SCIENTIFIC STATEMENT€¦ · underlying this “upward setting” or “resetting” of the defended level of body-fat mass, whether inherited or acquired. The ongoing study of how

Obesity Pathogenesis: An Endocrine SocietyScientific Statement

Michael W. Schwartz,1 Randy J. Seeley,2 Lori M. Zeltser,3 Adam Drewnowski,4 Eric Ravussin,5

Leanne M. Redman,5 and Rudolph L. Leibel3,6

ABSTRACT Obesity is among the most common and costly chronic disorders worldwide. Estimates suggest

that in the United States obesity affects one-third of adults, accounts for up to one-third of total mortality, is

concentrated among lower income groups, and increasingly affects children as well as adults. A lack of effective

options for long-term weight reduction magnifies the enormity of this problem; individuals who successfully

complete behavioral and dietary weight-loss programs eventually regain most of the lost weight. We included

evidence from basic science, clinical, and epidemiological literature to assess current knowledge regarding

mechanisms underlying excess body-fat accumulation, the biological defense of excess fat mass, and the

tendency for lost weight to be regained. A major area of emphasis is the science of energy homeostasis, the

biological process that maintains weight stability by actively matching energy intake to energy expenditure

over time. Growing evidence suggests that obesity is a disorder of the energy homeostasis system, rather than

simply arising from the passive accumulation of excess weight. We need to elucidate the mechanisms

underlying this “upward setting” or “resetting” of the defended level of body-fat mass, whether inherited or

acquired. The ongoing study of how genetic, developmental, and environmental forces affect the energy

homeostasis system will help us better understand these mechanisms and are therefore a major focus of this

statement. The scientific goal is to elucidate obesity pathogenesis so as to better inform treatment, public

policy, advocacy, and awareness of obesity in ways that ultimately diminish its public health and economic

consequences. (Endocrine Reviews 38: 1 – 30, 2017)

DefinitionRationale for a scientific statement onobesity pathogenesis

ENERGY HOMEOSTASIS AND THEPHYSIOLOGICAL CONTROL OFBODY-FAT STORESGeneral principlesBackgroundLeptin and energy homeostasisFuel partitioning, insulin, and obesity

Neurobiology of energy homeostasisBackgroundHypothalamic neurons controllingenergy balance

Hindbrain circuits and the parabrachialnucleus

Developmental considerationsIntegrative physiology of energyhomeostasisDeterminants of feeding behaviorDeterminants of energy expenditure

MECHANISMS OF OBESITYPATHOGENESIS

Genetic factors: evidence for and againstInteractions between genes,development, and environment

Role of epigenetic modificationsDevelopmental factors: evidence for andagainstRoles of parental body weight or dietUndernutrition

Overnutrition/obesity

Modes of transmission

Role of EDCsPFCsBPAModes of transmission of endocrinedisrupting chemical effects

GI factors, bariatric surgery, and themicrobiomeInsights from bariatric surgeryThe gut microbiome and other GIfactors

Social and economic factorsDiet composition, lifestyle, and obesityrisk

Impact of diet composition on obesityrisk

Roles of sedentary behavior, exercise,and nonexercise activity thermogenesis

Other factorsSmoking cessationInfectious factorsMechanisms for biological defense ofelevated body-fat mass

CONCLUDING REMARKS AND FUTUREDIRECTIONSThe two distinct components of obesitypathogenesis

Developmental determinants of thebiologically defended level of body-fat mass

Interactions between genetics,epigenetics, developmental influences,and the environment

Future directions for EDC researchLessons learned from the weight-reduced state

The gut–brain axisDietary influences

ISSN Print: 0163-769X

ISSN Online: 1945-7189

Printed: in USA

Copyright © 2017 Endocrine

Society

Received: 11 May 2017

Accepted: 12 May 2017

First Published Online:

26 June 2017

AFFILIATIONS1Diabetes Institute, University of

Washington, Seattle, Washington

981092Department of Surgery,

University of Michigan, Ann

Arbor, Michigan 481093Naomi Berrie Diabetes Center

and Department of Pathology

and Cell Biology, Columbia

University, New York, New York

100324Center for Public Health

Nutrition, University of

Washington, Seattle, Washington

981955John S. McIlhenny Skeletal

Muscle Physiology Laboratory,

Pennington Biomedical Research

Center, Baton Rouge, Louisiana

708086Division of Molecular Genetics,

Department of Pediatrics,

Columbia University, New York,

New York 10032

doi: 10.1210/er.2017-00111 https://academic.oup.com/edrv 1

SCIENTIFIC STATEMENT

http://dx.doi.org/10.1210/er.2017-00111

https://academic.oup.com/edrv

Stratification of obesity outcomesUnraveling mechanisms linking theenvironment to defense of elevatedbody weight

Identifying and mitigatingenvironmental risk factors

Translating basic science into moreeffective pharmacotherapy

T he need to integrate molecular, genetic, de-velopmental, behavioral, and environmental

factors highlights the substantial challenge inherent inachieving a comprehensive understanding of obesitypathogenesis. Mechanistic formulations must drawfrom disciplines that include: the neuroscience offeeding behavior; the psychology of food reward; themetabolic impact of specific nutrients and changes ofphysical activity; as well as genetics, epigenetics, anddevelopmental biology relevant to energy balancecontrol, and the influence of exposure to environ-mental variables ranging from endocrine-disruptingchemicals (EDCs) to socioeconomic factors. Whenprocessing this information, one must also be mindfulthat although there are many interventions that cancause obesity in an experimental setting, the keyquestion is whether they do cause obesity in a natu-ralistic environment. In this statement, we focus onfactors for which compelling evidence exists thatimplicates them in the pathogenesis of either theaccumulation or maintenance of excess body fat mass.

DefinitionObesity is broadly defined as an excess of body-fat mass. Reliable fat-mass quantitation requiressophisticated tools that are not widely available (e.g.,magnetic resonance imaging or dual energy X-rayabsorptiometry), and this has hampered efforts toarrive at a more specific definition. Consequently, anelevated body mass index (BMI), which expresses bodyweight (in kilograms) as a function of body height(in meters) as a surrogate measure of body fatness,is the most widely accepted definition of obesity.Population-based actuarial studies place the upperlimit of normal BMI in adults at kg/m, defineobesity as a BMI . kg/m, and designate a BMI

between these values to be “overweight.” The degree ofobesity can be further subcategorized into class (BMIof to ,), class (BMI of to ,), and class (BMI of .) (). Assessing BMI in children requiresadjusting for both age and gender.

BMI satisfies our need to estimate body-fat mass ata population level and thus gauge a group’s susceptibilityto complications of obesity. However, it is not a reliableclinical tool for assessing individual body fatness, becausevariations in skeletal muscle and other lean-body-masscomponents create substantial variations in total bodymass. For example, a heavily muscled individual withincreased body weight relative to height will have a BMIvalue that can erroneously place them into the over-weight or even obese category. Additionally, there aresignificant racial/ethnic differences in how BMI associ-ates with adverse medical consequences. (For addi-tional information, see the companion Endocrine SocietyScientific Statement titled “Obesity Management: Past,Present, and Future; Science and State of the Art.”)

Rationale for a scientific statement onobesity pathogenesisFor most endocrine disease, researchers have estab-lished effective therapeutic treatments based on un-derlying disease mechanisms. This is not the case withobesity; unlike most other endocrine disorders, wehave a very limited understanding of its pathogenesis,despite decades of research and billions of dollarsspent each year on its treatment. This gross expen-diture of time and money is undoubtedly linked to theextraordinarily high prevalence (affecting one-third ofthe adult United States population) (), ease of de-tection, and stigma associated with obesity. Thesefactors conspire to create an enormous demand forweight-loss products and services that continue to

ESSENTIAL POINTS

· Obesity pathogenesis involves two related but distinct processes: (1) sustained positive energy balance (energy intake.energy expenditure) and (2) resetting of the body weight “set point” at an increased value. The latter process explainswhy weight lost through changes of diet and/or lifestyle tends to be regained over time, a major obstacle to effectiveobesity treatment.

· How the increased body weight comes to be biologically defended remains uncertain, although ongoing research isbeginning to shed some light on underlying mechanisms. Therapeutic strategies that target these mechanisms have thepotential to reset the defended level of body weight at a lower, more normal value.

· The impact of diet on obesity risk is explained largely by its effect on calorie intake, rather than by changes of either energyexpenditure or the internal metabolic environment. Stated differently, "a calorie is a calorie.” Thus, habitual consumptionof highly palatable and energy-dense diets predispose to excess weight gain irrespective of macronutrient content.

· Beyond diet, environmental factors ranging from socioeconomic status to chemical exposures to sedentary lifestyle canconfer obesity risk. How these inputs interact with genetic, epigenetic and developmental factors that predispose toobesity is a key question for future research.

2 Schwartz et al Obesity Pathogenesis Endocrine Reviews, August 2017, 38(4):1–30


flourish, despite being largely ineffective, sometimesdangerous, and almost entirely unregulated ().

This situation is not unlike the medical practice ofa century ago in which “glandular extracts”were cleverlymarketed for a multitude of diseases, generating robustsales and profits for their manufacturers despite a lack ofefficacy or safety data (, ). It was largely in response tothe rise of this practice (termed “organotherapy”) thatthe Endocrine Society chose a different path. In thethird year of its existence, the Endocrine Society electedSir Harvey Cushing as President. In his presidentialaddress, he advocated strongly in favor of adopting thescientific method and abandoning empiricism to betterinform the diagnosis and treatment of endocrine disease(). In doing so, Cushing helped to usher in the modernera of endocrinology and with it, the end of organo-therapy. (In an interesting historical footnote, Cushing’s

presidential address was given in , the same yearthat insulin was discovered.)

Clearly, we need a well-defined, generally acceptedset of physiological, developmental, and environ-mental principles regarding body weight homeostasisthat will inform strong research and therapeuticstrategies regarding obesity pathogenesis. The currentlack of consensus regarding obesity pathogenesis hasresulted in competing and poorly justified claims bothfrom within and outside of the scientific community.These inconsistencies erode public trust and confi-dence in the scientific process as it pertains to obesityand its treatment, which only further supports non-scientific ideologies and products. To break this vi-cious cycle, and to identify effective treatments, weneed to establish clearly defined and reliable dataregarding obesity’s underlying causes.

Energy Homeostasis and the PhysiologicalControl of Body-Fat Stores

General principles

BackgroundAt its most basic level, the pathogenesis of obesityseems simple: Calories are consumed in amounts thatexceed ongoing energy expenditure. Based largely onthis concept, most people have historically perceivedobesity as the result of negative personal traits, such asgluttony, sloth, self-indulgence, laziness, and lack ofwill power. However, growing evidence indicates thatobesity pathogenesis involves processes far morecomplex than the passive accumulation of excesscalories. It is this complexity that lies at the heart ofwhy obesity is so difficult to treat. Fundamentally,humans have an “evolutionary physiology” that ispredisposed to conserve body fat as a factor of survival.This evolutionary physiology in today’s climate of easyaccess to virtually unlimited calories has created a largesegment of humanity that appears to be biologicallypredisposed to excessive weight gain. Hence, we seeupward trends of adiposity in developed and de-veloping communities.

How does the energy homeostasis system bear onthis issue? We see clear evidence of a properly op-erating energy homeostasis system in the remarkablebody-weight stability of individuals who are notobese over long periods of time. Evidence from anobservational study of , healthy Swedish women() indicates that participants were, on average,..% accurate in their annual matching of energyintake to expenditure for consecutive years ofobservation. To better understand the implicationsof this observation, consider that a healthy adultweighing pounds can be expected to gain ..pounds in a year if they expend fewer calories perday than they consumed. During years of adulthood[assuming a caloric intake of kcal/day (.M kcaltotal)], a weight gain of pound (~ kcal/pound

mixed body tissue) per year (, kcal total) isequivalent to a .% positive caloric balance. Thus, weinfer that such individuals are .% accurate inmatching energy intake to expenditure. There are caveatsto such calculations relating to the fact that increasedbody mass increases energy expenditure (further re-ducing the balance error), but from a thermodynamicperspective, it is clear that obesity is generally theconsequence of small, cumulative imbalances of energyintake and expenditure. Although the causes of theseimbalances can involve innumerable genetic, de-velopmental, and/or environmental factors, once in-dividuals who are obese and individuals who were neverobese achieve their “customary” body weights andcompositions, they tend to maintain and defend thoseweights by identical mechanisms.

Studies investigating the adaptive responses ofnormal-weight humans and animals to changes inbody weight support the concept of a physiologicallyimportant energy homeostasis system. Weight lossinduced by caloric restriction, for example, results inboth an increased drive to eat and a reduction ofenergy expenditure. These responses both resist fur-ther weight loss and favor recovery of lost weight,and they can persist for years, provided that body-fat stores have not returned to baseline (). Theseadaptive responses to weight loss are reported in bothindividuals who are obese and lean individuals (),therefore suggesting that obesity pathogenesis involvesthe physiological defense of a higher level of body fat.This perspective offers a plausible explanation for thevery frequent regain of lost weight that confoundsmost forms of obesity treatment (, ).

Conversely, normal-weight subjects respond toexperimental weight gain (induced by “forced over-feeding”) by increasing energy expenditure andreduced hunger. Once forced overfeeding is dis-continued, a combination of decreased drive to eat andincreased energy expenditure tends to restore bodyweight to normal (). Indeed, such overfeeding studiesshow that it is surprisingly difficult for normal-weight

3doi: 10.1210/er.2017-00111 https://academic.oup.com/edrv


http://dx.doi.org/10.1210/er.2017-00111


individuals to achieve and sustain experimentally in-duced weight gain (). Individuals who are obese alsoresist excess weight gain induced by forced overfeeding(). Therefore, their elevated levels of body-fat massappear to be similarly subject to biological defense.Stated differently, individuals who are obese and in-dividuals who are not obese appear to use the samehomeostatic mechanisms to defend different levels ofbody-fat mass. This observation suggests that dysfunc-tion of the energy homeostasis system is both necessaryand sufficient for the biological defense of elevated bodyweight in individuals who are obese (). What remainsunclear is how this dysfunction is linked to factors thatenable excess weight gain, such that excess body-fat masscomes to be biologically defended. This issue is central toobesity pathogenesis and therefore a central focus of thisscientific statement.

Leptin and energy homeostasisThe adipocyte hormone leptin, which circulates atconcentrations proportional to body-fat mass, playsa significant role in the relationship between obesity andenergy homeostasis. A deficiency of leptin causes severehyperphagia and obesity in both humans and animals(), with physiological leptin replacement amelioratingboth the hyperphagia and obesity in leptin-deficientindividuals (). Therefore, there can be no questionthat normal body-weight maintenance in humans re-quires intact leptin-regulated neurocircuits.

However, these observations do not indicate thatgenetic deficiencies of leptin or its cognate receptorare important causes of human obesity. Althoughsuch individuals exist, they are rare (). In contrast,most individuals who are obese have elevated plasmaleptin levels (in proportion to the increase of body-fatcontent), raising the possibility that common forms ofobesity are associated with “leptin resistance” (i.e., thatsupraphysiological plasma leptin levels are required toovercome tissue resistance to leptin and thereby enableenergy intake and energy expenditure to match oneanother). Because adipocytes secrete leptin in proportionto body-fat content, the only way to raise plasma leptinlevels in this setting is to become obese.

These considerations would seem to point toa causal role for leptin resistance in the pathogenesis ofcommon forms of obesity, but the matter remainsunsettled. For one, there is no uniformly agreed-upondefinition for leptin resistance (), and the presenceof hyperleptinemia per se cannot be taken as evidenceof its presence. Indeed, recent data suggest thatthe cellular response to leptin (e.g., activation of in-tracellular STAT signaling) is preserved in obese,hyperleptinemic rodents (). The circulating leptinconcentration needed to fully engage central nervoussystem responses likely differs among individualsbased on the influence of genetics, development, andpossibly diet. Thus, some individuals who are obesemay simply require more leptin (and hence body fat)to fully engage relevant neurocircuits ().

Given the evolutionary considerations alluded toabove, the primary role played by leptin-responsive

neurocircuits may be related more to preventing lossof body fat (communicated to the brain by a decreaseof leptin signaling) than to defending against its in-crease (conveyed by increased leptin levels). In thisformulation, genetics, development, and even envi-ronmental factors can influence the level of leptinsignaling (“threshold”) below which compensatoryincreases in food intake and reductions in energyexpenditure occur. Accordingly, this theory holds thatleptin circuitry is more sensitive to decreases than toincreases in the circulating leptin level, with the limitedresponse to leptin concentrations above the lowerthreshold offering a potential explanation for what somerefer to as leptin resistance. The apparent resistance inthis formulation is simply a reflection of the circuitry’sdesign (, ). More work in this area is clearly needed.

Fuel partitioning, insulin, and obesityGeneral perspectives on obesity pathogenesis haveswung from early conceptualizations of obesity asa storage disease of adipose tissue (analogized pre-sumably to some of the earliest identified inborn errorsof metabolism by Garrod) to more recent brain-centricmodels, in which the brain, by virtue of its operationalcontrol of food intake and energy expenditure, im-poses excess calories on passive adipocytes. Sometimesreferred to as “pull” and “push” models, respectively,each makes very different predictions regarding bothunderlying molecular mechanisms and how we shouldapproach obesity prevention and treatment. Embed-ded within this debate is the extent to which adipocyte-autonomous processes can pull substrate moleculespreferentially into adipocytes, and by “partitioning”calories in this way cause higher fractional depositionof calories as fat. A key prediction of this model is thatsome individuals who consume a diet in which caloricintake and energy expenditure are matched (e.g., anisocaloric diet) will preferentially deposit ingestedcalories as fat at the expense of their lean tissue mass,thereby becoming relatively fatter without enteringa state of positive energy balance per se (see the sectiontitled “Impact of diet composition on obesity risk”).

Although it is possible that variations in bodycomposition among individuals of the same bodyweight reflect (to some extent) the consequences ofsuch processes, achieving clinical obesity in thismanner must be rare, because most individuals whoare obese have absolute increases of both lean and fatmass. Nevertheless, when leptin-deficient ob/ob miceare pair-fed the same amount of food consumed bynormal controls, they gain more mass (due to reducedenergy expenditure) and preferentially deposit thatmass as fat, leading to an absolute and relative de-ficiency of lean mass in these animals (). Althoughthe mechanism underlying this partitioning effectmust relate to developmental and/or intercurrent ef-fects of leptin (perhaps involving interactions withinsulin as well), the precise biology is not clear ().The hypothesis that leptin plays a direct role in theseprocesses is supported by evidence of its direct effectson lipid partitioning in skeletal muscle ().



Lipoprotein lipase (LPL), an enzyme that hy-drolyzes circulating apolipoprotein-bound acylgly-cerides at the surface of many cell types (includingadipocytes and myocytes), also appears to affect thepartitioning of fatty acids in ways that affect bothabsolute and relative fat mass. Although miceoverexpressing LPL in skeletal muscle accumulatetriglyceride in muscle and are resistant to increasesof adipose tissue mass during overfeeding (),overexpression of LPL in adipocytes does not affectbody weight or adiposity in mice (). Restoration ofmuscle LPL in mice that otherwise lacks the enzymecreates animals functionally lacking LPL in adiposetissue; surprisingly, these mice are characterizedby normal body-fat mass, apparently because ofa compensatory increase of fatty acid synthesis inadipocytes (). Expression of LPL varies widely bydepot in humans, and it might account for differ-ences in adipose tissue distribution in some in-dividuals. As insulin increases the synthesis andactivity of LPL (while also stimulating adipocyteuptake of fatty acids and glucose), intrinsic differ-ences in insulin-mediated molecular processescould (in theory) play a determinative role in body-fat content and/or distribution. In this context,however, it is important to note that humans lackingLPL on a genetic basis have normal fat mass due tothe ability of adipocytes and muscle to take upcirculating free fatty acids in quantities sufficient toallow adequate aclyglyceride formation in adiposetissue or oxidation in muscle (). Therefore, al-though local activity of LPL (in adipocytes andmuscle) could play a role in partitioning of fatamong tissues, it appears to be neither necessary norsufficient for the uptake of fatty acids into adiposetissue.

