged around by reli-
gious prescription even for the prosperous few.
By contrast, clerical exemplars of gluttony seem to have been confined to fiction
(Friar Tuck in the legend of Robin Hood, and Chaucer’s monk, who ‘liked a fat swan
best, and roasted whole’) [8]. As with Classical painters and sculptors, though, it was
generally only those at the margins of the painter’s vision – working people, such as
the wine taster depicted here, the old, the sick and wrong-doers, who were depicted as
obese (fig. 4). This resonates with contemporary experience that in the West, it is the
lower socio-economic groups, with least disposable income, who are most prone to
obesity due to the cheapness of high-calorie foods.
Artistic realism, or at least the beginnings of it, is thought to originate with the
work of Jan van Eyck, claimed as the first real portrait painter [9], although the later
mediaeval convention of enacting great events of religious and secular history only by
stylised or idealised physical specimens went on for a long time – perhaps into the
mid-19th century [10]. Van Eyck’s contemporary, Robert Campin (c. 1375–1444),
could paint a realistic head and shoulders portrait of a stout aristocrat, Robert de
Masmines (fig. 5), but such realism was not extended to larger-scale works such as the
apocalyptic scenes of Hieronymous Bosch (c. 1450–1510) or the tableaux from peas-
ant life of his artistic heir Pieter Bruegel the Elder (1525–1569), which were intended
to portray character types rather than individual personalities.
But in one Bruegel painting, The Land of Cockaigne (1567; fig. 6), three plump
characters are depicted in a fairyland in which there is food in abundance (although
the winter of 1564 had been the coldest of the century, followed by harvest failure in
1565) [11]. Cockaigne or Luilekkerland, Dutch for ‘lazy-glutton land’, was a popular
fantasy in an age when food supplies were an obsessive concern. In addition to the
stock character of the greedy peasant, a soldier and a clerk have bedded down to sleep
after a feast. Beyond, a goose lies ready-cooked on a plate, a pig brings its own carving
knife with it, and the fence is made of sausages [12].
Cockaigne apart, there was a real dearth of images of body types at the extremes of
normal; shorter life expectancy may have contributed, because gross obesity, or con-
versely, emaciation, possibly suggesting serious disease states, would have been less
274 Woodhouse
Fig. 6. Peter Brueghel the Elder, The Land of Cockaigne (detail), 1576. Alte Pinakothek, Munich.
Reproduced with kind permission from the Alte Pinakothek, Munich.
sustainable in earlier times. On the other hand, there may have been less natural
variation in body type within localities – perhaps a corollary of restricted population
mobility, and greater consanguinity.
Obesity in the context of intermittent but unrelenting famines and plagues might
seem unlikely, but perhaps something to do with the repeated switching between
restricted and abundant fare, led to rebound weight gain like that observed in ‘yo-yo’
dieting today, especially as Western populations would have been largely of the
‘thrifty genotype’ variety. Also, the ever-present threat of food shortages, coupled with
the Church’s alternating seasons of feast and fast, may well have shaped people’s eat-
ing habits in ways difficult to imagine in affluent societies today; for instance, socially
sanctioned binge eating in times of plenty may have been the general rule, rather than
the exception [13]. There would have been an annual cycle of plenty in winter, when
the animals were slaughtered, and scarcity in spring and summer in the lead up to the
harvest, and people’s eating habits would have mirrored this, with excess at harvest
time and Christmas, and frugality during Lent.
Unpredictable food supplies may not have been the only stressor; long, cold win-
ters, and, for most people, extended periods of hard physical work with little time for
rest at harvest time, may have edged the population towards excess weight gain
related to relative sleep deprivation, which has recently been linked to up-regulation
of orexigenic ghrelin and down-regulation of anorexigenic leptin [14, 15].
Women’s body shape might have been more vulnerable to the prevailing condi-
tions than men’s, especially as women generally entered adulthood with a higher per-
centage of body fat than men. Menarche could be delayed if food supplies were
sparse, resulting, potentially, in fewer births before the menopause occurred. Again,
in theory, multiparous women would tend to lay down more body fat over the course
of their fertile lives (pregnancy itself being an obesogenic process) than the childless,
presumably leading to a greater difference in BMI, all other factors being equal,
between young, unmarried women, and older mothers, than between groups of men
of comparable ages. Social historian Peter Laslett has estimated the average number
of children per pre-1700 marriage in England as 5, or up to 8 provided that maternal
death did not supervene [16]. If the rich and powerful did became obese, though,
they could choose to disguise it with flattering clothing or flaunt it as a privilege of
rank, as in the many ‘swagger’ portraits of Henry VIII [17]. Antonio Moro’s painting
of Catherine of Austria, Queen of Portugal (fig. 7), though, is an example of the
Renaissance tradition of pregnancy portraiture, rather than a depiction of an obese
woman. Pregnancy portraits like these served both as a commemoration of the
important occasion of pregnancy and a visual ‘insurance policy’ in case the mother
did not survive: at the time of painting, Catherine of Austria was 45 years old, and as
a very elderly mother-to-be was probably at far greater risk from childbirth than a
younger woman.