Several investigators have proposed that the effectof specific diet components on insulin secretion oraction contributes to obesity pathogenesis througheffects on calorie deposition in adipocytes, rather than(or in addition to) effects on energy balance per se.Essentially, carbohydrates in general (and refined andpossibly naturally occurring sugars in particular) areproposed to promote hyperinsulinemia that in turndrives glucose and fatty acids into adipose tissue ().Accordingly, this process is proposed to cause obesityby both direct effects on adipocytes that favor fatdeposition and by lowering circulating metabolicsubstrates (and/or exerting effects on hepatic meta-bolism) that subsequently stimulate food intake. Addi-tional “lowering” effects on energy expenditure bythese dietary components are proposed to exacerbatethe tendency toward increased fat deposition (). Aswe discuss in greater detail in a later section (“Dietcomposition, lifestyle, and obesity risk”), this hy-pothesis remains controversial and has yet to receivethe level of support needed for broad acceptance.Among several concerns is that differences in dietcomposition have yet to be convincingly shown tocause differences in body composition when providedin an isocaloric manner (such that total calorie

consumption is matched between diets) (). This isnot to say that diets high in refined carbohydrate and/or fructose (soft drinks) do not predispose to obesity,but the underlying mechanism is likely to involveexcessive intake of calories, rather than nutrient-specific or hormonal effects on substrate partitioning.

Collectively, these data suggest that diet compo-sition per se (relative quantities and specific types ofcarbohydrates, sugars, and fatty acids, as distinct fromcaloric content) contributes far less to the etiology ofobesity than do contributions made by the net im-balance of intake and expenditure. It therefore followsthat although variations in diet composition canpowerfully affect palatability and hence hedonicallymotivated feeding, whether it can also influence foodintake via secondary metabolic consequences (e.g.,effects of insulin on circulating nutrient levels) remainsa highly controversial topic that seems unlikely to beresolved without clinical studies that will be costly andchallenging to undertake.

Neurobiology of energy homeostasis

BackgroundMore than years ago, to account for evidence thatbody weight is biologically defended, Kennedy() proposed that body-fat mass is regulated by an“adiposity negative feedback” system. Specifically, hesuggested that circulating signals inform the brainregarding the amount of body fuel stored in the formof fat. In response, the brain makes compensatoryadjustments to both food intake and energy expen-diture so as to promote weight stability. Compellingsupport for this hypothesis has emerged in the decadessince, including the identification of neurocircuits andsignaling molecules that sense, integrate, and trans-duce humoral input relevant to body-fuel stores intoadaptive changes of energy balance. Because a credibleformulation of obesity pathogenesis depends ona comprehensive understanding of how food intakeand energy balance are controlled, we explore thistopic in detail here.

Motivational aspects of feeding behavior providea useful context within which to introduce the topicof energy balance neurocircuitry. Both “metabolicneed” and “food reward” have long been consideredkey drivers of feeding. The former involves a desire toalleviate the discomfort associated with inadequatefood availability (“drive reduction”), whereas thelatter describes the anticipation of a rewarding ex-perience and subsequent fulfillment of that experi-ence (). Recent advances in neuroscience haveenabled the identification of neuronal substratesimplicated in these distinct but complementarysources of motivation.

Hypothalamic neurons controlling energy balancePerhaps the best studied subset of neurons involvedin feeding behavior are those that co-express neuro-peptide Y (NPY), agouti-related protein (AgRP) (anantagonist of melanocortin signaling), and the inhibitory



http://dx.doi.org/10.1210/er.2017-00111


neurotransmitter, g-aminobutyric acid (GABA);henceforth referred to as AgRP neurons, because theyare the only neurons that express AgRP. Located in anarea of the mediobasal hypothalamus known as thearcuate nucleus (ARC), AgRP neurons are activated inconditions of negative energy balance and weight loss(e.g., fasting), in part because such conditions reduceplasma concentrations of leptin and insulin, hormonesthat tonically inhibit these neurons (). Because thisinhibitory input becomes quiescent during fasting,AgRP neurons are activated and increase the drive toeat.

What evidence causally links AgRP neuron acti-vation to increased food intake? Thanks to recentadvances in optogenetics and related neurosciencemethods, researchers have been able to investigateresponses triggered by the selective activation or in-hibition of uniquely identified neuronal subsets inconscious, freely moving animals. This work has con-vincingly revealed the powerful hyperphagic responseelicited by selective AgRP neuron activation ().Combined with evidence that AgRP neurons areactivated across a variety of states of metabolic needthat drive hyperphagic feeding (e.g., fasting, un-controlled diabetes, genetic leptin deficiency, andhypoglycemia), researchers have suggested a causalrole for their activation in the associated hyperphagia(). This possibility received direct support fol-lowing the demonstration that experimental silenc-ing of AgRP neurons prevents hyperphagia elicitedby fasting ().

Recent work has identified several unique andunanticipated properties of AgRP neurons. Usingcell type–specific in vivo calcium imaging in con-scious, free-moving mice, Knight et al. () docu-mented that although AgRP neurons are activated infasted mice, as expected, they cease firing upon thesight of food, prior to feeding onset. Althoughoriginally interpreted to suggest that activation ofAgRP neurons merely prepares animals to eat,rather than driving feeding behavior per se, the samegroup showed that if food is made available onlyafter activation of AgRP neurons ceases, intake stillincreases markedly (as long as food is made availablewithin minutes) (). Thus, activation of AgRPneurons provides a robust stimulus to feeding thatcontinues throughout a meal, even though activityof these neurons ceases prior to meal onset.

Whether activation of these neurons increasesintake by enhancing the rewarding properties offood or whether it motivates feeding through drivereduction (a desire to alleviate the discomfort as-sociated with not eating) is an active and somewhatcontroversial area of study. Sternson et al. ()suggested a key role for the latter mechanism, im-plying that AgRP neuron activation is aversive whenfood is not available, and that feeding amelioratesthis effect. Conversely, Knight et al. () reportedthat as long as food is available, animals perceiveAgRP neuron activation as being highly rewarding.Whether animals perceive AgRP neuron activation

as a positive or a negative experience may thereforedepend on whether food is available.

Situated adjacent to AgRP neurons in the ARC areneurons that express pro-opiomelanocortin (POMC)and release the anorexic neuropeptide a-melanocyte–stimulating hormone. Food intake is reduced followingthe activation of melanocortin receptors expressedon “downstream” target neurons in the paraventricularhypothalamic nucleus and other brain areas, and leptinstimulates POMC neurons (). In conditions of nega-tive energy balance and leptin deficiency, therefore,POMC neurons are inhibited, whereas AgRP neuronsare activated (). This combination drives feeding inthe same way that a car is propelled forward by steppingon the accelerator pedal (AgRP neuron activation) whilealso removing one’s foot from the brake pedal (POMCneuron inactivation).

Just as leptin’s ability to reduce food intake andbody weight requires an intact melanocortin system,mutations that impair the melanocortin system causehyperphagic obesity in both humans and rodentmodels (). POMC neurons are also targets for theaction of certain anorexic agents, including seroto-nergic drugs (). Because the degree of hyperphagiaand obesity induced by defective melanocortin sig-naling is greatly enhanced by consuming a highlypalatable diet, the melanocortin system appears to playa physiological role to limit reward-based feeding ().Additionally, a recent study reported that adjacent toAgRP and POMC neurons in the ARC is a distinct andpreviously unrecognized subset of excitatory neuronsthat, when activated, powerfully and rapidly inhibitfeeding (). How these neurons fit into the biggerpicture of energy homeostasis will undoubtedly be thesubject of intensive additional study.

A relevant consideration here is that food intakeregulation involves distinct components operatingacross very different time periods. Some neurohu-moral mechanisms exert very rapid effects of shortduration (e.g., a single meal), whereas others are moremodest but sustained over long time intervals. Yeteach is somehow integrated to enable the precise long-term constancy of body weight mentioned above, andeach can respond in an integrated manner to per-turbations of body weight, although not necessarilythrough reciprocal effects on the same processes.

Delineating the mechanisms responsible for thisseamless integration is a critical issue for the futurestudy of the biology of energy homeostasis. Relatedissues are how homeostatic and hedonic drivers ofenergy intake interact, how this interaction is woveninto long-term control of energy balance, and whetherdefects in this integration contribute to obesitypathogenesis.

Hindbrain circuits and the parabrachial nucleusBecause experimental activation of POMC neurons isnot associated with rapid or potent feeding inhibition,the melanocortin system does not appear to be im-portant for meal termination under physiologicalconditions. The latter process involves meal-induced



secretion of gut-derived peptides, such as glucagon-like peptide and cholecystokinin, that, following theirrelease by enteroendocrine cells in the gastrointestinal(GI) tract, play a physiological role to promote satietyby activating an ascending visceral sensory circuit ().This circuit originates with vagal afferent neurons thatconvey GI signals to hindbrain areas, including thenucleus of the solitary tract. Some of these hindbrainneurons project to the parabrachial nucleus, which isa central node in this ascending pathway. Of particularrelevance are calcium gene-related peptide expressingneurons in the parabrachial nucleus (CGRPPBN)neurons located within the external lateral subnucleusof the parabrachial nucleus. A variety of stimulilinked to food consumption activate these neurons,such as gastric distention, secretion of cholecystokinin,and glucagon-like peptide . Activation of CGRPPBN

neurons is implicated not only in physiological satietyand normal meal termination, but also in anorexiaelicited by a variety of aversive stimuli. Because hy-pothalamic AgRP neurons inhibit CGRPPBN neurons(), activation of hypothalamic AgRP neurons ap-pears to stimulate feeding, in part, by inhibitingCGRPPBN neurons.

Unlike what is seen with POMC neurons, exper-imental activation of CGRPPBN neurons causes an-orexia that is rapid in onset, severe, and if sustainedcan lead to life-threatening weight loss (). Con-versely, inactivation of CGRPPBN neurons increasesmeal size and blocks the satiating effects of chole-cystokinin and glucagon-like peptide , implicatingthese neurons as physiological mediators of mealtermination (). However, because the increase ofmeal size induced by CGRPPBN neuron inactivation isoffset by a proportionate reduction in the number ofmeals, net food intake does not change. Therefore,normal energy homeostasis does not appear to requireactivation of these neurons, even though meal ter-mination does (). Unlike what is observed in micewith defective melanocortin signaling, CGRPPBN

neuron inactivation also does not increase intakeof a highly palatable diet, nor does it predispose todiet-induced obesity, whereas POMC neurons playa physiological role to limit food intake over long timeintervals. Therefore, CGRPPBN neurons provide animmediate and powerful brake to food consumptionduring individual meals.

Importantly, note that although these observationshighlight relevant recent advances in the neurobiologyof feeding, the substantial complexity inherent in foodintake regulation cannot be reduced to a small set ofinteracting neurocircuits, and much remains to belearned in this field. Adding to this complexity isevidence that some of these neurons can affect feedingin unexpected ways. For example, CGRPPBN neuronsare implicated not only in the perception of satiety butalso in the transmission of aversive experiences thatcan lead to fear conditioning and formation of threatmemory (). It therefore seems somewhat surprisingthat activation of these same neurons during a mealplays a key role in the physiological experience of

satiety. Despite its inherent complexity, further re-search in this area could lead to novel therapeuticagents for obesity.

Developmental considerationsIdentifying the contribution of developmental in-fluences to obesity risk is a daunting challenge because,as noted above, neurons regulating energy homeostasisare distributed throughout the brain (, ), and ourunderstanding of the ontogeny and plasticity of thesecircuits is incomplete. Nevertheless, available evidencesuggests that developmental influences can and docontribute to obesity pathogenesis in adults.

In rodents, circuits regulating distinct aspectsof feeding behavior develop asynchronously. Themost basic types of feeding regulation are presentat birth, whereas the development of progressivelymore complex systems extends into adolescence.The capacity to regulate food intake in response toshort-term signals associated with meal terminationdevelops prior to the maturation of systems governingenergy homeostasis. We can detect autonomic pro-jections at birth that link rodent hypothalamus andbrainstem to the stomach, and these projectionscontinue to increase during the lactation period (,). Food intake is suppressed in response to gastricdistension as early as postnatal day , but it is notinfluenced by postabsorptive nutritional signals fromthe gut until postnatal day to (, ). Ho-meostatic feeding circuits that sense, integrate, andrelay information about the availability of short- andlong-term energy stores develop in the periweaningperiod (–). Projections from ARC neurons thatconvey nutritional (e.g., glucose and fatty acids) andhormonal (e.g., leptin and insulin) signals of energystatus to preautonomic components of the feedingcircuitry form in the third week of life (), and re-sponsiveness to adiposity signals (such as leptin)emerges week after weaning (at weeks of age) ().Finally, processes that control motivated/rewardingaspects of feeding behavior are not established untilpostingestive consequences can be reinforced by theactions of corticolimbic circuits, which mature in thepostweaning period (, ). Although the onset ofcorresponding regulatory networks has yet to beparsed in humans, individual patterns of food intakeare apparently established between and years of age(, ). In both species, therefore, maturation offeeding circuits continues during the transition toindependent feeding of solid foods.

Progress toward an understanding of develop-mental events regulating the maturation of feedingcircuits is hampered by the fact that neurons withopposing effects on food intake are often interspersedwithin the same nucleus (). Furthermore, neuronswith similar peptidergic identities (e.g., NPY, AgRP,GABA, and POMC) can respond differently to thesame hormone and nutrient signals at different de-velopmental stages (), and they can also regulatefeeding via projections to multiple downstream targets(). The origin of ARC feeding circuits involves



http://dx.doi.org/10.1210/er.2017-00111


progenitor cells in the basal aspect of the third ven-tricle that differentiate into immature postmitoticneurons at midgestation in rodents (, ) and by theend of the first month of gestation in primates (). Inmice, most of these immature ARC neurons initiallyexpress the Pomc gene (), but during the course ofgestation and early part of lactation, Pomc expression isgradually extinguished in many of these cells. As thisoccurs, these neurons begin to express the combina-tion of neuropeptides and neurotransmitters thatcomprise the signaling outputs in the adult (e.g., NPY,AgRP, and GABA). In AgRP neurons, this processoccurs progressively, with expression of NPY turningon first, followed by GABA, and finally AgRP. Axonaloutgrowth from ARC neurons to downstream targetsbegins at the end of the first postnatal week and islargely complete by the end of lactation in the thirdpostnatal week (). In nonhuman primates, incomparison, ARC projections develop during the thirdtrimester of gestation (), consistent with the fact thatthe lactation period in rodents corresponds develop-mentally to late gestation in humans (). Conse-quently, many neurodevelopmental processes thatoccur in utero in humans do not take place until afterbirth in rodents.

Researchers have postulated a neurodevelopmentalrole for the surge of plasma leptin levels that occurs inrodents at the end of the first postnatal week ().Although food intake is not sensitive to leptin atthis age, developmental processes (such as axonaloutgrowth) apparently are (). Furthermore, duringlactation, AgRP neurons () require leptin to pro-ject from the ARC to other areas, such as theparaventricular hypothalamic nucleus (, ). Fol-lowing the transition to independent food intake atweaning, AgRP neurons begin to express adenosinetriphosphate–sensitive potassium channels, whichenable a switch in the response to leptin from exci-tation to inhibition (). Presynaptic modulation ofAgRP neuronal activity develops with the sametemporal pattern as the postsynaptic systems outlinedabove, with excitatory inputs to AgRP becoming fullydeveloped by the second postnatal week, whereas thenumber of inhibitory synapses increases after weaning.The onset of homeostatic regulation of feeding co-incides with this final maturation step. These obser-vations raise the possibility of a “developmentalwindow” for the maturation of energy homeostasisneurocircuitry. We clearly need additional work in thisarea to further elucidate these developmental factors.

Signals relevant to the external nutritional envi-ronment (usually transmitted by the mother) caninfluence the maturation of ARC neurons. Duringgestation, for example, the metabolic status of thedam (e.g., the presence of obesity-associated hy-perinsulinemia) can influence the number of ARCneurons that adopt an anorexigenic POMC vs anorexigenic NPY cell fate (), and during lactation,such influences appear to be magnified. For example,differences in the availability or composition of milkduring lactation can affect the onset and the strength

of the pup-derived leptin surge (, ). Exposure tomaternal obesity or overnutrition during lactationcan also reduce the number of neurons that expressleptin receptors, with lasting impacts on leptin re-sponsiveness (, ). Furthermore, the extent ofaxonal outgrowth from ARC neurons to their varioustarget sites appears to be influenced by both milk-derived (insulin) and pup-derived (leptin) hormones(, ). Finally, if undernutrition sufficient to limitgrowth occurs during lactation, the maturation ofsystems that provide presynaptic (GABA) and post-synaptic (adenosine triphosphate–sensitive potassiumchannels) inhibitory signals to AgRP neurons isdelayed (). The existence of multiple steps at whichARC circuits can recalibrate to match the anticipatedexternal environment likely underlies the extraordi-nary capacity of these circuits to compensate for earlydevelopmental deficits ().

Although developmental influences on food in-take regulatory circuits could certainly impact obesitysusceptibility in adults, maternal programming ofobesity susceptibility in rodents appears to arise morefrom reductions of resting () and activity-dependentenergy expenditure () than from persistent effects onfood intake (). To better understand these effects,a brief focus on the ontogeny of autonomic nervoussystem neurocircuits regulating thermogenesis iswarranted. The components of energy expenditurerelevant to this discussion include heat produced ei-ther in the service of thermoregulation or in responseto food consumption (diet-induced thermogenesis).The ontogeny of sympathetic circuits regulating non-shivering thermogenesis hinges on the coordinateddevelopment of both the nervous system and its targetorgan. In both neonatal rodents and humans, theprimary means of heat generation is via activation ofinterscapular brown adipose tissue (iBAT) (–). Inhumans, the iBAT depot appears to be maximallyactive during infancy, before the development ofsystems that increase or decrease core body tem-perature by shivering or sweating, respectively. Inrodents, iBAT remains the primary source of ther-moregulatory heat production throughout the life-span. This distinction between species is critical tointerpreting how rodent studies relate to humanphysiology.

Whereas baselines of activity in feeding circuits arelikely established by to weeks of age in rodents ()and by to years of age in humans (), circuitsregulating resting energy expenditure (also referred toas basal metabolic rate) appear to mature later. Forexample, when a period of caloric restriction is im-posed on obese mice between and weeks of age, theresult is a paradoxical (and maladaptive) increase ofresting energy expenditure (), implying incompletematuration of circuits controlling this response.Consistent with this observation, weight loss in youngchildren who are obese is not necessarily followed bythe compensatory decrease of circulating levels of freetriiodothyronine (the active form of thyroid hormonethat helps to determine basal metabolic rate) that



occurs in adults (). Teenagers appear to be hyper-metabolic compared with adults (), potentiallyowing to higher endogenous brown/beige fat activity(–).

Unlike what occurs in rodents and other smallanimals, iBAT activity appears to decline with age inhumans (, , ). Thermogenic circuits develop inseveral distinct phases (). In both species, the iBATdepot is formed and produces key components of thethermogenic machinery (e.g., uncoupling protein )during gestation, but it is not active until after birth.An immature phase of brown adipose tissue (BAT)thermogenesis is induced both by hormones releasedat parturition (e.g., glucocorticoids and prolactin) andby factors released in response to the ketogenic diet oflactation (e.g., FGF) (, ). Hypothalamic neu-rocircuits that project to the brainstem and modulatesympathetic nervous system output regulate themature phase of BAT thermogenesis; certain hor-monal signals (e.g., thyroid hormone) are also re-quired. In rodents, both sympathetic projections andsympathetic nervous system–dependent stimulationof BAT activity develop during the suckling period(–).

Sensitive periods for development of circuits reg-ulating BAT thermogenesis are distinct from thosethat characterize the development of ARC feeding-relevant circuits, in that neither maternal signals ofmetabolic status (, ) nor maternal environmentalexposures () during gestation influence the forma-tion of the iBAT depot or the expression of its as-sociated thermogenic machinery. Nevertheless, leptindeficiency () or severe intrauterine growth retar-dation can cause diminished iBAT capacity in rodents(), and leptin administration early in the postnatalperiod (postnatal days to ) can reverse these effects.Paradoxically, the same treatment leads to impairedBAT thermogenesis and increased susceptibility todiet-induced obesity when applied to wild-type micenourished normally (, ).

Circuits regulating iBAT thermogenesis remainplastic throughout the weaning period. For example,weaning onto a high-fat diet (HFD) or during coldexposure leads to increased catecholaminergic in-nervation of iBAT (, ) with lasting impacts onthermogenic capacity and sensitivity to diet-inducedobesity (-). Environmental programmingof the maximal capacity of iBAT thermogenesis iscorrelated with lasting effects on the number ofiBAT-innervating neurons in the sympatheticganglia (). In addition to upstream regulation ofprocesses controlling sympathetic innervation, iBATactivation itself may alter the expression of neuro-trophic factors that promote the outgrowth or survivalof innervating sympathetic neurons. This notion ispredicated on established evidence that signals fromperipheral targets influence innervation by sensorynerves, and an analogous system operating in ther-mogenic circuits would provide a means of tuningneuronal outputs to match the anticipated need foriBAT thermogenesis.

Integrative physiology of energy homeostasis

Determinants of feeding behaviorIn most mammals, food intake is organized into in-dividual bouts (meals), the frequency and size of whichcan vary greatly to accommodate the needs of theorganism. For example, predatory hunters, such aslions and wolves, may eat only every couple of days,provided that they can eat the entirety of their kill inwhat amounts to a single meal. In contrast, mosthumans eat multiple meals per day, and each mealconstitutes a modest fraction of total daily caloricintake, with the number and size of meals per dayranging widely across populations and cultures. Ul-timately, however, moment-to-moment regulation ofintake serves the larger goal of maintaining adequatefuel stores to support life, and meal size and frequencycan be adjusted so as to meet this goal ().

The identification of neuromolecular mechanismsthat integrate short-term and long-term control offeeding behavior, such that calorie intake preciselymatches energy expenditure over long time intervals,will almost certainly enable better preventive andtherapeutic approaches to obesity. The origins of theflexibility in meal size and frequency that serve thisgoal can be traced to differences in the biologicalunderpinnings of meal initiation and meal cessation.Specifically, although researchers have proposed thatchanges in the internal milieu trigger meal initiation(e.g., a decrease in circulating levels of glucose or anincrease of plasma ghrelin levels) (–), the de-cision to begin a meal is also strongly impacted bya wide range of external variables, including foodavailability and palatability, the cost and risk associatedwith acquiring food (e.g., threats from predators), andso forth.

In contrast to the considerable flexibility inherentin meal onset, meal termination appears to be highlyregulated by postingestive feedback signals andother physiological variables. Accordingly, meal sizeis greatly increased when postingestive feedback isminimized, such as draining ingested food from thestomach (). One implication of this observation isthat meal initiation and its associated consequences(e.g., activation of oral taste receptors) are themselvesinsufficient to elicit meal termination. However,available evidence points to a host of afferent humoraland neural signals arising from the interaction offood with the GI tract as primary determinants of thesize of individual meals (see the earlier section titled“Hindbrain circuits and the parabrachial nucleus”).Both the time since the last meal and the status ofintercurrent fat stores in the body can strongly in-fluence the capacity of these physiologic signals toterminate a meal. Consequently, when energy stores(and therefore leptin levels) are low, these physiologic“satiety signals” are less capable of terminating a meal,thereby resulting in larger meal size (). Such in-teractions among signals arising from the GI tract andthose generated in proportion to stored fuel adjustingestive behavior on a meal-to-meal basis, so as to



http://dx.doi.org/10.1210/er.2017-00111


maintain stable body-fat stores. This process is fun-damental to energy homeostasis, although it is in-completely understood.

Beyond this physiological control system, varia-tions in both the type and amount of food availableand the environment in which it is eaten can alter howmuch food is consumed at a single meal. For example,delivery of the same food in the same form repeatedlycan reduce consumption, a phenomenon known as“sensory-specific satiety” (). A more variable pre-sentation of the same food can reverse this effect, eventhough overall diet composition is unaffected. Evenfactors such as the size of one’s plate, the type ofserving utensils, and the number of people in the roomcan influence the number of calories consumed ina single meal (). What has been harder to de-termine is the extent to which such factors contributeto sustained alteration of overall energy balance. Al-though meal-size variation does not by itself appear tohave much impact on body weight (because changes inmeal frequency can compensate for this to maintainstable caloric intake), few studies have investigatedthe long-term consequences of changing meal sizes.Ongoing investigation into the effects of restrictedfeeding and intermittent fasting paradigms may proveinteresting in this regard.

Determinants of energy expenditureAs noted earlier, in free-living adults, energy intakeand expenditure are tightly coupled over long timeintervals. A mismatch between intake and expenditureof as little as % can result in involuntary changes ofbody weight amounting to several pounds per annum,which (over time) could result in profound obesity.As weight increases, so does total energy expenditure.Consequently, total energy intake must increasegradually over time for a fixed caloric excess to persist.Although energy intake adjusts well to increased en-ergy expenditure, this compensation appears to be lessaccurate at low levels of energy expenditure, favoringweight gain in sedentary individuals (). Similarly,energy expenditure can compensate for a changeof energy intake, although the “coupling” may bestronger when weight is reduced vs when it is in-creased. Therefore, whereas weight loss resulting fromreduced caloric intake is strongly resisted by decreasesof energy expenditure (both during and after weightloss), increases of energy expenditure induced byoverfeeding tend to be more modest and short-lived().

Although the mechanisms underlying these re-sponses (and the metabolic/endocrine connectionslinking intake and expenditure) are of major clinicalimportance, they are incompletely understood. Insedentary individuals, the following factors are theprimary determinants of energy expenditure: cardio-respiratory activity and the maintenance of cellular iongradients (resting energy expenditure, %); the di-gestion and initial distribution of food substrates (%);and both planned or voluntary activity and low-level unplanned physical activity, including fidgeting

(nonresting energy expenditure, %). Overfeedingraises energy expenditure in each of these compartments(due in part due to increases of both thyroid hormonelevels and sympathetic autonomic activity); conversely,weight loss due to imposed caloric restriction reducesenergy expenditure in each compartment.

Net gain in stored energy cannot occur unlessenergy intake exceeds expenditure. Cell-autonomouscharacteristics of adipocytes and skeletal myocytes, thechemical composition of the ingested calories, and thehormonal responses to these factors may (in theory)influence the chemical composition of stored energy(fat or lean mass). Although a low rate of restingenergy expenditure predicts subsequent weight gain insome studies (), and despite growing interest in thecontribution made by brown and beige adipose tissueto energy expenditure in humans (), the majorcause of the energy imbalance implicated in thecurrent obesity epidemic is excessive food intake(relative to energy expenditure) in the context ofsedentary lifestyles. It would be interesting to identifythe different contributions that autonomic activity, thethyroid axis, and brown/beige adipocytes make toresting energy expenditure in individuals who arepreobese. However, it is unlikely that such differences(if they exist) will account for a substantial amount ofrisk variance in comparison with physical activity (forexample). Furthermore, variations in energy intake(driven by a complex mix of endogenous and exog-enous factors described earlier) typically have a muchlarger effect than variations in energy expenditure onoverall energy balance. Another consideration is thattherapeutic interventions that raise energy expendituresufficiently to cause weight loss eventually triggerincreased food intake as a compensatory response. Forthese reasons, the ability to influence and clinicallymanipulate energy intake is the more pressing goalwhere obesity treatment is concerned.

Mechanisms of Obesity Pathogenesis

Genetic factors: evidence for and againstConcordance rates for obesity in studies of both twinpairs and in adopted children suggest that % to %of the risk for obesity is heritable (). Reasonablearguments can be made on evolutionary groundsthat the current Homo sapiens genome is enrichedfor “thrifty” alleles that conserve calories and resistdownward perturbations of weight that would—byvirtue of effects on fat mass—impair reproductiveefficiency (). Others have argued that “predationrelease” (the reduced threat of predation brought on byadvances in social behavior, weapons, and the use offire) enabled more obese and therefore less agilehominids to escape predation, leading to changes inpopulation adiposity as a result of random mutationsand genetic drift (). These are not mutually ex-clusive formulations.

The search for the genetic basis of the apparent“lipostat” for body fat led ultimately to the molecular



cloning of the ob and dbmutations (in genes encodingleptin and the leptin receptor, respectively). These inturn led to the identification of a canonical molecular/cellular signaling pathway: LEP → LEPR → POMC,AgRP → PC → MCR.

With the exception of MCR, obesity-causingcoding mutations in these genes are rare in hu-mans. During the past decades, genome-wide as-sociation studies (GWASs) and exome/genomesequencing studies have identified a large number ofgene variants associated with more prevalent instancesof obesity. It is disappointing that these strategies havebeen able to account for only a small fraction (~% to% of interindividual variation) of the implied geneticrisk variance for obesity. Sample size has been an issuewith most candidate gene studies, but for some genes(e.g., MCR, ADRB, BDNF, and PC) the associationstudies are convincing, especially because hypomor-phic alleles of each of these genes cause obesity in mice.

Since , GWASs of obesity have tested asso-ciations of millions of relatively common (.%) singlenucleotide polymorphisms spaced more or less evenlyacross the genome with obesity-related phenotypessuch as BMI, body-fat content, waist/hip ratio, re-sponse to bariatric surgery, and other phenotypes.Most of the subjects in these studies have been whiteadults. However, some studies included AfricanAmericans, East Asians, and children, and associationspresent in one adult population were generally presentin others, as well as in children.

GWASs have identified many loci of small effectsize. Importantly, note that single nucleotide poly-morphisms themselves simply implicate a geneticinterval and do not necessarily identify the relevantgene or allele, even if the single nucleotide poly-morphism is within the coding region of a gene. Thispoint aside, many of the ~ GWAS-implicated locifor obesity () are preferentially expressed in brain,consistent with the large amount of research sup-porting primacy of the central nervous system in thecontrol of energy homeostasis. Loci associated withbody-fat distribution, in comparison, primarily markgenes implicated in adipose tissue biology.

The weak explanatory/predictive power of thealleles implicated by GWASs in obesity has led to thesuggestion that other (as yet undiscovered) alleles(genetic “dark matter”) exist that are of much lowerfrequency and higher phenotypic impact. Sequencingof the entire exomes of individuals has the potential toaddress this possibility. An alternative explanation isthat individual risk alleles do not act in isolation.Rather, it is the interaction among different risk allelesor between these alleles and environmental factors thatresults in increased obesity risk. Quantifying theseinteractions, however, is a far more daunting challenge(requiring much larger sample sizes) that has yet to bemet effectively ().

Another area that has yet to be explored (mostlydue to difficulties finding suitable subjects) is whetheralleles of genes that protect against obesity existand can be identified. Such genes/alleles would,

presumably, be residue of earlier selection for phe-notypes enabling predation avoidance, as we alluded toabove. To the extent that these are hypomorphic al-leles, the genes could provide attractive targets forinactivating drugs.

Interactions between genes, development,and environmentAlthough genetic factors acting in isolation are un-likely to explain the rapid increase of obesity preva-lence during the past years, it remains quite possiblethat certain genetic factors enhance the risk of obesityconferred by environmental influences in ways thatfavor positive energy balance (higher calorie intake,less physical activity, or both) and/or result in thebiological defense of increased fat mass. The long listof potentially relevant environmental factors includeschanges of diet composition and lifestyle, environ-mental toxins, infections, changes in the microbiome,and many others as well. Superimposed on theseinfluences are the potential roles played by maternalobesity and diabetes. As discussed in the next section,the mechanisms responsible for transmitting sucheffects range from changes in maternal substrateprovision and endocrine factors to effects conveyed byplacental secretions into fetal circulation and effects onmilk provided during lactation. Vertical “transmission”of phenotype in this fashion could exacerbate (ormitigate) shared genetic predispositions betweenmother and offspring while also affecting the phe-notypes of progeny in the absence of primary geneticpredisposition. Factors such as these almost certainlyconfound some of the efforts to quantify genetic riskfor obesity and diabetes.

An example of interactions of genetic predis-position and lifestyle characteristics that influenceobesity risk can be found in how levels of physicalactivity and diet composition strongly influence theimpact of obesity risk alleles of the FTO gene (whichencodes “fat mass and obesity-associated protein”)(). Additionally, single base pair sequence varia-tions in noncoding portions of the first intron of theFTO gene have the strongest association with obesityin human populations yet detected (). The effectsize is not large, but the susceptibility alleles are veryfrequent in the population. There are apparentlyseveral mechanisms by which these noncodingvariants affect obesity risk, mediated by effects onneighboring genes that influence brain developmentand/or function, as well as the development of beigeadipose tissue ().

Much of the sequence variation contributing toobesity risk will also likely be found in noncodingportions of the genome. Unfortunately, our grasp ofthe molecular genetics of noncoding DNA sequencevariation (including epigenetic influences conveyedby these sequences) is insufficient for a clear under-standing of how these factors might relate to humandisease susceptibility.

Related to this issue and to the future identificationof obesity therapeutics is an apparent conceptual bias



http://dx.doi.org/10.1210/er.2017-00111


regarding the biological consequences of sequencevariants implicated in these studies. The canonicalpathway identified above relates primarily to secretedpeptides/neurotransmitters and their cognate recep-tors. Even though interruptions of signaling can causeacute changes of energy expenditure and intake,these pathways also affect hypothalamic structure/connectivity both during development and in post-natal life (). Thus, the consequences of congenital oracquired disruption may be long-lived. A potentiallyimportant example of a complex structure/pathwayexemplifying a likely combination of such effects isthe primary cilium of hypothalamic and other neuronsthat conveys both acute signaling and structuralguidance in the development of circuits affecting foodintake (, ).

These considerations highlight the possibility thatgenes that contribute to obesity susceptibility throughdirect effects on energy intake and expendituremay also influence the response to developmental/environmental factors, such as intrauterine andperinatal exposures to “obesogenic” diets, toxins, andothers. By such mechanisms, genes/alleles not impli-cated in responses to earlier evolutionary factorsmight be implicated in responses to historically recentand novel factors. Future research regarding geneticand environmental/developmental factors that affectobesity will need to consider these pathways andmechanisms.

Role of epigenetic modificationsEpigenetic modification of genes typically involveschanges in how transcriptional complexes accessregulatory elements in the genome and can occurduring development and throughout life. That sub-strates of intermediary metabolism convey someepigenetic modifications implies a sensitivity to nu-tritional status (). Although examples exist inwhich developmental methylation changes are tieddirectly to obesity pathogenesis (, ), such effectsare the exception rather than the rule. Epigeneticchanges can program phenotypes in adulthood byaffecting placental function, fetal growth rate, organfunction, and the expression of metastable orimprinted genes involved in energy balance regulation.

At fertilization, there is a near-global resetting ofthe epigenome that accompanies the process of cel-lular differentiation from totipotent progenitors tospecific cell identities that is mandatory for cell typespecification (). MicroRNAs also play a role inreinforcing cell fate decisions that progressively restrictthe potential cell fates of progenitor cells. One form ofepigenetic modification involves methylation of CpGsites that are established during development, andthese can be inherited via semiconservative replicationin progenitor cells. Once established, they are highlystable in differentiated cells (). Some of these earlymethylation events occur on metastable alleles orimprinted genes, which can impart a lasting impact onnumerous tissues, whereas others are restricted tospecific cell lineages.

That maternal undernutrition increases obesityrisk only during the first trimester is consistent witha role for early epigenetic processes (). Methyldonor supplementation during rodent gestation canreverse adverse metabolic consequences programmedby undernutrition (, ), and methylation statusin the periconception period is particularly sensitiveto undernutrition. This is because high levels ofhomocysteine (due to folate deficiency) suppress theexpression of DNA methylase (a key enzyme formaintaining methylation during mitosis) (), andsupplementation with either folic acid or methyldonors during gestation can reverse these effects (,, ). Studies in both sheep and humans havereported hypomethylation of imprinted genes (IGF)and metastable alleles (POMC) when the subject isseverely undernourished early in gestation, but notwhen undernourished only in late gestation (–).Although early influences on the epigenetic modifi-cation of metastable or imprinted alleles can persist toadulthood, these changes do not correlate with laterobesity risk within groups of similarly exposed in-dividuals (, , ). Genome-wide epigenomicanalysis of children of underweight mothers hasidentified thousands of hypomethylated loci (), andsome EDCs can induce hypomethylation that is re-versible with maternal methyl donor supplementation(). Future studies that clarify whether global hypo-methylation results from impaired placental functionor nutrient transport may help to explain why so manydifferent maternal exposures yield overlapping phe-notypic outcomes.

Although studies have reported associations be-tween adult obesity and methylation marks on can-didate loci in either cord blood or peripheral blood(, –), these findings are often either dis-cordant with one another or inconsistent with dif-ferentially methylated loci identified in a genome-widescreen (or both). Impacts of maternal obesity maytherefore be conveyed via tissue-specific mechanismsthat cannot be assessed in the analyses of globalmethylation patterns in blood samples.

In some brain regions, as well as peripheral tissues(such as skeletal muscle and BAT), site-specificdemethylation/remethylation driven by transcrip-tional or neuronal activity reorganizes the methylome(–). In tissues in which this remodeling isrelatively restricted to the postnatal period, epigeneticmarks reflective of the postnatal environment canpersist to adulthood (). For example, studiesof several animal models of maternal overnutrition/obesity reported CpG hypermethylation and increasedrepressive histone marks on the Pomc locus, whichpersist to adulthood (–). These observations areconsistent with the idea that epigenetic marks on thePomc promoter underlie alterations in hypothalamicfeeding circuits that diminish leptin responsivenessand consequently predispose not only to weight gainbut to biological defense of elevated body weight.However, some studies involving maternal HFDfeeding reported hypomethylation of the Pomc locus,



as well as dopamine- and opioid-related genes (,). Even if we can potentially explain these differ-ences via complex interactions between the effects ofmaternal nutrition on early methylation patterns andthe later influences of the postnatal and postweaningenvironments on the remodeling of these epigeneticmarks (, , ), they nevertheless challenge thereliability of epigenetic marks as a biomarker of futureobesity risk.

Exposure to maternal obesity during gestation isassociated with both an increased risk of obesity inoffspring and decreased methylation of a develop-mental gene (Znf) that promotes adipocyte dif-ferentiation (–). As this effect is associated withenhanced adipogenic potential of white adipose tissue(–), it could conceivably contribute to sub-sequent obesity risk. Although some early epigeneticmarks are retained in adipose tissue, the adipose tissuemethylome is sensitive to changes in diet, exercise, andweight loss throughout life (–), perhaps owingto the many different cell types represented in adiposetissue.

Collectively, this evidence supports the possibilityof a causal link between the stable transmission ofepigenetic marks, parental nutritional status in thepericonception period, and programming of sub-sequent obesity risk. However, this hypothesis has verylimited experimental support and hence only modestpotential to explain the complex biology observedeither in the laboratory or in human populations.Going forward, promising avenues of research includeinterdisciplinary approaches that combine epigeneticand developmental approaches to clarify how specificepigenetic changes influence overall fetal growth andtissue-specific developmental processes. Such studiesare needed to delineate whether these processescontribute to the impact of maternal undernutrition orovernutrition on obesity risk in offspring.

Developmental factors: evidence for and againstDuring sensitive periods of development, ontogenicprocesses in both brain and peripheral organs can bemodified so as to match anticipated environmentalconditions. Although many exposures during devel-opment could potentially predispose to obesity inadulthood, we focus here on two that some researchersthink contribute to the secular trends in obesity: pa-rental obesity and exposure to EDCs.

Roles of parental body weight or dietUndernutrition. The impact of developmental

exposure to reduced maternal food availability hasbeen examined both in animals and in human pop-ulations around the globe (). Although thesestudies often report sustained effects on obesity risk,outcomes tend to be somewhat variable and sensitiveto both the timing of developmental exposure and therelative abundance of food in the postnatal environ-ment. Specifically, early gestational exposure to un-dernutrition followed by an abundant food supply inthe postnatal period is reliably associated with an

increased risk of obesity (, , ), whereasexposure late in gestation or the persistence of limitednutrient availability after birth is often protectiveagainst obesity. These observations are consistent withthe theory that the undernourished fetus experienceschanges in the energy homeostasis system that areadaptive when limited nutrient availability persists, butbecome maladaptive in a nutrient-rich environment(, ). We have not yet identified the mechanismunderlying such a proposed change.