276 Woodhouse
8 9
During the later mediaeval period, the introduction of calorie-dense food crops
from the New World, most importantly rice, maize and the potato [18], innovations
in the preservation of meat and fish by salting, and the weakening of religious embar-
gos on consumption, led to an enrichment of the common diet. At the same time, the
upheaval of the Reformation, urbanisation and the beginnings of capitalism, led to
big changes in who commissioned works of art and why. Church patronage of art
continued, but increasing numbers of lay people with the funds to engage artists often
chose from a wider range of secular and religious themes: the body ceased to be a
focus of shame, to be depicted modestly, in a stylised way, and started to be a mar-
ketable commodity, to be promoted by the sitter or the artist, depending on who held
the real power in the transaction.
By the early 17th century, Amsterdam had overtaken Antwerp as the great interna-
tional port of the north and the chief banking centre of Europe, and it was here, in the
Dutch Republic, that there was a great boom in portraiture, and evidence in these pic-
tures of the results of a more stable food supply and better diet. Art historian Kenneth
Clark wrote that we know more about what the 17th-century Dutch looked like than
we do about any other society, except perhaps the 1st-century Romans, and he singled
out Frans Hals’ work as the exemplification of this [19]. Interestingly, there is at least
one occasion where Hals (1582–1666) painted the same individual more than once,
enabling us to see his figure bulking up over the years: Nicolaes van der Meer, a
278 Woodhouse
wealthy Haarlem brewer and burgomaster, depicted first, aged about 42 in 1616, in a
large group portrait, Banquet of the Officers of the St George Militia (fig. 8), and subse-
quently in a portrait of 1631 when he would have been around 57 years of age (fig. 9).
The well-fleshed appearance of people from all social strata in Holland reflected
the absence of the severe famines experienced elsewhere in Europe during this period
[20] and gives an air of bustling prosperity to genre paintings such as the market
scene by Emmanuel de Witte (c. 1616–1691/1692) (fig. 10). Allegedly it is with 17th-
century Dutch portraiture that breasts are seen as attractive for the first time: previ-
ously, unobtrusive, small breasts were to be preferred; large breasts being linked with
moral laxity and even witchcraft [21]: this novel voluptuousness gives another clue to
Holland’s being a well-nourished society.
A handful of prominent Flemish and Dutch artists stand out as advocates of the larger
body. Among the best known, Rubens (1577–1640) and Rembrandt (1606–1669) both
dwelt on the texture of flesh, but whereas Rubens concentrated on allegorical portraits or
tableaux featuring the legendary or the lusciously nubile (fig. 11, 12), Rembrandt, chron-
icler of life’s misfortunes, depicted all conditions of men and women, and was accused in
his day of seeking out the gratuitously unappealing (fig. 13). Later, Degas (1834–1917)
used a similarly wide range of (mainly female) subjects as Rembrandt for his mainstream
work, but it is in his less well-known brothel monotypes that he explores a very different
kind of nude: squat, short-necked (fig. 14) and very far from his pastel studies of ballet
dancers (fig. 15). Renoir’s (1841–1919) women ‘massive, ruddy … with the weight and
unity of great sculpture’ (fig. 16) [22] continue in the tradition of Rubens, rather than
emulating Degas’ realism. However, in both instances there was an emphasis on fleshi-
ness, either as a mark of sensuality or as a reflection of moral turpitude.
Heaviness and obesity had now been synonymous with wealth, success and ele-
vated social status, for centuries, but the link was soon to be put into reverse by a
complex blend of economic and social developments: by the later 19th century,
advances in agricultural methods, food processing and transportation, had brought a
varied, calorie-rich diet within the reach of all but the poorest, across much of Europe
and North America, opening up the possibility of fatness as a life ‘choice’ for the
many, for the first time ever. However, for the elites,
‘Slimness, together with speed, productivity and efficiency, were beginning to be advocated as a
new aesthetic and cultural model. A new Puritanism, which shared obvious traits with traditional
Christian penitence, re-launched the image of a lean, slender and productive body; the bourgeois
body which ‘sacrifices itself ’ to the production of goods and wealth’ [23].