The observation that some of the adverse conse-quences of exposure to gestational undernutrition canbe reversed by treatment with leptin () or folic acid() during the neonatal (but not the peripubertal)period () supports the existence of discrete sensitiveperiods (referred to as “critical windows”) duringwhich influences on developmental processes can havea lasting impact on metabolic disease risk. Althoughthe early postnatal period in rodents likely representsone such window, whether corresponding processesoccur in humans is unclear, especially because manydevelopmental processes that occur postnatally inrodents take place during late gestation in humans(). Although epidemiological studies suggest thatmaternal undernutrition predisposes to obesity inoffspring in humans (), we need additional in-formation to determine whether such a mechanismcontributes to the increased prevalence of humanobesity in recent decades, and, if so, whether there isa critical window during which the energy homeostasissystem is impacted in ways that predispose to obesityin adulthood.

Overnutrition/obesity. Although parentalobesity is associated with an increased obesity risk inoffspring, parsing contributions made by develop-mental exposure vs genetic or environmental factorsis a difficult challenge. Studies in a genetically rela-tively homogeneous population at high risk for obesity(Pima Indians) point to a link between exposure tomaternal (but not paternal) diabetes during gestationand an earlier onset of obesity (), whereas maternalweight loss due to bariatric surgery prevents thetransmission of increased obesity risk (). Combinedwith evidence that the strongest predictor of childhoodobesity is pregravid maternal BMI (, ), thesestudies support the idea that an obese gestational en-vironment programs susceptibility to obesity.

In rodents, exposure throughout gestation andlactation to maternal consumption of an obesogenicHFD is correlated with persistent increases of adiposityin offspring (, ). However, this effect appears tobe explained by exposure specifically during lactation,which leads to increased adiposity, irrespective ofwhether offspring are weaned onto chow or an HFD(, , ). Because the lactation period in rodentscorresponds developmentally to late gestation inhumans, there are very few scenarios in which obesityin human parents would be limited to only earlyor late gestation. Thus, offspring will typically beexposed throughout development. Whether theamount or type of food consumed by the mother has



http://dx.doi.org/10.1210/er.2017-00111


developmental consequences that predispose to obe-sity independently of maternal obesity per se awaitsfurther study.

Modes of transmission. Although initial ef-forts to model developmental exposure were focusedon maternal transmission to progeny (F generation),recent studies have also explored the possibility of bothpaternal (F) and transgenerational (F) transmission.Paternal obesity is reported to impair placental andfetal growth in mice (), but consequences foradiposity in offspring are variable (–). Althoughpaternal exposure to famine has been suggested toprogram increased adiposity in humans (), thecontribution of paternal BMI to later obesity risk inbroad-based population studies is weaker than thematernal contribution (, ), as might be ex-pected. Although transgenerational transmission ofobesity in humans has been suggested—through thepaternal lineage in a largely homogeneous populationof humans () or inbred rodents ()—such effectsare difficult to quantify, owing to the influence of manyother variables that are difficult to control for.

These data collectively support the notion thatdevelopmental exposure to either undernutrition orobesity can increase susceptibility to obesity in off-spring, and that the timing of exposure strongly in-fluences such effects. However, we need more evidenceto conclude whether these developmental exposureshave contributed to the increase of obesity prevalencein human populations in recent decades.

Role of EDCsNumerous studies link exposure to EDCs to a varietyof outcomes of potential relevance to obesity, in-cluding stimulation of adipogenesis and changes ofinsulin secretion, insulin sensitivity, and liver meta-bolism. A recent Endocrine Society Scientific State-ment provides a comprehensive review on this topic(). In many cases, however, the relevance of thesefindings to obesity pathogenesis per se remains un-certain. Our focus here is limited to evidence thatspecifically links EDC exposure to the accumulationand maintenance of excess body fat mass.

That the increase of human exposure to EDCsparallels the rise in obesity rates in the United States(, ) raises the possibility of a causal link betweenthe two. Many of these chemicals are classified asEDCs based on their capacity to mimic or alter re-ceptor signaling by endogenous hormones, includingestrogen, testosterone, and thyroid hormone ().Whereas some EDCs are chemically unstable [e.g.,bisphenol A (BPA) and pthalates], several are highlypersistent in the environment and bioaccumulate.These include brominated flame retardants; poly-chlorinated biphenyls; organotins, such as tributyltin;organochlorine pesticides, such as dichlorodiphenyl-trichloroethane; and perfluorinated chemicals (PFCs)(, ). Although most of the latter chemicals arecurrently banned from use (with the exception offlame retardants), low-level exposure is widespread inhuman populations ().

Maternal exposure to ECDs usually occurs viaingestion of contaminated food or beverages, althoughcontact with personal care items, plastics, or otherproducts that contain these chemicals can also con-tribute. These chemicals can also be transmitted to thefetus across the placenta or to an infant via breast milk(). Compared with adults, relative levels of expo-sure as a function of body weight are higher for fetusesand infants (), and exposure to low, environ-mentally relevant levels can program lasting effects(,). Given that some EDCs can act directly onadipocytes to promote adipogenesis (–), thepossibility that low levels of these chemicals mightprogram increased susceptibility to obesity later in lifehas been raised ().

The potential contribution to obesity risk of de-velopmental exposure to EDCs has been extensivelyreviewed, including two scientific statements pub-lished by the Endocrine Society (the first in andthe second in ) (, ), a workshop sponsoredby National Toxicology Program of the NationalInstitute of Environmental Health Sciences (), andelsewhere in the scientific literature (, , ). Yetunlike the consensus that exists regarding maternalmetabolic influences on obesity risk, links betweenEDC exposure and obesity risk are a focus of ongoingresearch, with many key questions remaining unan-swered. A brief comparison of how concern about theimpact of maternal influences vs EDC exposure onobesity risk first came to light, and how the two re-search areas developed thereafter, offers insight intothe origins of some of this uncertainty.

Compelling support for the involvement of ma-ternal influences in programming metabolic outcomesfirst emerged from epidemiological observations inhumans. These observations led to the development ofanimal models that have largely recapitulated theseobservations across multiple species, thereby creatinga foundation upon which underlying cellular mech-anisms can be investigated (for examples see earliersections on both epigenetics as well as developmentalfactors). In comparison, evidence of a link betweenEDC exposure and obesity risk began with in vitroeffects on master regulators of adipogenesis. Theseobservations generated justifiable concern regardingrisk to human populations and hence led to a searchfor evidence of a causal link through epidemiologicalstudies and animal-based research, with mixed results.To illustrate the challenges inherent in transitioningfrom in vitro studies to animal models that are sur-rogates for human exposure, we focus on two specificEDCs—PFCs and BPAs.

PFCsPFCs are widely used tomake products more resistant tostains, grease, and water. Because they break down in theenvironment very slowly, they tend to bioaccumulateand can therefore persist in human tissues for years(). Some classes of PFCs can bind to and activatethe nuclear receptors peroxisome proliferator–activatedreceptor a and/or peroxisome proliferator–activated



receptor g (a master regulator of adipogenesis), effectsthat can promote adipocyte proliferation and differen-tiation (, , ). These chemicals also have thepotential to alter the methylation status of peroxisomeproliferator–activated receptor g or its target genes inways that promote (or retard) adipocyte differentiation(, ). Other evidence suggests that some PFCs canalter thyroid function (), which can secondarilyimpact both adipocyte biology and energy balance. Fi-nally, some PFCs increase glucocorticoid concentrationsby inhibiting the degrading enzyme b-hydroxysteroiddehydrogenase (), which also has the potential toincrease obesity risk.

Another relevant consideration is that, as notedearlier, obesity risk in adults appears to be increased byany intervention that results in a period of growthrestriction in early development followed by sub-sequent catch-up growth. Relevant to this issue is thatmeta-analysis of studies in rodents () and ninestudies in humans () supports the hypothesis thatdevelopmental exposure to PFCs reduces fetal growth.In some cases, therefore, the obesogenic effects ofPFCs may be secondary to nonspecific effects on fetalgrowth rather than involving direct effects (e.g., onadipogenesis) that predispose to obesity later in life.

BPAAs BPA is used in the production of polycarbonateplastics and epoxy resins, the main exposure route inhumans is via bottles, food-can linings, and foodpackaging. Maternal BPA is transported across theplacenta, with significantly higher concentrations in malefetuses (). Young children can also be exposed throughhuman milk, as well as through bottle-feeding ().

There is little question that, unlike PFCs, BPAdirectly promotes adipogenesis in a peroxisomeproliferator–activated receptor g-independent man-ner (). Although such an effect cannot, in and ofitself, be assumed to cause or predispose to obesity,other potentially obesogenic effects of BPA includeactivation of estrogen receptor a (, ) or in-creased glucocorticoid synthesis (through actionson b-hydroxysteroid dehydrogenase type ) ().BPA-induced estrogen receptor signaling is alsoassociated with increased expression of histone-modifying enzymes that cause global increases inrepressive epigenetic marks (HK trimethylation)(, ) and changes of mitochondrial metabolismresulting from altered DNA methylation (). Thereis therefore little question as to whether BPA can havebiologically relevant and potentially concerning effectsin vivo as well as in vitro.

Where obesity pathogenesis is concerned, how-ever, data from animal models of low-level BPA ex-posure have been inconsistent. Thus, whereas severalstudies reported increased postnatal growth in rodentsexposed to low-dose BPA during development(–), others reported no differences (–)or even reduced postnatal growth rates (,).In , a review of the existing literature sponsoredby the National Toxicology Program identified

methodological and statistical concerns that maycontribute to these discrepancies () and raisedawareness about the potential impact of differences inexperimental design that include diets, sex, species,strain and dose, duration, and route of exposure ().The same organization subsequently convened a fol-low-up workshop in to reassess the state of thefield and highlighted both continuing inconsistenciesin animal data and the need for better metrics ofobesity (fat mass, adipose tissue cellularity, and re-sponse to an HFD challenge), rather than relying solelyon body weight ().

This workshop also recommended the need forcomparable dose-response information across studiesto address peculiarities in the observed relationshipbetween BPA exposure and excessive weight gain. Forexample, lower BPA-exposure levels often producemore profound effects than are observed at higherlevels (, , ), and the dose of BPA ( mg/kg/d) associated with increased adiposity in rats ()elicits reduced adiposity in mice (, ). Furthercomplicating matters are sexually dimorphic effectsof BPA exposure, with increased body weight and liverweight but unchanged adiposity in males, and de-creased body weight, liver weight, and adiposity infemales (). Such effects may reflect actions of BPAon brown adipose tissue and physical activity, as wellas on white adipose tissue adipogenesis ().

Additional concerns pertain to the nonmonotonicdose relationship between effects of BPA on liver vsadipose tissue. Specifically, doses that promote adi-posity typically have no effect on the liver, whereasdoses that impact the liver can have opposite effects onadipose tissue mass (, ). Based on these con-siderations, the consequences of developmental ex-posure to any particular dose of BPA on growthtrajectories, body weight, and fat mass would appear to beinfluenced by the sum of its various effects on a variety oftissues. Metabolic and physiological outcomes can befurther influenced by changes of BPA exposure duringgestation vs during lactation and/or bottle-feeding.

Efforts to extend these findings to the conse-quences of BPA exposure for human obesity risk haveencountered similar inconsistencies. For example,whereas evidence from cross-sectional studies showsa link between urinary BPA concentrations andchildhood adiposity (–), prospective studieshave reported reduced growth at birth or at years ofage (with a stronger effect in girls than boys) (, ),increased adiposity (), or no effect (, ). Incomparison, a link between BPA exposure and be-havioral disturbances in boys is a more consistentfinding in prospective epidemiological studies ().Such effects seem worthy of investigation independentof the impact of BPA exposure on obesity risk.

Modes of transmission of endocrine disruptingchemical effectsIn rodents, maternal exposure to organotins andorganochloride pesticides (dichlorodiphenyltrichloro-ethane and methyloxychlor) is linked to increased



http://dx.doi.org/10.1210/er.2017-00111


obesity risk in the F generation (–), despiteinconsistent effects on F offspring (). These ob-servations raise the possibility of effects transmitted viaepigenetic changes in the germline and transgenera-tional transmission of exposure-specific epigeneticmarks (). Should this interpretation prove correct,a relevant consideration is that most rodent studies oftransgenerational transmission involve intercrosses be-tween offspring with the same developmental exposure,a condition that is less common in humans.

GI factors, bariatric surgery, and the microbiome

Insights from bariatric surgeryThere is little question that bariatric surgical pro-cedures can produce profound and long-lastingeffects on body weight. Chief among these pro-cedures are Roux-en-Y gastric bypass and verticalsleeve gastrectomy, each of which can produceprofound and sustained weight loss that cannot bereliably achieved by other means (). Althoughthese effects have historically been attributed toeither caloric restriction (converting the stomachinto a small pouch) or malabsorption (the loss ofcalories in the feces), neither explanation can ac-count for the aggregate effects of these procedures.Of particular relevance is that patients report beingless (rather than more) hungry following theseprocedures, even in the face of pronounced weightloss. In comparison, weight loss due to restrictedfood access or to conditions causing GI malab-sorption clearly increases hunger, suggesting thatthese bariatric procedures suppress appetite even inthe face of greatly reduced energy stores—ordinarily,a potent stimulus to food intake.

Among plausible mechanisms underlying the un-expected effects of Roux-en-Y gastric bypass andvertical sleeve gastrectomy on appetite is the possibilitythat these procedures alter communication betweenthe GI tract and energy homeostasis neurocircuits(referred to as the “gut–brain axis”). Researchers havesuggested that signals that might contribute to theseeffects include: gastric hormones, such as ghrelin;intestinal hormones, such as glucagon-like peptide-and peptide tyrosine tyrosine; and alterations in thelevel and composition of bile acids and/or the in-testinal microbiome (). These various mechanismsare not mutually exclusive, and each may contribute toeffects on food intake control that result from surgicalalteration of the GI tract. The larger point is that wecannot attribute the effect of bariatric surgery onenergy homeostasis to simple mechanical alterationsof the physical capacity of the GI tract to ingest foodor absorb nutrients. Rather, these bariatric surgeriesappear to lower the defended level of body-fat mass,presumably through effects involving the gut–brainaxis.

This conclusion is consistent with rodent studies inwhich animals subjected to bariatric surgery activelydefend a lower body weight, even when this involvesincreasing food intake (, ). It is tempting to

speculate that this reduction of the defended level ofbody-fat mass reflects a reversal of pathogenic pro-cesses that led to obesity in the first place. However, itis also possible that the original pathological processesare not themselves altered by bariatric procedures, andthat, instead, a host of nonphysiological responsescollectively serve to lower the defended level of bodyweight. In either case, it is evident that the response ofthe gut–brain axis to these procedures can powerfullyimpact not only food intake but also the homeostaticregulation of body-fat mass. Studies that clarify howthis occurs are a high priority, as are studies to identifycell types in the gut that convey these effects to thebrain and the role of the liver as a potential in-termediary in these processes.

The gut microbiome and other GI factorsHow might such information inform our un-derstanding of obesity pathogenesis? The importanceof GI signals in meal termination, the potent effects ofbariatric surgery on some of these signals, and the largesustained reductions in body weight from bariatricsurgery support the hypothesis that obesity patho-genesis involves changes in the secretion or responseto GI signals related to the composition and quantityof food. Yet despite the large number of experimentsmeasuring gut hormone responses to different types ofnutrients in individuals who are obese vs lean, a clearrole for GI factors has yet to emerge, owing to bothbiological variation inherent in these responses andissues related to study design that make it difficultto compare results across studies. Thus, one can findexamples of differences in almost any gut hormoneresponse between individuals who are obese vs leanthat are contradicted by examples where there are nomeaningful differences. That obesity risk alleles haveyet to be directly linked to GI signals by GWAS alsoargues against a causal relationship between GI factorsand obesity pathogenesis ().

The composition of the ~ pounds of gut bacteriain humans has also been linked to obesity risk ().Alterations in diet can profoundly affect the compo-sition of gut bacteria at multiple levels of the GI tract,and obesity itself may also affect the composition ofgut bacteria.

In mouse studies, transferring bacteria into the GItract of germ-free mice causes weight gain, and theeffect is modestly greater when the source is an obesedonor vs bacteria from a lean animal (). These andsimilar observations raise the possibility that the“obese” microbiome is capable of harvesting morerealizable calories from ingested food than is the leanmicrobiome. The problem with this explanation is thatsimply increasing the energy harvested from ingestedfoods should elicit compensatory adjustments else-where in the energy homeostasis system (e.g., reducedenergy intake or increased energy expenditure) thatlimit weight gain, as detailed above. An alternativepossibility is that gut bacteria generate biologicalsignals (such as butyrate and other short-chain fattyacids) that impact the energy homeostasis system, and



that the composition of gut bacteria influences thenature of these signals (, ).

Evidence linking specific gut microbiota to obesityfalls well short of establishing a causal relationship.Indeed, a recent analysis of available literature wasunable to identify a reliable bacterial compositiondifference between humans who are lean and humanswho are obese across different studies and differentpopulations (). Furthermore, a recent clinical trialfound no evidence of a change of energy balance orother metabolic alterations arising from long-termadministration of antibiotics that profoundly im-pacted the gut microbiome (). An additionalconcern is the reliance on germ-free mice as a main-stay of basic research in this area. These animals atbaseline have a lean, hypermetabolic phenotype andare resistant to weight gain when exposed to a HFD;the mechanistic basis for this phenotype remainsuncertain. Although these phenotypic features tend tonormalize following fecal microbiome transfer, thebody weight of the donor does not dramatically affectthe outcome. This observation is consistent with fecaltransfer experiments designed to supplement an or-ganism’s existing microbiome. Whether conducted inrodents or humans, the effect on body weight tends tobe relatively small ().

One other experimental limitation that impacts thelink between obesity and the gut microbiome pertainsto differences between the microbiome of the largeintestine and that of the small intestine. Whereas thedonated microbiome usually is almost exclusivelyfrom fecal (or sometimes cecal) samples, variations inthe microbiome that affect energy balance by influ-encing signals from the GI tract are likely to involveorganisms in the small intestine as well as those inthe large intestine. These considerations highlight themany questions that must be answered before theimpact of the microbiome on obesity pathogenesis andits role in future interventions for obesity preventionor treatment are understood.

Social and economic factorsThere is little question that in the United States obesityrates are linked inversely to socioeconomic status(SES), especially among women (–). Deducingcause and effect is complicated by the difficulty in-herent in controlling for numerous potentially con-founding variables (e.g., genetic or epigenetic factorsand environmental exposures that impact develop-ment). The focus of this section, therefore, is not onwhether obesity is caused by economic insecurity(which cannot be ascertained from the available evi-dence) but rather on the extent to which highlyprevalent social, economic, and cultural conditionsinfluence obesity risk. Once we identify factors thatimpart the greatest obesity risk, future studies canbegin to investigate whether and how they impact theenergy homeostasis system.

Unlike obesity statistics at the national or countylevels, which can obscure social disparities, new geo-localization and mapping approaches (–) are

able to track the patterns of social, racial, and resi-dential segregation and their impact on body weightand health. At the census tract level, obesity prevalencerates can vary from . to ., depending on wherepeople live. Analyses that accounted for variations inresidential property values, education, and incomeshave accounted for % of the variance in census tractobesity rates in Seattle/King County (). Suchgeolocalized data provide a clear link (although notnecessarily causal) between social and economic fac-tors and obesity risk at the level of individualneighborhoods (with the caveats noted above).

Disparities in the types and amounts of foodconsumed are obvious candidates to explain this as-sociation. Low-cost foods are typically highly pro-cessed, composed of refined grains with added sugarsand added fats, inexpensive, and palatable. Such foodscombine low cost with high energy density and highreward value, are readily available in underserved areas(), and tend to be preferentially selected by lowerincome groups (, ). Excess consumption of suchfoods is linked to rising obesity rates (, ), and ithas been suggested that (for some individuals) con-sumption of such foods mitigates the stress of daily life(–), thereby adding to their high inherent re-ward value.

Studies that link diverse aspects of the built en-vironment with diet quality metrics, reported physicalactivity, and weight and health outcomes have iden-tified proximity to supermarkets, grocery stores, andother services, as well as access to parks and otheropportunities for physical activity as independentpredictors of obesity risk (, –). These andother studies (–) collectively identified an as-sociation between lower obesity rates and locationsthat have pedestrian safety; low crime; attractivestreets; well-maintained parks; and homes within closephysical proximity to supermarkets, parks, sidewalkcafes, and landmark buildings. Conversely, locationswith physical disorder; poor sidewalk quality; closeproximity to bars, liquor stores, fast food, and con-venience stores; and the presence of garbage, litter,and graffiti were associated with higher BMI in some(–) but not other studies (–).

Residential property values provide insight intovariations in the wealth of an individual or area; theyalso offer a useful alternative metric with which toassess the relationship between the built environmentand obesity risk (, ). A Seattle-based study thatexamined the associations among perceived measuresof the environment, residential property values, andBMIs () found that for each $, increase inproperty value, obesity prevalence was .% lower(). Factors related to poverty or wealth maytherefore account for at least some of the link betweenthe built environment and obesity risk.

Access to healthy foods is one such measure. Theconcept of a “food desert,” defined as a low-incomearea in which the nearest supermarket is at least mileaway (), has become a focus of public healthpolicies aimed at improving both diet and health (,



http://dx.doi.org/10.1210/er.2017-00111


, ). Studies suggest that BMI tends to be lowerin areas where the consumption of vegetables andfruits is higher (–). Researchers tend to gaugea specific population’s access to healthy food and foodchoices largely in terms of whether their neighborhoodhas a high density of (or long distance to) healthy foodsources (–). Yet the risk of obesity is not reliablyassociated with distances between home and multiplefood sources (). Rather, it is the type of super-market by price that is significantly associated withobesity rates, even after adjusting for the distance tothe store and other sociodemographic and lifestylevariables.

Where obesity risk is concerned, therefore, accessto healthier foods might be associated more witheconomics than with distance from home. This pos-sibility highlights the need to focus beyond neigh-borhood geographic boundaries. One way to do this isto use GPS tracking methods that capture actual foodshopping behaviors (). Whether improved accessof low-income populations to healthy foods will re-duce obesity prevalence is an important unansweredquestion that the food desert formulation does notcompletely address.

Studies in the United States (), United Kingdom(), France (, ), Finland (), Belgium (),Ireland (), and Australia () have reported thatdiet quality is directly linked to SES. Studies on foodpattern modeling suggest that imposing a cost con-straint leads to food choices that are energy-dense butnutrient-poor, similar in composition to diets con-sumed by lower income groups (, ). Recenteconomic analyses of empirical data from the UnitedKingdom provide a relevant example (). In thewake of the economic recession of , more Britishconsumers turned to foods with lower cost per calorie(), and obesity rates in the United Kingdom in-creased dramatically in recent years; today they are thehighest in Europe (). We need additional studies toinvestigate the extent to which economically drivendietary changes contributed to the observed increase ofobesity prevalence.

Overall, these observations support the testablehypothesis that obesity risk increases among lowerincome groups when their food budgets are in-sufficient to ensure access to a healthy diet. Thishypothesis points to the larger question of how suchan effect ultimately causes not only weight gain butalso the biological defense of elevated fat mass—thetwo processes that define the obese state. To reiteratea key point, factors that predispose to positive energybalance alone are insufficient to explain how obesitypersists, because individuals who are obese defendtheir body-fat stores as effectively as do lean in-dividuals (), and switching to a “healthy” diet andlifestyle is insufficient to restore elevated body weightto normal in the vast majority of individuals (). Inaddition to predisposing to weight gain, therefore,dietary, behavioral, and other factors linking SES statusto obesity risk must also affect the energy homeostasissystem in ways that raise the defended level of body

adiposity. Studies that clarify how this occurs are a keypriority for the field.

Diet composition, lifestyle, and obesity riskAlthough an increase of average energy intake relativeto energy expenditure during the past to decadescan be inferred from the mean increase of adult bodyweight (. kg) in the United States (), the relativecontributions of increased energy intake and reducedphysical activity to this increase cannot be known withcertainty, nor can the extent to which energy balance isimpacted by other variables [e.g., sleep deprivation,decreased variation in environmental temperature(owing to heating and air conditioning), drugs causingweight gain, decreased smoking, and possible de-velopmental exposures discussed earlier] (). Al-though data from Swinburn et al. (–) suggestthat increased energy intake is sufficient to account forrecent population increases of body weight withoutinvoking large decreases of physical activity, epide-miological evidence points to a concomitant increaseof sedentary behaviors. For example, Church et al.() reported that energy expenditure related tooccupation has drastically decreased over time, andthat the associated reduction of energy expenditurelikely contributes to the increase of mean body weightin the United States.

Impact of diet composition on obesity riskThe extent to which changes in diet composition drivethe obesity epidemic has been a matter of considerablecontroversy for decades. To what extent does con-sumption of highly processed foods (especially snackfoods) with higher levels of sugars and fats (and rel-atively low fiber) play a role? What about excessconsumption of sugar-sweetened beverages, includingsoda, juice-based drinks, and sports drinks? These arequestions of great public interest, particularly wherediets high in fat vs carbohydrate content are concerned(, ).

Diets of different macronutrient composition cantheoretically affect energy balance by altering overallcaloric intake, energy expenditure, or both. However,available data argue against major effects on accu-mulation of body-fat mass, so long as energy intake isheld constant (discussed further below). For thisreason, the focus of the debate has shifted toward theeffects of diet composition on caloric intake.

Substantial debate surrounds the question ofwhether the effects of dietary lipids to increase bothpalatability and caloric density contribute to theiroverconsumption (, ) or whether increaseddietary carbohydrate content (and the associated in-sulin response, in particular) plays a uniquely im-portant role in obesity pathogenesis (, –). Asnoted earlier, researchers have suggested that dietarycarbohydrates—and refined sugars in particular—increase insulin secretion. This suppresses lipolysis andthe associated release of fatty acids from the adiposetissue while also preferentially directing dietary fattoward storage. Some researchers propose that the



ensuing depletion of circulating fatty acids triggersa state of “cellular starvation” for metabolically activetissues, such as heart, muscle, and liver. Such changesmight then induce both an adaptive decrease of energyexpenditure () and an increase of food intake (,, ) that together conspire to cause obesity.Extending this logic, replacing dietary carbohydratewith fat should reduce insulin secretion while in-creasing fat mobilization and the oxidation of circu-lating free fatty acids, effects that together protectagainst obesity.

A number of observations challenge this premise.For one, despite the known ability of insulin to inhibitadipocyte lipolysis, and despite the fact that in un-controlled diabetes severe insulin deficiency causesunrestrained lipolysis, there is little direct evidence thatinsulin’s antilipolytic action is an independent de-terminant of fat mass. Furthermore, the hyper-insulinemia of obesity is typically associated withnormal or elevated circulating glucose and fatty acidlevels (), a combination inconsistent with a state ofinsulin-mediated cellular starvation.

The foregoing discussion raises the perennialquestion: Is a calorie a calorie? In other words, canconsumption of different types of foods predispose toweight gain independently of the number of caloriesconsumed? To address this question, one study placed volunteers who were obese or overweight on anisocaloric ketogenic diet (% carbohydrate) for weeks.This study revealed a marginal (~ kcal/d) butstatistically significant effect of the ketogenic diet toincrease -hour energy expenditure measured ina respiratory chamber (), but the effect waned overtime and was not associated with significant loss of fatmass. This finding is consistent with other studiesshowing that carbohydrate restriction does not increaseenergy expenditure unless accompanied by an increase inprotein content (as is the case in most low carbohydratediets, but not the study described above) (–).

When calorie intake is held constant, therefore,body-fat accumulation does not appear to be affectedby even very pronounced changes in the amount of fatvs carbohydrate in the diet. With regard to obesity risk,the clear implication is that the impact of a change ofdiet composition is primarily due to the number ofcalories consumed, with changes of energy expendi-ture, fat oxidation, or other aspects of nutrient han-dling being less important. Given that diets high insimple sugars and processed carbohydrates tend tobe calorie-dense, low in satiety-promoting fiber andother nutrients, affordable, widely available, and oftenheavily marketed, it is perhaps not surprising that suchdiets can favor an increase of overall energy intake. Incomparison, mechanisms related to increased insulinsecretion, nutrient partitioning, cellular starvation,or other internal processes do not appear to explainthe putative benefits of low-carbohydrate or low-glycemic index diets. Furthermore, a recent reviewof weight-loss diets () reported that although low-carbohydrate, higher fat diets led to slightly greaterweight loss than did low-fat diets (~ kg), the overall

difference was trivial and does not justify recom-mending one diet over the other for weight-losspurposes. Although low-carbohydrate diets have beensuggested to be helpful—by virtue of increased energyexpenditure—for maintaining reduced body weight(), differences in protein content of the comparisondiets confound this conclusion.

Roles of sedentary behavior, exercise, andnonexercise activity thermogenesisAs mentioned earlier, the increasing global prevalenceof obesity may involve sedentary behavior in additionto changes of food availability and/or diet composi-tion. Analyzing the contribution of physical activity tototal daily energy expenditure and obesity risk isa complex challenge. Although originally describedas a single component of total daily energy expendi-ture (in addition to basal metabolic rate and diet-induced thermogenesis), overall physical activity canbe subdivided into three distinct components: sed-entary behaviors, nonexercise activity thermogenesis(NEAT), and planned/structured activity (exercise).Researchers estimate that during the past yearsreductions in occupational physical activity have de-creased total daily energy expenditure by . cal-ories per day ().

Physical activity is now widely accepted to be themost variable component of daily energy expenditurein people. Surprisingly, however, increased physicalactivity is largely ineffective as a stand-alone weightloss intervention, even though it should promotenegative energy balance as effectively as does dietaryenergy restriction (). The explanation for thisparadox is presumably that among those individualssusceptible to obesity, the energy homeostasis systemcompensates for increased energy expenditure (byincreasing energy intake and thus resisting weigh loss)better than it compensates for increased energy intake(by increasing energy expenditure to protect againstweight gain). Nevertheless, adherents to vigorousphysical activity programs are more likely to keep lostweight off than are those who remain relatively in-active. For example, according to the National WeightControl Registry, individuals who lose a substantialamount of weight and keep it off for an extendedperiod of time, on average, are those who engage in atleast minutes per day of physical activity (). Thiseffect of exercise may involve a partial reversal of theeffects of weight loss to increase muscle work efficiency(). The key point is that although increased physicalactivity has not proven effective as a stand-alonetreatment of obesity, it can help to sustain weightloss achieved by other means.

In addition to structured activity, Levine et al. ()reported that energy expended passively in activities ofdaily living is both a determinant of total daily energyexpenditure and an obesity risk factor. NEAT activitiesare those that occur during normal daily life, ratherthan in structured bouts of exercise () (e.g., sitting;standing; fidgeting; walking; computer work; house-related chores; activities associated with personal



http://dx.doi.org/10.1210/er.2017-00111


hygiene, such as bathing; and occupation-related ac-tivities). Evidence that reduced NEAT contributesindependently to obesity pathogenesis stems fromthe observations that NEAT tends to be lower inindividuals who are obese vs lean, and it does notappear to change in response to either weight gain(in normal weight subjects) or weight loss (in subjectswho are obese) (). With regard to the magnitudeof this effect, a study of individuals conducted inthe late s using a room calorimeter () foundthat energy expenditure attributable to spontaneousphysical activity averaged kcal/d but varied from to kcal/d between individuals. We need ad-ditional research to establish the extent to whichvariations in NEAT play a causative role in obesitypathogenesis.

Other factors

Smoking cessationAnother potentially causative factor in the steadyincrease of obesity prevalence in the United States andother Westernized societies is a substantial concom-itant reduction in rates of cigarette smoking. Smokingcessation is reliably associated with weight gain (),presumably owing to withdrawal of the pharmaco-logical effect of nicotine to suppress food intake andweight gain. Nicotinic acetylcholine receptors arefound on hypothalamic POMC neurons, and activa-tion of these receptors can reduce food intake andbody weight in animal models (). Although re-duced tobacco use can therefore be expected to in-crease the average body weight of a population, obesitynevertheless remains a problem among both currentsmokers and those who have never smoked.

Infectious factorsThe emergence of obesity as a global pandemic hasspurred interest in the hypothesis that one or moreinfectious agents play a causal role. Consistent withthis view, individuals who are obese have a reducedimmune response to some vaccines, raising the pos-sibility that susceptibility to infections could play a rolein the development of obesity. For example, infectionwith the AD- virus is reported to cause adipocyteproliferation and increased body weight in a variety ofpreclinical models, and individuals who are obese havesignificantly higher antibody titers to this virus than dolean individuals (). However, without better evi-dence of a causal relationship between such infectiousagents and human obesity, this mechanism seemsunlikely to be a major factor underlying the currentobesity epidemic. Should such evidence one dayemerge, it would profoundly impact current ap-proaches to obesity treatment and prevention.

Mechanisms for biological defense of elevatedbody-fat massAlthough the many intrinsic (e.g., genetic) andextrinsic (e.g., diet composition, lifestyle, or SES)variables discussed herein can favor positive energy

balance and predispose to weight gain, a fundamentalunanswered question is how elevated body-fat masscomes to be biologically defended in individuals whoare obese. Under certain circumstances (e.g., mutationof genes encoding leptin or POMC) we can predictthat the biologically defended level of body fat willincrease due to the direct, deleterious effects on theenergy homeostasis system. In the vast majority ofindividuals who are obese, however, researchers haveyet to identify a clearly definable energy homeostasisdefect (genetic or otherwise) to explain this phe-nomenon. Part of the experimental problem is thatsubtle differences in energy intake and/or expenditurecan have large effects on adiposity over time, and oncean individual is in weight equilibrium, any etiologicaldifferences are no longer present.

A very modest but persistent energy balancemismatch (~% to % more calories consumed thanexpended per year) can explain the slow but contin-uous accumulation of additional body fat over manyyears that is characteristic of most humans who areobese. Although an acute increase of body-fat mass isoften reversible, incremental, sustained increases ofbody fat typically end up becoming part of the totalbody-fat mass that is biologically defended. It is for thisreason that the weight loss induced by a change of dietor lifestyle (e.g., changes that might be expected toremedy an acquired defect in the energy homeostasissystem arising from a maladaptive diet, sedentarybehavior) is often ultimately regained, even in the faceof adherence to a more healthy diet and lifestyle. Theseobservations imply that although the mechanismunderlying the gradual increase in the defended levelof body fat may have been triggered by one or more ofthe many environmental exposures discussed earlier,simple withdrawal of the offending exposure is un-likely to reverse the increased body fat once it becomesestablished. Instead, the energy homeostasis systemhas been upwardly reset, so that the higher level ofbody fat is relatively resistant to lifestyle interven-tions, similar to a genetically determined increaseof adiposity (although not necessarily by the samemechanisms).

How might such a change in the energy homeo-stasis system be acquired? Although we still wait fordefinitive answers, recent work offers potential in-sights. In many tissues (including liver, skeletal muscle,adipose tissue, and the vasculature) obesity is associ-ated with activation of inflammatory processes markedby the invasion of macrophages or related immunecells and an associated increase in the expression ofproinflammatory cytokines, such as tumor necrosisfactor-a or interleukin- (). Because the onset ofthis inflammation either coincides with obesity onsetor occurs after obesity is established, it is likelya consequence rather than a cause of obesity ().However, the hypothalamus is an exception. Inthis region of the brain inflammation and tissueinjury occur in discrete areas involved in energy ho-meostasis, and this effect is evident before obesitydevelops. When rats or mice are placed on an HFD,



inflammatory markers become detectable in the hy-pothalamic ARC within to hours, well beforebody-fat mass has increased (). Moreover, thisinflammatory response is associated with expansionand activation of hypothalamic glial cells, a processreferred to as “reactive gliosis” (the prototypical brainresponse to neuron injury). Specifically, switchingeither mice or rats from standard chow to a HFD(which predisposes to diet-induced obesity) inducesboth microgliosis and astrogliosis (activation ofmicroglia and astrocytes, respectively) in the ARCwithin week (, ).

Although it is tempting to draw a causal linkbetween the evidence of injury to a key brain area forenergy homeostasis and defense of elevated body-fatmass, we emphasize that neither the causes nor theconsequences of this local hypothalamic reaction toHFD feeding are known. Thus, the idea that thedefense of elevated fat mass results from this hy-pothalamic gliosis (by impairing the capacity of keyneurons to respond to input from leptin and/or otherpertinent humoral/neural signals) is a hypothesis thatneeds further testing. A working model posits that insusceptible animals or humans, an increase in theconsumption of saturated fat () and/or othernutrients induces injury of hypothalamic neuronsinvolved in energy homeostasis, which in turn trig-gers reactive gliosis. Alternatively, the diet switch maydirectly activate local microglia (a macrophage-likecell found only in the central nervous system), a re-sponse that sets off a vicious cycle by causing neuroninjury and triggering more gliosis. In either case, it isnot difficult to envision how this type of local reactionmight dampen the capacity of ARC neurons to re-spond to humoral (leptin) or neural inputs relevantto body weight control and thereby favor the defenseof elevated body-fat mass. It is noteworthy thatstudies have reported radiological evidence of gliosisin the mediobasal hypothalamus in humans who areobese (, ), as well as in rodent models. Whatdistinguishes this type of mechanism from thosediscussed earlier is its potential to account for thedefense of elevated body weight in acquired (envi-ronmental) forms of obesity. The extent to which itdoes so is a key priority for future work.

Concluding Remarks and Future Directions

In closing, we attempt to distill from foregoing sectionsa set of key areas for possible further investigation.

The two distinct components ofobesity pathogenesisTo be viable, theories of obesity pathogenesis mustaccount not only for how excess body fat is acquired,but also for how excess body fat comes to be bi-ologically defended. To date, the preponderance ofresearch has focused on the former. However, wemust consider the possibility that some (perhapseven most) mechanisms underlying weight gain are

distinct from those responsible for the biologicaldefense of excess fat mass. A key question, therefore,is how the energy homeostasis system comes todefend an elevated level of fat mass (analogous tothe defense of elevated blood pressure in patientswith hypertension). Answering this question re-quires an improved understanding of the neuro-molecular elements that underlie a “defended” levelof body fat. What are the molecular/neuroanatomicpredicates that help establish and defend a “setpoint” for adiposity? How do these elements reg-ulate feeding behavior and/or energy expenditure,so as to achieve long-term energy balance? By whatmechanisms is an apparently higher set pointestablished and defended in individuals who areobese?

Given that recovery of lost weight (the normal,physiological response to weight loss irrespective ofone’s starting weight) is the largest single obstacleto effective long-term weight loss, we cannot overstatethe importance of a coherent understanding ofobesity-associated alterations of the energy homeo-stasis system.

Developmental determinants of the biologicallydefended level of body-fat massBy what means do intrauterine, perinatal, and laterdevelopmental processes influence the defendedlevel of body fat or set point? How do these factorsinteract with underlying genetic determinants? Isovernutrition during development associated withpremature maturation of feeding circuits? Can in-sight into key developmental influences be leveragedinto strategies to reset the defended level at a lowervalue? Can these insights enable us to prevent thehomeostatic system from being reset at a higherlevel in the first place? Do these developmentalexposures act directly on energy balance neuro-circuitry, or do such effects occur indirectly asa result of nonspecific changes in rates of placentalor fetal growth that impact adult body weight andcomposition? Answers to these questions will po-tentially help us develop maternal/prenatal in-terventions that could protect against obesity inadulthood.

Interactions between genetics, epigenetics,developmental influences, and the environmentAlthough many studies (e.g., those based on identicaltwins reared apart) identify a major role for heritablefactors as determinants of obesity susceptibility,GWAS data indicate that only a small fraction (~%)of obesity risk is attributable to identifiable allelicvariants. One potential explanation for this discrep-ancy is that nongenetic factors that contribute toheritability, such as epigenetic modification or de-velopmental influences, make a major contribution toobesity risk. Another possibility is that interactionsamong risk alleles themselves and/or between riskalleles, epigenetic factors, and developmental factorsplay a role. Beyond these considerations, each of these



http://dx.doi.org/10.1210/er.2017-00111


potential contributors to obesity risk can interact withenvironmental factors as well. In other words, manyheritable factors may increase obesity risk, not bycausing obesity so much as by conferring susceptibilityto obesogenic environmental factors. It seems plau-sible, for example, that certain obesity risk allelesexpressed in the brain, perhaps through an interactionwith neurodevelopmental consequences of gestationalevents, enable energy homeostasis neurocircuits tobecome reset around a higher level of body fat stores.It is also possible that the likelihood of such anoutcome is maximized by consuming a diet that ishighly palatable, energy dense, and in abundant supply.Establishing suitable experimental models with which totest such concepts is a necessity.