For the privileged minority, eating to excess as a way of displaying wealth and priv-
ilege was gradually being replaced by other forms of conspicuous consumption: the
concept of overweight being a hindrance to living a newer, more fluid kind of ‘good
life’ dawned first with those at the top of society, and diffused downwards [24].
In the early stages of this transformation, slimness would have been seen as a novelty
(just as obesity had once been) for the majority of people. Fleshiness was still linked to
prosperity in the collective mind, and thinness was often associated with tuberculosis
280 Woodhouse
and chronic ill health [25]. It was only at the very end of the 19th century that slimness,
rather than obesity, began to be seen (first in the United States) as a desirable standard
for all, heralded by a dramatic increase in the numbers of advertisements for diets, and
editorials in the popular and medical press on the dangers of excess weight [26].
Suggested reasons for why disapproval of obesity gained ground so quickly have
included deliberate manipulation by an emergent diet industry, an increased middle-class
interest in athleticism (manifested by the bicycle craze of the 1880s and 1890s), both
perhaps springing from accelerating industrialisation and urbanisation, and an overall
speeding up of the pace of life for everyone. Now, to aspire to slimness meant to stand
out from the crowd, just at a time when most people had managed to achieve the req-
uisite standard of living to be plump. In addition, photographic portraiture was
becoming increasingly accessible and revealing to its subjects their true, three-dimen-
sional shape for the first time. More reliable methods of contraception were becoming
generally available; women could more easily limit family size, and so reduce the
amount of weight they put on during their fertile years. In this new ‘machine age’, car-
rying excess weight might become a hindrance to ‘staying ahead’ [27].
Once the connection between slimness and social advantage had been formed, it
was just a short step to obesity becoming socially undesirable, stigmatising even: by
the early 20th century, this had been reinforced by insurance companies’ promoting
ideal body weight to height tables [28], showing that the medical establishment was
now ‘on board’ also.
During the century since then, the desirability of slimness over heaviness has been
accepted more or less unchallenged by Western societies: its cachet has increased as
the average Western citizen has become fatter and the population profile has aged –
slimness now being associated with youth. ‘Super-sizing’ of the human body in art
has continued, but has been reserved for monumental sculpture such as the socialist
realist statuary of countries of the former Soviet bloc, for surrealist interpreters of the
human form or for artists with agendas involving humour, social comment, or a
voyeuristic take on the outsized.
Lucien Freud (b. 1922), speaking about his model for Benefits Supervisor Resting,
admitted that he ‘had perhaps a predilection towards people of unusual or strange pro-
portions’ and had become aware of ‘all kinds of spectacular things to do with her size,
like amazing craters and things one’s never seen before’ [29]. (Indeed, one art historian
found her body proportions to be identical to those of the Willendorf Venus [30]).
Jenny Saville (b. 1970) was well-known for her massive, uncompromising canvases
of naked, obese women even before the controversial 1997 ‘Sensation’ exhibition at
the Royal Academy, where she shared billing with others similarly engaged with
mutilated and abnormal forms. She started painting in the 1980s when ‘everyone was
obsessed with the body – it was all about dieting, the gym, the body beautiful.
Pornography and AIDS were the big debates’ [31]. Her huge images were always
female, ‘massive as the Eiger’, ‘daunting’ and ‘confrontational’ [32]. Sponsored by
Charles Saatchi, she spent part of 1994 in New York observing and photographing
282 Woodhouse
Fig. 14. Edgar Degas, The Customer,
1876–1877. Musée d’Orsay, Paris. Reproduced
with kind permission from the Musée d’Orsay,
Paris.
Fig. 15. Edgar Degas, Dance Lesson, 1872. Musée d’Orsay, Paris. Reproduced with kind permission
from the Musée d’Orsay, Paris.
Fig. 16. Pierre-Auguste Renoir, The Great Bathers (The Nymphs), 1918–1919. Musée d’Orsay, Paris.
Reproduced with kind permission from the Musée d’Orsay, Paris.
Fig. 17. Jenny Saville, Plan, 1993. Gagosian Gallery, New York, N.Y. © Jenny Saville. Courtesy of
Gagosian Gallery, New York, N.Y. Photograph by Robert McKeever.