Future directions for EDC researchAlthough available evidence suggests that EDCs canimpact the function of genes important for the controlof energy balance and adipocyte function, humandata and results from in vivo animal studies have yetto clearly demonstrate an increased risk of obesityconferred by developmental EDC exposure. Amongthe confounding factors that may underlie this un-certainty are differential developmental impacts ofEDC exposure on metabolically relevant organs (e.g.,liver and adipose tissue) (, ), variation stem-ming from sex- and dosage-specific effects, and thepotential for combinations of EDCs to cause syner-gistic effects (, , , ). These concernshighlight the need to focus animal research on thoseexposures most closely linked to untoward impacts onchild health. Meta-analyses of prospective epidemio-logical data may ultimately help to identify thosecombinations and doses of EDC exposures that aremost often associated with increased adiposity andthat are observed consistently across species. Studiesthat define conditions in mice that recapitulate themolecular and/or epigenetic signatures of EDC ex-posure in humans will also be important (, ).Given difficulties inherent in obtaining human tissue,the identification of biomarkers for these effects (e.g.,changes of liver function) may also help to accelerateprogress in this field. Lastly, the question of whetherdevelopmental EDC exposure increases obesity riskthrough nonspecific effects on fetal growth (),rather than or in addition to direct effects on energyhomeostasis system components, must be addressed.

Lessons learned from the weight-reduced stateThe weight-reduced state is a distinct metabolic/behavioral condition created when body weight isreduced below its biologically defended level. Studiesduring the past years (, , –) have shownthat maintenance of a reduced body weight (e.g., %or more) is associated with reductions in energy ex-penditure (in part conveyed by increased contractileefficiency of skeletal muscle) due to reduced sympa-thetic autonomic activity and circulating thyroidhormones. Reduced body weight increases hunger aswell, creating a “perfect metabolic storm” for weight

regain (). It also reduces circulating leptin con-centrations in proportion to lost body fat, and (con-sistent with the leptin threshold formulation describedearlier) providing exogenous leptin in doses justsufficient to restore circulating leptin to preweight lossconcentrations relieves some of the autonomic, en-docrine, energy expenditure, and behavioral pheno-types associated with the weight-reduced state (,). Whether such “physiological leptin replacement”strategies can protect against recovery of lost weight isan important unanswered question.

Rather than viewing recovery of lost weight asa therapeutic failure or as evidence of noncompliancewith a prescribed treatment regimen, patients andpractitioners alike should view this phenomenon as anexpected physiological response to weight loss. Thisperspective emphasizes the importance of identifyingstrategies to subvert these responses. It is likely that thepharmacology involved in maintaining the weight-reduced state is different—qualitatively and quanti-tatively—from that relevant to the induction of weightloss. The Food and Drug Administration should allowfor such differences in the licensing of pharmacologicagents for the treatment of obesity.

The gut–brain axisThere are a number of key questions pertaining to thegut microbiome. Does it influence the defended levelof body-fat mass and, if so, to what extent? How domicrobial influences interact with the energy ho-meostasis system? To what extent do host genetic orother variables influence such effects? Is the micro-biome a potential target for obesity treatment? Howdo bariatric surgical procedures, such as Roux-en-Ygastric bypass or vertical sleeve gastrectomy, reduce thebiologically defended level of body-fat stores? Whichspecific signals emanating from the GI tract accountfor this phenomenon?

Identifying the relevant signals and delineatinghow they influence energy homeostasis neurocircuitrywill ultimately inform potential drug-therapy targets.

Dietary influencesGrowing evidence suggests that (for practical pur-poses) the answer to the question, “Is a calorie a cal-orie?” is “yes.” Calories derived from different dietaryconstituents (fats, carbohydrates, and proteins) do notdiffer significantly in their inherent capacity to pro-mote weight gain by affecting energy expenditure ornutrient partitioning, so long as total calorie intake isheld constant (). From this we infer that the effectsof diet composition per se on metabolic variablessuggested to contribute to obesity pathogenesis (e.g.,those related to plasma levels of glucose, insulin, orfree fatty acids; inherent differences in the propensityof adipocytes to store fat; or the gut microbiome) donot play a clinically significant causal role unless theypromote increased calorie intake (). The thera-peutic potential of interventions primarily targetingthese metabolic processes per se, therefore, seemslimited.



Stratification of obesity outcomesGiven that not all individuals who are obese are subjectto the same level of risk of comorbidities, we needimproved strategies for identifying biomarkers pre-dictive of these comorbidities (e.g., diabetes, hyper-tension, dyslipidemia, and cardiovascular disease).Similarly, strategies for identifying both predictors offuture obesity comorbidity risk (to enable effectivepreventive intervention) and responsiveness to a spe-cific therapeutic intervention (to improve outcomes byidentifying what intervention is best suited for eachindividual) are also a high priority.

Unraveling mechanisms linking the environmentto defense of elevated body weightA seemingly unlimited number of environmentalvariables appear to predispose to weight gain. Povertyis a case in point. How does its influence on foodchoices, behavior, and activity translate to increasedobesity risk? How can we best account for the impacton obesity risk of differences in the racial and ethnicmix between low and higher SES communities? Dodevelopmental factors also contribute?

Disarticulating differences in genetic backgroundfrom the impact of diet, lifestyle, and other environ-mental variables is a daunting challenge. That theimpact of low SES on obesity risk is greater amongwomen than men () points to poorly understoodinteractions between environment and biology that

influence obesity risk. Studies that delineate how in-teractions among these factors impact the energyhomeostasis system are critical.

Identifying and mitigating environmentalrisk factorsTaking poverty again as an example, does an identi-fiable set of motivations, attitudes, and beliefs con-tribute to obesity susceptibility? Which specificelements related to SES predispose not only to weightgain but to the biological defense of elevated body-fatstores, and how might they be mitigated? Can suchmitigation efforts reduce obesity risk?

Translating basic science into moreeffective pharmacotherapyCan anatomic and functional imaging of the humanbrain be used, in combination with suitable behavioralmeasures, to translate cognitive, regulatory, and hedonicaspects of human feeding behavior into effective newpharmacologic approaches to the treatment of obesity?An improved understanding of those aspects of ingestivebehavior that are not driven directly by homeostaticmechanisms controlling body weight is critical tostrategies for obesity prevention. Insight gained fromanimal models of these “hedonic” processes mayeventually identify relevant pathways and molecules.However, we will need better behavioral and imagingtools to fully understand these phenotypes in humans.

References1. Centers for Disease Control and Prevention. Defining

adult overweight and obesity. Available at: www.cdc.gov/obesity/adult/defining.html. Accessed 7 June 2017.

2. Bray G. Battle of the Bulge: A History of Obesity Research.Philadelphia, PA: Dorrance Publishing; 2007.

3. Cohen PA, Maller G, DeSouza R, Neal-Kababick J.Presence of banned drugs in dietary supplements fol-lowing FDA recalls. JAMA. 2014;312:1691–1693.

4. Schwartz TB. Henry Harrower and the turbulent begin-nings of endocrinology.Ann InternMed. 1999;131:702–706.

5. Schwartz MW. Can the history of modern endocri-nology shape the future of obesity?Mol Endocrinol. 2015;29:155–157.

6. Cushing H. Disorders of the pituitary gland: retro-spective and prophetic. JAMA. 1921;76:1721–1726.

7. NorbergM, Lindvall K, Jenkins PL, EmmelinM, Lonnberg G,Nafziger AN. Self-rated health does not predict 10-yearweight change amongmiddle-aged adults in a longitudinalpopulation study. BMC Public Health. 2011;11:748.

8. Keys A. Human starvation and its consequences. J AmDiet Assoc. 1946;22:582–587.

9. Leibel RL, Rosenbaum M, Hirsch J. Changes in energyexpenditure resulting from altered body weight. N Engl JMed. 1995;332:621–628.

10. Butryn ML, Webb V, Wadden TA. Behavioral treatmentof obesity. Psychiatr Clin North Am. 2011;34:841–859.

11. Alamuddin N, Wadden TA. Behavioral treatment of thepatient with obesity. Endocrinol Metab Clin North Am.2016;45:565–580.

12. Pasquet P, Apfelbaum M. Recovery of initial bodyweight and composition after long-term massiveoverfeeding in men. Am J Clin Nutr. 1994;60:861–863.

13. Diaz EO, Prentice AM, Goldberg GR, Murgatroyd PR,Coward WA. Metabolic response to experimentaloverfeeding in lean and overweight healthy volunteers.Am J Clin Nutr. 1992;56:641–655.

14. Rosenbaum M, Leibel RL. Physiological adaptations fol-lowing weight reduction. In: Brownell KD, Walsh BT, eds.Eating Disorders and Obesity: A Comprehensive Handbook.3rd ed. New York, NY: Guilford Press; 2017:51–56.

15. Flier JS. Clinical review 94: What’s in a name? In search ofleptin’s physiologic role. J Clin Endocrinol Metab. 1998;83:1407–1413.

16. Farooqi IS, Jebb SA, Langmack G, Lawrence E, CheethamCH,Prentice AM, Hughes IA, McCamishMA, O’Rahilly S. Effectsof recombinant leptin therapy in a child with congenitalleptin deficiency. N Engl J Med. 1999;341:879–884.

17. Farooqi IS, O’Rahilly S. Monogenic human obesitysyndromes. Recent Prog Horm Res. 2004;59:409–424.

18. Myers MG, Jr, Heymsfield SB, Haft C, Kahn BB, LaughlinM, Leibel RL, Tschop MH, Yanovski JA. Challenges andopportunities of defining clinical leptin resistance. CellMetab. 2012;15:150–156.

19. Myers MG, Jr. Leptin keeps working, even in obesity. CellMetab. 2015;21:791–792.

20. Leibel RL. The role of leptin in the control of bodyweight. Nutr Rev. 2002;60(Suppl 10):S15–S19.

21. Zhang Y, Leibel RL. Leptin and body weight. In: BrownellKD, Walsh BT, eds. Eating Disorders and Obesity: AComprehensive Handbook. 3rd ed. New York, NY:Guilford Press; 2017:15–21.

22. Muoio DM, Dohm GL, Fiedorek FT, Jr, Tapscott EB,Coleman RA. Leptin directly alters lipid partitioning inskeletal muscle. Diabetes. 1997;46:1360–1363.

23. Voshol PJ, Jong MC, Dahlmans VE, Kratky D, Levak-Frank S, Zechner R, Romijn JA, Havekes LM. Inmuscle-specific lipoprotein lipase-overexpressingmice, muscle triglyceride content is increasedwithout inhibition of insulin-stimulated whole-bodyand muscle-specific glucose uptake. Diabetes. 2001;50:2585–2590.

24. Shimada M, Shimano H, Gotoda T, Yamamoto K,Kawamura M, Inaba T, Yazaki Y, Yamada N. Over-expression of human lipoprotein lipase in transgenicmice. Resistance to diet-induced hypertriglyceridemiaand hypercholesterolemia. J Biol Chem. 1993;268:17924–17929.

25. Wang H, Eckel RH. Lipoprotein lipase: from gene toobesity. Am J Physiol Endocrinol Metab. 2009;297:E271–E288.

26. Brunzell JD. Familial lipoprotein lipase deficiency. In:Pagon RA, Adam MP, Ardinger HH, Wallace SE,Amemiya A, Bean LJH, Bird TD, Ledbetter N, MeffordHC, Smith RJH, Stephens K, eds. GeneReviews [Internet].Seattle, WA: University of Washington; 1993–2017. 1999Oct 12 [updated 2011 Dec 15].

27. Ludwig DS, Friedman MI. Increasing adiposity: conse-quence or cause of overeating? JAMA. 2014;311:2167–2168.

28. Hall KD, Chen KY, Guo J, Lam YY, Leibel RL, Mayer LES,Reitman ML, Rosenbaum M, Smith SR, Walsh TB,Ravussin E. Energy expenditure and body compositionchanges after an isocaloric ketogenic diet in overweightand obese men. Am J Clin Nutr. 2016;104(2):324–333.

29. Kennedy GC. The hypothalamus and obesity. Proc R SocMed. 1966;59:1276–1277.



http://dx.doi.org/10.1210/er.2017-00111


30. Seeley RJ, Berridge KC. The hunger games. Cell. 2015;160:805–806.

31. Boden G, Chen X, Mozzoli M, Ryan I. Effect of fasting onserum leptin in normal human subjects. J Clin EndocrinolMetab. 1996;81:3419–3423.

32. Krashes MJ, Koda S, Ye C, Rogan SC, Adams AC, CusherDS, Maratos-Flier E, Roth BL, Lowell BB. Rapid, reversibleactivation of AgRP neurons drives feeding behavior inmice. J Clin Invest. 2011;121:1424–1428.

33. Morton GJ, Meek TH, Schwartz MW. Neurobiology offood intake in health and disease. Nat Rev Neurosci.2014;15:367–378.

34. Chen Y, Lin YC, Kuo TW, Knight ZA. Sensory detectionof food rapidly modulates arcuate feeding circuits. Cell.2015;160:829–841.

35. Chen Y, Lin YC, Zimmerman CA, Essner RA, Knight ZA.Hunger neurons drive feeding through a sustained,positive reinforcement signal. eLife. 2016;5:e18640.

36. Betley JN, Xu S, Cao ZF, Gong R, Magnus CJ, Yu Y,Sternson SM. Neurons for hunger and thirst transmita negative-valence teaching signal. Nature. 2015;521:180–185.

37. Schwartz MW, Seeley RJ, Woods SC, Weigle DS,Campfield LA, Burn P, Baskin DG. Leptin increaseshypothalamic pro-opiomelanocortin mRNA expressionin the rostral arcuate nucleus. Diabetes. 1997;46:2119–2123.

38. Xu Y, Jones JE, Lauzon DA, Anderson JG, Balthasar N,Heisler LK, Zinn AR, Lowell BB, Elmquist JK. A serotoninand melanocortin circuit mediates D-fenfluramine an-orexia. J Neurosci. 2010;30:14630–14634.

39. Butler AA, Marks DL, Fan W, Kuhn CM, Bartolome M,Cone RD. Melanocortin-4 receptor is required for acutehomeostatic responses to increased dietary fat. NatNeurosci. 2001;4:605–611.

40. Fenselau H, Campbell JN, Verstegen AM, Madara JC, XuJ, Shah BP, Resch JM, Yang Z, Mandelblat-Cerf Y, LivnehY, Lowell BB. A rapidly acting glutamatergic ARC→PVHsatiety circuit postsynaptically regulated by a-MSH. NatNeurosci. 2017;20:42–51.

41. Grill HJ, Hayes MR. Hindbrain neurons as an essentialhub in the neuroanatomically distributed control ofenergy balance. Cell Metab. 2012;16:296–309.

42. Roman CW, Derkach VA, Palmiter RD. Genetically andfunctionally defined NTS to PBN brain circuits medi-ating anorexia. Nat Commun. 2016;7:11905.

43. Campos CA, Bowen AJ, Schwartz MW, Palmiter RD.Parabrachial CGRP Neurons Control Meal Termination.Cell Metab. 2016;23:811–820.

44. Han S, Soleiman MT, SodenME, Zweifel LS, Palmiter RD.Elucidating an affective pain circuit that creates a threatmemory. Cell. 2015;162:363–374.

45. Rinaman L, Roesch MR, Card JP. Retrograde trans-ynaptic pseudorabies virus infection of central auto-nomic circuits in neonatal rats. Brain Res Dev Brain Res.1999;114:207–216.

46. Rinaman L. Ontogeny of hypothalamic-hindbrainfeeding control circuits. Dev Psychobiol. 2006;48:389–396.

47. Phifer CB, Browde JA, Jr, Hall WG. Ontogeny of glucoseinhibition of independent ingestion in preweanling rats.Brain Res Bull. 1986;17:673–679.

48. Swithers SE, Hall WG. A nutritive control of in-dependent ingestion in rat pups emerges by nine daysof age. Physiol Behav. 1989;46:873–879.

49. Ahima RS, Prabakaran D, Flier JS. Postnatal leptin surgeand regulation of circadian rhythm of leptin by feeding.Implications for energy homeostasis and neuroendo-crine function. J Clin Invest. 1998;101:1020–1027.

50. Mistry AM, Swick A, Romsos DR. Leptin alters meta-bolic rates before acquisition of its anorectic effect indeveloping neonatal mice. Am J Physiol. 1999;277:R742–R747.

51. Ahima RS, Hileman SM. Postnatal regulation of hy-pothalamic neuropeptide expression by leptin: impli-cations for energy balance and body weight regulation.Regul Pept. 2000;92:1–7.

52. Steculorum SM, Bouret SG. Developmental effects ofghrelin. Peptides. 2011;32:2362–2366.

53. Grove KL, Smith MS. Ontogeny of the hypothalamicneuropeptide Y system. Physiol Behav. 2003;79:47–63.

54. Rozin P, Kalat JW. Specific hungers and poison avoid-ance as adaptive specializations of learning. Psychol Rev.1971;78:459–486.

55. Ogawa H, Hasegawa K, Ohgushi M, Murayama N.Changes in properties of neuronal responses in twocortical taste areas in rats of various ages. Neurosci Res.1994;19:407–417.

56. Deheeger M, Akrout M, Bellisle F, Rossignol C, Rolland-Cachera MF. Individual patterns of food intake devel-opment in children: a 10 months to 8 years of agefollow-up study of nutrition and growth. Physiol Behav.1996;59:403–407.

57. Skinner JD, Ziegler P, Pac S, Devaney B. Meal and snackpatterns of infants and toddlers. J Am Diet Assoc. 2004;104(Suppl. 1):65–70.

58. Cone RD, Cowley MA, Butler AA, FanW, Marks DL, LowMJ. The arcuate nucleus as a conduit for diverse signalsrelevant to energy homeostasis. Int J Obes Relat MetabDisord. 2001;25(Suppl 5):S63–S67.

59. Hill JW, Elias CF, Fukuda M, Williams KW, Berglund ED,Holland WL, Cho YR, Chuang JC, Xu Y, Choi M, LauzonD, Lee CE, Coppari R, Richardson JA, Zigman JM, Chua S,Scherer PE, Lowell BB, Bruning JC, Elmquist JK. Directinsulin and leptin action on pro-opiomelanocortinneurons is required for normal glucose homeostasisand fertility. Cell Metab. 2010;11:286–297.

60. Betley JN, Cao ZF, Ritola KD, Sternson SM. Parallel,redundant circuit organization for homeostatic controlof feeding behavior. Cell. 2013;155:1337–1350.

61. Shimada M, Nakamura T. Time of neuron origin inmouse hypothalamic nuclei. Exp Neurol. 1973;41:163–173.

62. Padilla SL, Carmody JS, Zeltser LM. Pomc-expressingprogenitors give rise to antagonistic neuronal pop-ulations in hypothalamic feeding circuits. Nat Med.2010;16:403–405.

63. Grove KL, Grayson BE, Glavas MM, Xiao XQ, Smith MS.Development of metabolic systems. Physiol Behav. 2005;86:646–660.

64. Koutcherov Y, Mai JK, Ashwell KW, Paxinos G. Orga-nization of human hypothalamus in fetal development.J Comp Neurol. 2002;446:301–324.

65. Bouret SG, Draper SJ, Simerly RB. Trophic action ofleptin on hypothalamic neurons that regulate feeding.Science. 2004;304:108–110.

66. Baquero AF, de Solis AJ, Lindsley SR, Kirigiti MA, SmithMS, Cowley MA, Zeltser LM, Grove KL. Developmentalswitch of leptin signaling in arcuate nucleus neurons.J Neurosci. 2014;34:9982–9994.

67. Bouyer K, Simerly RB. Neonatal leptin exposure specifiesinnervation of presympathetic hypothalamic neuronsand improves the metabolic status of leptin-deficientmice. J Neurosci. 2013;33:840–851.

68. Carmody JS, Wan P, Accili D, Zeltser LM, Leibel RL.Respective contributions of maternal insulin resistanceand diet to metabolic and hypothalamic phenotypes ofprogeny. Obesity (Silver Spring). 2011;19(3):492–499.

69. Yura S, Itoh H, Sagawa N, Yamamoto H, Masuzaki H,Nakao K, Kawamura M, Takemura M, Kakui K, OgawaY, Fujii S. Role of premature leptin surge in obesityresulting from intrauterine undernutrition. Cell Metab.2005;1:371–378.

70. Bautista CJ, Boeck L, Larrea F, Nathanielsz PW, Zam-brano E. Effects of a maternal low protein isocaloric dieton milk leptin and progeny serum leptin concentration

and appetitive behavior in the first 21 days of neonatallife in the rat. Pediatr Res. 2008;63:358–363.

71. Bouret SG, Gorski JN, Patterson CM, Chen S, Levin BE,Simerly RB. Hypothalamic neural projections are per-manently disrupted in diet-induced obese rats. CellMetab. 2008;7:179–185.

72. Glavas MM, Kirigiti MA, Xiao XQ, Enriori PJ, Fisher SK,Evans AE, Grayson BE, Cowley MA, Smith MS, Grove KL.Early overnutrition results in early-onset arcuate leptinresistance and increased sensitivity to high-fat diet.Endocrinology. 2010;151:1598–1610.

73. Cottrell EC, Mercer JG, Ozanne SE. Postnatal devel-opment of hypothalamic leptin receptors. Vitam Horm.2010;82:201–217.

74. Vogt MC, Paeger L, Hess S, Steculorum SM, Awazawa M,Hampel B, Neupert S, Nicholls HT, Mauer J, Hausen AC,Predel R, Kloppenburg P, Horvath TL, Bruning JC.Neonatal insulin action impairs hypothalamic neuro-circuit formation in response to maternal high-fatfeeding. Cell. 2014;156(3):495–509.

75. Juan De Solis A, Baquero AF, Bennett CM, Grove KL,Zeltser LM. Postnatal undernutrition delays a key stepin the maturation of hypothalamic feeding circuits.MolMetab. 2016;5:198–209.

76. Luquet S, Perez FA, Hnasko TS, Palmiter RD. NPY/AgRPneurons are essential for feeding in adult mice but canbe ablated in neonates. Science. 2005;310:683–685.

77. Lerea JS, Ring LE, Hassouna R, Chong AC, Szigeti-Buck K,Horvath TL, Zeltser LM. Reducing adiposity in a criticaldevelopmental window has lasting benefits in mice.Endocrinology. 2016;157(2):666–678.

78. Baker MS, Li G, Kohorst JJ, Waterland RA. Fetal growthrestriction promotes physical inactivity and obesity infemale mice. Int J Obes. 2015;39:98–104.

79. Lagisz M, Blair H, Kenyon P, Uller T, Raubenheimer D,Nakagawa S. Little appetite for obesity: meta-analysis ofthe effects of maternal obesogenic diets on offspringfood intake and body mass in rodents. Int J Obes (Lond).2015;39(1):1669–1678.

80. Aherne W, Hull D. Brown adipose tissue and heatproduction in the newborn infant. J Pathol Bacteriol.1966;91:223–234.

81. Heaton JM. The distribution of brown adipose tissue inthe human. J Anat. 1972;112:35–39.

82. Cannon B, Nedergaard J. Brown adipose tissue: functionand physiological significance. Physiol Rev. 2004;84:277–359.

83. Oscai LB, McGarr JA. Evidence that the amount of foodconsumed in early life fixes appetite in the rat. Am JPhysiol. 1978;235:R141–R144.

84. Fisher JO, Cai G, Jaramillo SJ, Cole SA, Comuzzie AG,Butte NF. Heritability of hyperphagic eating behaviorand appetite-related hormones among Hispanic chil-dren. Obesity (Silver Spring). 2007;15:1484–1495.

85. Wolters B, Lass N, Reinehr T. TSH and free tri-iodothyronine concentrations are associated withweight loss in a lifestyle intervention and weight regainafterwards in obese children. Eur J Endocrinol. 2013;168:323–329.

86. Wang Z, Heymsfield SB, Ying Z, Pierson RN, Jr, GallagherD, Gidwani S. A cellular level approach to predictingresting energy expenditure: evaluation of applicability inadolescents. Am J Hum Biol. 2010;22:476–483.

87. Gelfand MJ, O’Hara SM, Curtwright LA, Maclean JR. Pre-medication to block [18F]FDG uptake in the brownadipose tissue of pediatric and adolescent patients.Pediatr Radiol. 2005;35:984–990.

88. Drubach LA, Palmer EL III, Connolly LP, Baker A, Zur-akowski D, Cypess AM. Pediatric brown adipose tissue:detection, epidemiology, and differences from adults.J Pediatr. 2011;159:939–944.

89. Gilsanz V, Smith ML, Goodarzian F, Kim M, Wren TA,Hu HH. Changes in brown adipose tissue in boys and



girls during childhood and puberty. J Pediatr 2012;160(4):604–609.e1.

90. Tanuma Y, Tamamoto M, Ito T, Yokochi C. The oc-currence of brown adipose tissue in perirenal fat inJapanese. Arch Histol Jpn. 1975;38:43–70.

91. Symonds ME, Henderson K, Elvidge L, Bosman C,Sharkey D, Perkins AC, Budge H. Thermal imaging toassess age-related changes of skin temperature withinthe supraclavicular region co-locating with brown ad-ipose tissue in healthy children. J Pediatr. 2012;161:892–898.

92. Zeltser LM. Developmental influences on circuitsprogramming susceptibility to obesity. Front Neuro-endocrinol. 2015;39:17–27.

93. Villarroya F, Vidal-Puig A. Beyond the sympathetic tone:the new brown fat activators. Cell Metab. 2013;17:638–643.

94. Berry DM, Daniel H. Sympathetic nerve development inthe brown adipose tissue of the rat. Can J PhysiolPharmacol. 1970;48:160–168.

95. Xiao XQ, Williams SM, Grayson BE, Glavas MM, CowleyMA, Smith MS, Grove KL. Excess weight gain during theearly postnatal period is associated with permanentreprogramming of brown adipose tissue adaptivethermogenesis. Endocrinology. 2007;148:4150–4159.

96. Hull D, Vinter J. The development of cold-inducedthermogenesis and the structure of brown adipocytemitochondria in genetically-obese (ob/ob) mice. Br JNutr. 1984;52:33–39.

97. Bazin R, Eteve D, Lavau M. Evidence for decreased GDPbinding to brown-adipose-tissue mitochondria of obeseZucker (fa/fa) rats in the very first days of life. Biochem J.1984;221:241–245.

98. Ashwell M, Holt S, Jennings G, Stirling DM, Trayhurn P,York DA. Measurement by radioimmunoassay of themitochondrial uncoupling protein from brown adiposetissue of obese (ob/ob) mice and Zucker (fa/fa) rats atdifferent ages. FEBS Lett. 1985;179:233–237.

99. Felipe A, Villarroya F, Mampel T. Effects of maternalhypocaloric diet feeding on neonatal rat brown adiposetissue. Biol Neonate. 1988;53:105–112.

100. Vickers MH, Gluckman PD, Coveny AH, Hofman PL,Cutfield WS, Gertler A, Breier BH, Harris M. Neonatalleptin treatment reverses developmental programming.Endocrinology. 2005;146:4211–4216.

101. Vickers MH, Gluckman PD, Coveny AH, Hofman PL,Cutfield WS, Gertler A, Breier BH, Harris M. The effect ofneonatal leptin treatment on postnatal weight gain inmale rats is dependent on maternal nutritional statusduring pregnancy. Endocrinology. 2008;149:1906–1913.

102. Ashwell M, Dunnett SB. Fluorescent histochemicaldemonstration of catecholamines in brown adiposetissue from obese (ob/ob) and lean mice acclimated atdifferent temperatures. J Auton Nerv Syst. 1985;14:377–386.

103. Morrison SF, Ramamurthy S, Young JB. Reduced rearingtemperature augments responses in sympathetic out-flow to brown adipose tissue. J Neurosci. 2000;20:9264–9271.

104. Rothwell NJ, Stock MJ. Effects of early overnutrition andundernutrition in rats on the metabolic responses toovernutrition in later life. J Nutr. 1982;112:426–435.

105. Doi K, Kuroshima A. Lasting effect of infantile coldexperience on cold tolerance in adult rats. Jpn J Physiol.1979;29:139–150.

106. Woods SC. The eating paradox: how we tolerate food.Psychol Rev. 1991;98:488–505.

107. Campfield LA, Smith FJ. Blood glucose dynamics andcontrol of meal initiation: a pattern detection andrecognition theory. Physiol Rev. 2003;83:25–58.

108. Cummings DE. Ghrelin and the short- and long-termregulation of appetite and body weight. Physiol Behav.2006;89:71–84.

109. Rasmussen BA, Breen DM, Luo P, Cheung GW, Yang CS,Sun B, Kokorovic A, Rong W, Lam TK. Duodenal ac-tivation of cAMP-dependent protein kinase inducesvagal afferent firing and lowers glucose production inrats. Gastroenterology. 2012;142(4):834–843.e3.

110. Davis JD, Smith GP, Miesner J. Postpyloric stimuli arenecessary for the normal control of meal size in realfeeding and sham feeding rats. Am J Physiol. 1993;265:R888–R895.

111. Woods SC. Gastrointestinal satiety signals I. An over-view of gastrointestinal signals that influence food in-take. Am J Physiol Gastrointest Liver Physiol. 2004;286:G7–G13.

112. Rolls BJ. Sensory-specific satiety. Nutr Rev. 1986;44:93–101.

113. Wansink B. Environmental factors that increase thefood intake and consumption volume of unknowingconsumers. Annu Rev Nutr. 2004;24:455–479.

114. Mayer J, Thomas DW. Regulation of food intake andobesity. Science. 1967;156:328–337.

115. Tataranni PA, Harper IT, Snitker S, Del Parigi A, VozarovaB, Bunt J, Bogardus C, Ravussin E. Body weight gain infree-living Pima Indians: effect of energy intake vs ex-penditure. Int J Obes Relat Metab Disord. 2003;27:1578–1583.

116. Cohen P, Spiegelman BM. Brown and beige fat: mo-lecular parts of a thermogenic machine. Diabetes. 2015;64:2346–2351.

117. O’Rahilly S, Farooqi IS. Human obesity as a heritabledisorder of the central control of energy balance. Int JObes. 2008;32(Suppl 7):S55–S61.

118. Frisch RE. Body fat, menarche, fitness and fertility. HumReprod. 1987;2:521–533.

119. Speakman JR. A nonadaptive scenario explaining thegenetic predisposition to obesity: the “predation re-lease” hypothesis. Cell Metab. 2007;6:5–12.

120. Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR,Powell C, Vedantam S, Buchkovich ML, Yang J, Croteau-Chonka DC, Esko T, Fall T, Ferreira T, Gustafsson S,Kutalik Z, Luan J, Magi R, Randall JC, Winkler TW, WoodAR, Workalemahu T, Faul JD, Smith JA, Hua Zhao J, ZhaoW, Chen J, Fehrmann R, Hedman AK, Karjalainen J,Schmidt EM, Absher D, Amin N, Anderson D, BeekmanM, Bolton JL, Bragg-Gresham JL, Buyske S, Demirkan A,Deng G, Ehret GB, Feenstra B, Feitosa MF, Fischer K, GoelA, Gong J, Jackson AU, Kanoni S, Kleber ME, KristianssonK, Lim U, Lotay V, Mangino M, Mateo Leach I, Medina-Gomez C, Medland SE, Nalls MA, Palmer CD, Pasko D,Pechlivanis S, Peters MJ, Prokopenko I, Shungin D,Stancakova A, Strawbridge RJ, Ju Sung Y, Tanaka T,Teumer A, Trompet S, van der Laan SW, van Setten J,Van Vliet-Ostaptchouk JV, Wang Z, Yengo L, Zhang W,Isaacs A, Albrecht E, Arnlov J, Arscott GM, Attwood AP,Bandinelli S, Barrett A, Bas IN, Bellis C, Bennett AJ, BerneC, Blagieva R, Bluher M, Bohringer S, Bonnycastle LL,Bottcher Y, Boyd HA, Bruinenberg M, Caspersen IH, IdaChen YD, Clarke R, Daw EW, de Craen AJ, Delgado G,Dimitriou M, Doney AS, Eklund N, Estrada K, Eury E,Folkersen L, Fraser RM, Garcia ME, Geller F, Giedraitis V,Gigante B, Go AS, Golay A, Goodall AH, Gordon SD,Gorski M, Grabe HJ, Grallert H, Grammer TB, Grassler J,Gronberg H, Groves CJ, Gusto G, Haessler J, Hall P, HallerT, Hallmans G, Hartman CA, Hassinen M, Hayward C,Heard-Costa NL, Helmer Q, Hengstenberg C, Holmen O,Hottenga JJ, James AL, Jeff JM, Johansson A, Jolley J,Juliusdottir T, Kinnunen L, Koenig W, Koskenvuo M,Kratzer W, Laitinen J, Lamina C, Leander K, Lee NR,Lichtner P, Lind L, Lindstrom J, Sin Lo K, Lobbens S,Lorbeer R, Lu Y, Mach F, Magnusson PK, Mahajan A,McArdle WL, McLachlan S, Menni C, Merger S, MihailovE, Milani L, Moayyeri A, Monda KL, Morken MA, MulasA, Muller G, Muller-Nurasyid M, Musk AW, Nagaraja R,Nothen MM, Nolte IM, Pilz S, Rayner NW, Renstrom F,Rettig R, Ried JS, Ripke S, Robertson NR, Rose LM, Sanna

S, Scharnagl H, Scholtens S, Schumacher FR, Scott WR,Seufferlein T, Shi J, Vernon Smith A, Smolonska J,Stanton AV, Steinthorsdottir V, Stirrups K, StringhamHM, Sundstrom J, Swertz MA, Swift AJ, Syvanen AC, TanST, Tayo BO, Thorand B, Thorleifsson G, Tyrer JP, UhHW, Vandenput L, Verhulst FC, Vermeulen SH, VerweijN, Vonk JM, Waite LL, Warren HR, Waterworth D,Weedon MN, Wilkens LR, Willenborg C, Wilsgaard T,Wojczynski MK, Wong A, Wright AF, Zhang Q. LifeLinesCohort S, Brennan EP, Choi M, Dastani Z, Drong AW,Eriksson P, Franco-Cereceda A, Gadin JR, Gharavi AG,Goddard ME, Handsaker RE, Huang J, Karpe F, Kathir-esan S, Keildson S, Kiryluk K, Kubo M, Lee JY, Liang L,Lifton RP, Ma B, McCarroll SA, McKnight AJ, Min JL,Moffatt MF, Montgomery GW, Murabito JM, NicholsonG, Nyholt DR, Okada Y, Perry JR, Dorajoo R, Reinmaa E,Salem RM, Sandholm N, Scott RA, Stolk L, Takahashi A,Tanaka T, Van’t Hooft FM, Vinkhuyzen AA, Westra HJ,Zheng W, Zondervan KT, Consortium AD, Group A-BW, Consortium CAD, Consortium CK, Glgc, Icbp,Investigators M, Mu TC, Consortium MI, Consortium P,ReproGen C, Consortium G, International Endogene C,Heath AC, Arveiler D, Bakker SJ, Beilby J, Bergman RN,Blangero J, Bovet P, Campbell H, Caulfield MJ, Cesana G,Chakravarti A, Chasman DI, Chines PS, Collins FS,Crawford DC, Cupples LA, Cusi D, Danesh J, de Faire U,den Ruijter HM, Dominiczak AF, Erbel R, Erdmann J,Eriksson JG, Farrall M, Felix SB, Ferrannini E, Ferrieres J,Ford I, Forouhi NG, Forrester T, Franco OH, GansevoortRT, Gejman PV, Gieger C, Gottesman O, Gudnason V,Gyllensten U, Hall AS, Harris TB, Hattersley AT, HicksAA, Hindorff LA, Hingorani AD, Hofman A, Homuth G,Hovingh GK, Humphries SE, Hunt SC, Hypponen E, IlligT, Jacobs KB, Jarvelin MR, Jockel KH, Johansen B, Jou-silahti P, Jukema JW, Jula AM, Kaprio J, Kastelein JJ,Keinanen-Kiukaanniemi SM, Kiemeney LA, Knekt P,Kooner JS, Kooperberg C, Kovacs P, Kraja AT, Kumari M,Kuusisto J, Lakka TA, Langenberg C, Le Marchand L,Lehtimaki T, Lyssenko V, Mannisto S, Marette A, MatiseTC, McKenzie CA, McKnight B, Moll FL, Morris AD,Morris AP, Murray JC, Nelis M, Ohlsson C, OldehinkelAJ, Ong KK, Madden PA, Pasterkamp G, Peden JF, PetersA, Postma DS, Pramstaller PP, Price JF, Qi L, Raitakari OT,Rankinen T, Rao DC, Rice TK, Ridker PM, Rioux JD,Ritchie MD, Rudan I, Salomaa V, Samani NJ, Saramies J,Sarzynski MA, Schunkert H, Schwarz PE, Sever P,Shuldiner AR, Sinisalo J, Stolk RP, Strauch K, Tonjes A,Tregouet DA, Tremblay A, Tremoli E, Virtamo J, VohlMC, Volker U, Waeber G, Willemsen G, Witteman JC,Zillikens MC, Adair LS, Amouyel P, Asselbergs FW,Assimes TL, Bochud M, Boehm BO, Boerwinkle E,Bornstein SR, Bottinger EP, Bouchard C, Cauchi S,Chambers JC, Chanock SJ, Cooper RS, de Bakker PI,Dedoussis G, Ferrucci L, Franks PW, Froguel P, Groop LC,Haiman CA, Hamsten A, Hui J, Hunter DJ, Hveem K,Kaplan RC, Kivimaki M, Kuh D, Laakso M, Liu Y, MartinNG, Marz W, Melbye M, Metspalu A, Moebus S,Munroe PB, Njolstad I, Oostra BA, Palmer CN, PedersenNL, Perola M, Perusse L, Peters U, Power C, QuertermousT, Rauramaa R, Rivadeneira F, Saaristo TE, Saleheen D,Sattar N, Schadt EE, Schlessinger D, Slagboom PE,Snieder H, Spector TD, Thorsteinsdottir U, Stumvoll M,Tuomilehto J, Uitterlinden AG, Uusitupa M, van derHarst P, Walker M, Wallaschofski H, Wareham NJ,Watkins H, Weir DR, Wichmann HE, Wilson JF, Zanen P,Borecki IB, Deloukas P, Fox CS, Heid IM, O’Connell JR,Strachan DP, Stefansson K, van Duijn CM, Abecasis GR,Franke L, Frayling TM, McCarthy MI, Visscher PM,Scherag A, Willer CJ, Boehnke M, Mohlke KL, LindgrenCM, Beckmann JS, Barroso I, North KE, Ingelsson E,Hirschhorn JN, Loos RJ, Speliotes EK. Genetic studies ofbody mass index yield new insights for obesity biology.Nature. 2015;518:197–206.



http://dx.doi.org/10.1210/er.2017-00111


121. Loos RJF, Leibel RL. Genetics of obesity and related traits.In: Brownell KD, Walsh BT, eds. Eating Disorders andObesity: A Comprehensive Handbook. 3rd ed. New York,NY: Guilford Press; 2017:22–30.

122. Kim JY, DeMenna JT, Puppala S, Chittoor G, Schneider J,Duggirala R, Mandarino LJ, Shaibi GQ, Coletta DK.Physical activity and FTO genotype by physical activityinteractive influences on obesity. BMC Genet. 2016;17:47.

123. Yeo GS. The role of the FTO (fat mass and obesityrelated) locus in regulating body size and composition.Mol Cell Endocrinol. 2014;397:34–41.

124. Claussnitzer M, Dankel SN, Kim KH, Quon G, Meule-man W, Haugen C, Glunk V, Sousa IS, Beaudry JL,Puviindran V, Abdennur NA, Liu J, Svensson PA, Hsu YH,Drucker DJ, Mellgren G, Hui CC, Hauner H, Kellis M. FTOobesity variant circuitry and adipocyte browning inhumans. N Engl J Med. 2015;373:895–907.

125. Oh EC, Vasanth S, Katsanis N. Metabolic regulation andenergy homeostasis through the primary cilium. CellMetab. 2015;21:21–31.

126. Vaisse C, Reiter JF, Berbari NF. Cilia and obesity. ColdSpring Harb Perspect Biol. 2017:a028217.

127. Kaelin WG, Jr, McKnight SL. Influence of metabolism onepigenetics and disease. Cell. 2013;153:56–69.

128. Waterland RA, Jirtle RL. Transposable elements: targetsfor early nutritional effects on epigenetic gene regula-tion. Mol Cell Biol. 2003;23:5293–5300.