284 Woodhouse
Fig. 18. Thomas Rowlandson, A Little Tighter,
1791. British Museum, London. © The Trustees of
the British Museum, London. With kind
permission.
activities [38]. In other words, once the daily business of survival ceased to be a
struggle, conspicuous markers of success other than extreme obesity would have
been sought. Even so, from the evidence, the ideal weights for both men and women
would probably have been heavier than present day ones for most of the period
under consideration.
The trim, athletic body proportions of classical art were feasible in the context of
stable, prosperous, mercantile societies with good and varied food supplies and a
warm climate, but when, subsequently, contentious northern European civilisations
operating in colder climates became pre-eminent, the depicted ideal was replaced by
a stockier, less elegant model. From early modern times until just over a century ago,
excess weight had positive associations with wealth, success, physical strength and
health, and none of its current negative associations with sudden death, chronic dis-
ease, shorter life expectancy and ‘loser’ social status. Intimations that excess weight
might have drawbacks coincided with a ‘democratisation’ of obesity, as a high-calorie
diet came within the reach of the majority, and led to slimness soon replacing heavi-
ness as a mark of social distinction, initially just for women, but subsequently for men
as well.
Fat in art could no longer be so ‘mainstream’, once the body ideal had shifted, but
still remained as an important theme in painting, particularly in naive art. The para-
dox now is that while thinness has become ever more valued, in real life and in the
media, the prevalence of obesity in society has soared, and even Saville and Freud
nudes no longer outscale everyone around them.
Copyright. The author and the publisher have made every effort to obtain permission for all copyright-protected
material. Any omissions are entirely unintentional. The publisher would be pleased to hear from anyone whose rights
unwittingly have been infringed.
286 Woodhouse
Author Index
287
Subject Index
288
leptin regulation 109 DNA microarray analysis 124, 125
melanocortin system regulation 110 fatty acid-induced apoptosis 122, 123
neuroendocrine control 68 genetics 120–122
neuropeptide Y regulation 109, 110, 139, glucolipotoxicity 122
169 mitochondrial oxidative metabolism
regulation in large animals and birds 125, 126
108–110 peroxisome proliferator-activated
2-Arachidonoylglycerol (2-AG), receptor-␥ 124
endocannabinoid system 137, 138 signaling processes 125
Area postrema (AP), peripheral signals of SREBP1c 123
energy balance 167 loss in diabetes 119
ARNT, expression in diabetic -cells 124 microRNA regulation of function 127
Art, obesity representations stimulus-secretion coupling 118, 119
ancient Greece and Rome 271–273 Body mass index (BMI)
18th century 279 children 86–88
famine and plague 275 classification 73, 229
medieval period 273–277 elderly 97, 98
19th century 279, 281, 283 Börjeson-Forssman-Lehmann syndrome
prospects 285 (BFLS), features 52
17th century 277–280 Bupropion, obesity management 251
20th century 281, 282, 284
women’s body shape 276, 284 Candidate gene
Association analysis, genetic dissection of approach for genetic dissection of complex
complex diseases 16 diseases 14
diabetes type 2 risk genes
Bardet-Biedl syndrome (BBS) CDKAL1 26
association studies 40 CDKN2A 26
clinical features 37–40 genome-wide scans 25, 26
obesity 40 IGFBP2 26
Bariatric surgery TCF7L2 24, 25
appetite reduction effects 248–250 gene-gene interactions 26, 27
childhood obesity management 94 obesity risk genes
complications 243, 244 ENPP1 21, 22
diabetes type 2 response 245, 250, 251 FTO 22, 23
elderly patients 104, 105 INSIG2 20, 21
gut hormone response 179, 248–250 peroxisome proliferator-activated
malabsorptive and hybrid procedures 242, receptor-␥ 18–20
243 Cannabinoids, see Endocannabinoid system
metabolic syndrome response 245 Cardiovascular disease, AMP-activated
non-alcoholic fatty liver disease response protein kinase role 207
246 Carpenter syndrome, features 42
obstructive sleep apnea response 246 Catecholamines, thyroid hormone
restrictive bariatric surgery 241, 242 interactions 216
success criteria 240, 241 CDKAL1, diabetes type 2 candidate gene
weight loss outcomes 243 26
Beta-cell CDKN2A, diabetes type 2 candidate gene
AMP-activated protein kinase activation 26
effects 202 Children
diabetes type 2 defective insulin secretion genetic dissection of complex diseases 17
mechanisms growth importance on long-term health
dense core vesicle exocytosis 126, 127 85, 86