129. Goldstone AP. Prader-Willi syndrome: advances in ge-netics, pathophysiology and treatment. Trends Endo-crinol Metab. 2004;15:12–20.

130. Reik W. Stability and flexibility of epigenetic generegulation in mammalian development. Nature. 2007;447:425–432.

131. Cedar H, Bergman Y. Epigenetic silencing during earlylineage commitment. In: StemBook [Internet]. Cam-bridge, MA: Harvard Stem Cell Institute; 2008–. doi:10.3824/stembook.1.42.1.

132. Ravelli GP, Stein ZA, Susser MW. Obesity in young menafter famine exposure in utero and early infancy. N Engl JMed. 1976;295:349–353.

133. Lillycrop KA, Phillips ES, Jackson AA, Hanson MA,Burdge GC. Dietary protein restriction of pregnant ratsinduces and folic acid supplementation prevents epi-genetic modification of hepatic gene expression in theoffspring. J Nutr. 2005;135:1382–1386.

134. Sharp GC, Lawlor DA, Richmond RC, Fraser A, SimpkinA, Suderman M, Shihab HA, Lyttleton O, McArdle W,Ring SM, Gaunt TR, Davey Smith G, Relton CL. Maternalpre-pregnancy BMI and gestational weight gain, off-spring DNA methylation and later offspring adiposity:findings from the Avon Longitudinal Study of Parentsand Children. Int J Epidemiol. 2015;44:1288–1304.

135. Lillycrop KA, Burdge GC. Epigenetic changes in early lifeand future risk of obesity. Int J Obes. 2011;35:72–83.

136. Waterland RA, Travisano M, Tahiliani KG. Diet-inducedhypermethylation at agouti viable yellow is notinherited transgenerationally through the female. FASEBJ. 2007;21:3380–3385.

137. Tobi EW, Lumey LH, Talens RP, Kremer D, Putter H,Stein AD, Slagboom PE, Heijmans BT. DNA methylationdifferences after exposure to prenatal famine arecommon and timing- and sex-specific. Hum Mol Genet.2009;18:4046–4053.

138. Heijmans BT, Tobi EW, Stein AD, Putter H, Blauw GJ,Susser ES, Slagboom PE, Lumey LH. Persistent epigeneticdifferences associated with prenatal exposure to faminein humans. Proc Natl Acad Sci USA. 2008;105:17046–17049.

139. Stevens A, Begum G, Cook A, Connor K, Rumball C,Oliver M, Challis J, Bloomfield F, White A. Epigeneticchanges in the hypothalamic proopiomelanocortin andglucocorticoid receptor genes in the ovine fetus after

periconceptional undernutrition. Endocrinology. 2010;151:3652–3664.

140. Souren NY, Tierling S, Fryns JP, Derom C, Walter J,Zeegers MP. DNA methylation variability at growth-related imprints does not contribute to overweight inmonozygotic twins discordant for BMI. Obesity (SilverSpring). 2011;19:1519–1522.

141. Dolinoy DC, Huang D, Jirtle RL. Maternal nutrientsupplementation counteracts bisphenol A-inducedDNA hypomethylation in early development. Proc NatlAcad Sci USA. 2007;104:13056–13061.

142. Gemma C, Sookoian S, Alvarinas J, Garcia SI, Quintana L,Kanevsky D, Gonzalez CD, Pirola CJ. Maternal pre-gestational BMI is associated with methylation of thePPARGC1A promoter in newborns. Obesity (SilverSpring). 2009;17:1032–1039.

143. Morales E, Groom A, Lawlor DA, Relton CL. DNAmethylation signatures in cord blood associated withmaternal gestational weight gain: results from theALSPAC cohort. BMC Res Notes. 2014;7:278.

144. Godfrey KM, Sheppard A, Gluckman PD, Lillycrop KA,Burdge GC, McLean C, Rodford J, Slater-Jefferies JL,Garratt E, Crozier SR, Emerald BS, Gale CR, Inskip HM,Cooper C, Hanson MA. Epigenetic gene promotermethylation at birth is associated with child’s lateradiposity. Diabetes. 2011;60:1528–1534.

145. Huang RC, Garratt ES, Pan H, Wu Y, Davis EA, Barton SJ,Burdge GC, Godfrey KM, Holbrook JD, Lillycrop KA.Genome-wide methylation analysis identifies differen-tially methylated CpG loci associated with severeobesity in childhood. Epigenetics. 2015;10:995–1005.

146. Tateishi K, Okada Y, Kallin EM, Zhang Y. Role of Jhdm2ain regulating metabolic gene expression and obesityresistance. Nature. 2009;458:757–761.

147. Caldji C, Hellstrom IC, Zhang TY, Diorio J, Meaney MJ.Environmental regulation of the neural epigenome.FEBS Lett. 2011;585:2049–2058.

148. Guo JU, Ma DK, Mo H, Ball MP, Jang MH, Bonaguidi MA,Balazer JA, Eaves HL, Xie B, Ford E, Zhang K, Ming GL,Gao Y, Song H. Neuronal activity modifies the DNAmethylation landscape in the adult brain. Nat Neurosci.2011;14:1345–1351.

149. Plagemann A, Harder T, Brunn M, Harder A, Roepke K,Wittrock-Staar M, Ziska T, Schellong K, Rodekamp E,Melchior K, Dudenhausen JW. Hypothalamic proo-piomelanocortin promoter methylation becomes al-tered by early overfeeding: an epigenetic model ofobesity and the metabolic syndrome. J Physiol. 2009;587:4963–4976.

150. Cho CE. Role of methyl group vitamins in hypothalamicdevelopment of food intake regulation in Wistar rats.Appl Physiol Nutr Metab. 2014;39:844.

151. Marco A, Kisliouk T, Tabachnik T, Meiri N, Weller A.Overweight and CpG methylation of the Pomc pro-moter in offspring of high-fat-diet-fed dams are not“reprogrammed” by regular chow diet in rats. FASEB J.2014;28:4148–4157.

152. Marco A, Kisliouk T, Tabachnik T, Weller A, Meiri N.DNA CpG methylation (5-methylcytosine) and itsderivative (5-hydroxymethylcytosine) alter histoneposttranslational modifications at the Pomc promoter,affecting the impact of perinatal diet on leanness andobesity of the offspring. Diabetes. 2016;65(8):2258–2267.

153. Vucetic Z, Kimmel J, Totoki K, Hollenbeck E, Reyes TM.Maternal high-fat diet alters methylation and geneexpression of dopamine and opioid-related genes. En-docrinology. 2010;151:4756–4764.

154. Carlin J, George R, Reyes TM. Methyl donor supple-mentation blocks the adverse effects of maternal highfat diet on offspring physiology. PLoS One. 2013;8:e63549.

155. Marco A, Kisliouk T, Weller A, Meiri N. High fat dietinduces hypermethylation of the hypothalamic Pomc

promoter and obesity in post-weaning rats. Psycho-neuroendocrinology. 2013;38:2844–2853.

156. Borengasser SJ, Zhong Y, Kang P, Lindsey F, Ronis MJ,Badger TM, Gomez-Acevedo H, Shankar K. Maternalobesity enhances white adipose tissue differentiationand alters genome-scale DNA methylation in male ratoffspring. Endocrinology. 2013;154:4113–4125.

157. Yang QY, Liang JF, Rogers CJ, Zhao JX, Zhu MJ, Du M.Maternal obesity induces epigenetic modifications tofacilitate Zfp423 expression and enhance adipogenicdifferentiation in fetal mice. Diabetes. 2013;62:3727–3735.

158. Desai M, Beall M, Ross MG. Developmental origins ofobesity: programmed adipogenesis. Curr Diab Rep. 2013;13:27–33.

159. Bouchard L, Rabasa-Lhoret R, Faraj M, Lavoie ME, Mill J,Perusse L, Vohl MC. Differential epigenomic andtranscriptomic responses in subcutaneous adiposetissue between low and high responders to caloricrestriction. Am J Clin Nutr. 2010;91:309–320.

160. Barres R, Yan J, Egan B, Treebak JT, Rasmussen M, Fritz T,Caidahl K, Krook A, O’Gorman DJ, Zierath JR. Acuteexercise remodels promoter methylation in humanskeletal muscle. Cell Metab. 2012;15:405–411.

161. Multhaup ML, Seldin MM, Jaffe AE, Lei X, Kirchner H,Mondal P, Li Y, Rodriguez V, Drong A, Hussain M,Lindgren C, McCarthy M, Naslund E, Zierath JR, WongGW, Feinberg AP. Mouse-human experimental epige-netic analysis unmasks dietary targets and genetic lia-bility for diabetic phenotypes. Cell Metab. 2015;21:138–149.

162. Lumey LH, Stein AD, Susser E. Prenatal famine and adulthealth. Annu Rev Public Health. 2011;32:237–262.

163. Ravelli ACJ, van der Meulen JHP, Michels RPJ, OsmondC, Barker DJP, Hales CN, Bleker OP. Glucose tolerance inadults after prenatal exposure to famine. Lancet. 1998;351:173–177.

164. Painter RC, Roseboom TJ, Bleker OP. Prenatal exposureto the Dutch famine and disease in later life: anoverview. Reprod Toxicol. 2005;20:345–352.

165. Hales CN, Barker DJ. The thrifty phenotype hypothesis.Br Med Bull. 2001;60:5–20.

166. Gluckman PD, Hanson MA. Living with the past:evolution, development, and patterns of disease. Sci-ence. 2004;305:1733–1736.

167. Burdge GC, Lillycrop KA, Phillips ES, Slater-Jefferies JL,Jackson AA, Hanson MA. Folic acid supplementationduring the juvenile-pubertal period in rats modifies thephenotype and epigenotype induced by prenatal nu-trition. J Nutr. 2009;139:1054–1060.

168. Dabelea D, Hanson RL, Lindsay RS, Pettitt DJ, ImperatoreG, Gabir MM, Roumain J, Bennett PH, Knowler WC.Intrauterine exposure to diabetes conveys risks for type2 diabetes and obesity: a study of discordant sibships.Diabetes. 2000;49:2208–2211.

169. Kral JG, Biron S, Simard S, Hould FS, Lebel S, Marceau S,Marceau P. Large maternal weight loss from obesitysurgery prevents transmission of obesity to childrenwho were followed for 2 to 18 years. Pediatrics. 2006;118:e1644–e1649.

170. Catalano PM, Farrell K, Thomas A, Huston-Presley L,Mencin P, de Mouzon SH, Amini SB. Perinatal riskfactors for childhood obesity and metabolic dysregu-lation. Am J Clin Nutr. 2009;90:1303–1313.

171. Guo F, Jen KL. High-fat feeding during pregnancy andlactation affects offspring metabolism in rats. PhysiolBehav. 1995;57:681–686.

172. Bayol SA, Simbi BH, Stickland NC. A maternal cafeteriadiet during gestation and lactation promotes adiposityand impairs skeletal muscle development and meta-bolism in rat offspring at weaning. J Physiol. 2005;567:951–961.

173. Gorski JN, Dunn-Meynell AA, Hartman TG, Levin BE.Postnatal environment overrides genetic and prenatal



factors influencing offspring obesity and insulin re-sistance. Am J Physiol Regul Integr Comp Physiol. 2006;291:R768–R778.

174. McPherson NO, Bell VG, Zander-Fox DL, Fullston T, WuLL, Robker RL, Lane M. When two obese parents areworse than one! Impacts on embryo and fetal devel-opment. Am J Physiol Endocrinol Metab. 2015;309:E568–E581.

175. Ng SF, Lin RC, Laybutt DR, Barres R, Owens JA, MorrisMJ. Chronic high-fat diet in fathers programs b-celldysfunction in female rat offspring. Nature. 2010;467:963–966.

176. Fullston T, Ohlsson Teague EM, Palmer NO, DeBlasioMJ, Mitchell M, Corbett M, Print CG, Owens JA, Lane M.Paternal obesity initiates metabolic disturbances in twogenerations of mice with incomplete penetrance to theF2 generation and alters the transcriptional profile oftestis and sperm microRNA content. FASEB J. 2013;27:4226–4243.

177. Huypens P, Sass S, Wu M, Dyckhoff D, Tschop M, TheisF, Marschall S, Hrabe de Angelis M, Beckers J. Epigeneticgermline inheritance of diet-induced obesity and insulinresistance. Nat Genet. 2016;48:497–499.

178. Veenendaal MV, Painter RC, de Rooij SR, Bossuyt PM,van der Post JA, Gluckman PD, Hanson MA, RoseboomTJ. Transgenerational effects of prenatal exposure to the1944–45 Dutch famine. BJOG. 2013;120:548–554.

179. Pembrey ME, Bygren LO, Kaati G, Edvinsson S,Northstone K, Sjostrom M, Golding J, Team AS. Sex-specific, male-line transgenerational responses inhumans. Eur J Hum Genet. 2006;14:159–166.

180. Gore AC, Chappell VA, Fenton SE, Flaws JA, Nadal A,Prins GS, Toppari J, Zoeller RT. EDC-2: the EndocrineSociety’s Second Scientific Statement on endocrine-disrupting chemicals. Endocr Rev. 2015;36:E1–E150.

181. Casals-Casas C, Desvergne B. Endocrine disruptors: fromendocrine to metabolic disruption. Annu Rev Physiol.2011;73:135–162.

182. Baillie-Hamilton PF. Chemical toxins: a hypothesis toexplain the global obesity epidemic. J Altern Comple-ment Med. 2002;8:185–192.

183. Bergman A, Heindel JJ, Kasten T, Kidd KA, Jobling S,Neira M, Zoeller RT, Becher G, Bjerregaard P, Bornman R,Brandt I, Kortenkamp A, Muir D, Drisse MN, Ochieng R,Skakkebaek NE, Bylehn AS, Iguchi T, Toppari J, WoodruffTJ. The impact of endocrine disruption: a consensusstatement on the state of the science. Environ HealthPerspect. 2013;121:A104–A106.

184. Heindel JJ, Newbold R, Schug TT. Endocrine disruptorsand obesity. Nat Rev Endocrinol. 2015;11:653–661.

185. World Health Organization. State of the science ofendocrine disrupting chemicals—2012. Available at:www.who.int/ceh/publications/endocrine/en/. Accessed15 March 2017.

186. Chapin RE, Adams J, Boekelheide K, Gray LE, Jr, HaywardSW, Lees PS, McIntyre BS, Portier KM, Schnorr TM,Selevan SG, Vandenbergh JG, Woskie SR. NTP-CERHRexpert panel report on the reproductive and de-velopmental toxicity of bisphenol A. Birth Defects Res BDev Reprod Toxicol. 2008;83:157–395.

187. Diamanti-Kandarakis E, Bourguignon JP, Giudice LC,Hauser R, Prins GS, Soto AM, Zoeller RT, Gore AC.Endocrine-disrupting chemicals: an Endocrine SocietyScientific Statement. Endocr Rev. 2009;30:293–342.

188. Grun F, Blumberg B. Environmental obesogens: orga-notins and endocrine disruption via nuclear receptorsignaling. Endocrinology. 2006;147:S50–S55.

189. Vanden Heuvel JP, Thompson JT, Frame SR, Gillies PJ.Differential activation of nuclear receptors by per-fluorinated fatty acid analogs and natural fatty acids:a comparison of human, mouse, and rat peroxisomeproliferator-activated receptor-a, -b, and -g, liver Xreceptor-b, and retinoid X receptor-a. Toxicol Sci. 2006;92:476–489.

190. Taxvig C, Dreisig K, Boberg J, Nellemann C, Schelde AB,Pedersen D, Boergesen M, Mandrup S, Vinggaard AM.Differential effects of environmental chemicals andfood contaminants on adipogenesis, biomarker releaseand PPARg activation. Mol Cell Endocrinol. 2012;361:106–115.

191. Grun F, Blumberg B. Perturbed nuclear receptor sig-naling by environmental obesogens as emerging factorsin the obesity crisis. Rev Endocr Metab Disord. 2007;8:161–171.

192. Thayer KA, Heindel JJ, Bucher JR, Gallo MA. Role ofenvironmental chemicals in diabetes and obesity:a National Toxicology Program workshop review. En-viron Health Perspect. 2012;120:779–789.

193. de Cock M, van de Bor M. Obesogenic effects of en-docrine disruptors, what do we know from animal andhuman studies? Environ Int. 2014;70:15–24.

194. Braun JM. Early-life exposure to EDCs: role in childhoodobesity and neurodevelopment. Nat Rev Endocrinol.2017;13:161–173.

195. Buck RC, Franklin J, Berger U, Conder JM, Cousins IT, deVoogt P, Jensen AA, Kannan K, Mabury SA, vanLeeuwen SP. Perfluoroalkyl and polyfluoroalkyl sub-stances in the environment: terminology, classification,and origins. Integr Environ Assess Manag. 2011;7:513–541.

196. Bastos Sales L, Kamstra JH, Cenijn PH, van Rijt LS,Hamers T, Legler J. Effects of endocrine disruptingchemicals on in vitro global DNA methylation andadipocyte differentiation. Toxicol In Vitro. 2013;27:1634–1643.

197. Kirchner S, Kieu T, Chow C, Casey S, Blumberg B.Prenatal exposure to the environmental obesogentributyltin predisposes multipotent stem cells to be-come adipocytes. Mol Endocrinol. 2010;24:526–539.

198. Boas M, Feldt-Rasmussen U, Main KM. Thyroid effectsof endocrine disrupting chemicals. Mol Cell Endocrinol.2012;355:240–248.

199. Ye L, Guo J, Ge RS. Environmental pollutants andhydroxysteroid dehydrogenases. Vitam Horm. 2014;94:349–390.

200. Koustas E, Lam J, Sutton P, Johnson PI, Atchley DS, Sen S,Robinson KA, Axelrad DA, Woodruff TJ. The NavigationGuide—evidence-based medicine meets environmen-tal health: systematic review of nonhuman evidence forPFOA effects on fetal growth. Environ Health Perspect.2014;122:1015–1027.

201. Johnson PI, Sutton P, Atchley DS, Koustas E, Lam J, Sen S,Robinson KA, Axelrad DA, Woodruff TJ. The NavigationGuide—evidence-based medicine meets environmen-tal health: systematic review of human evidence forPFOA effects on fetal growth. Environ Health Perspect.2014;122:1028–1039.

202. Chamorro-Garcia R, Kirchner S, Li X, Janesick A, CaseySC, Chow C, Blumberg B. Bisphenol A diglycidyl etherinduces adipogenic differentiation of multipotentstromal stem cells through a peroxisome proliferator-activated receptor gamma-independent mechanism.Environ Health Perspect. 2012;120:984–989.

203. Ozgyin L, Erdos E, Bojcsuk D, Balint BL. Nuclear receptorsin transgenerational epigenetic inheritance. Prog Bio-phys Mol Biol. 2015;118:34–43.

204. Li Y, Luh CJ, Burns KA, Arao Y, Jiang Z, Teng CT, Tice RR,Korach KS. Endocrine-disrupting chemicals (EDCs):in vitro mechanism of estrogenic activation and dif-ferential effects on ER target genes. Environ HealthPerspect. 2013;121:459–466.

205. Wang J, Sun B, Hou M, Pan X, Li X. The environmentalobesogen bisphenol A promotes adipogenesis by in-creasing the amount of 11b-hydroxysteroid de-hydrogenase type 1 in the adipose tissue of children. IntJ Obes. 2013;37:999–1005.

206. Bhan A, Hussain I, Ansari KI, Bobzean SA, Perrotti LI,Mandal SS. Histone methyltransferase EZH2 is

transcriptionally induced by estradiol as well as estro-genic endocrine disruptors bisphenol-A and di-ethylstilbestrol. J Mol Biol. 2014;426:3426–3441.

207. Doherty LF, Bromer JG, Zhou Y, Aldad TS, Taylor HS. Inutero exposure to diethylstilbestrol (DES) or bisphenol-A (BPA) increases EZH2 expression in the mammarygland: an epigenetic mechanism linking endocrinedisruptors to breast cancer. Horm Cancer. 2010;1:146–155.

208. Strakovsky RS, Wang H, Engeseth NJ, Flaws JA, HelferichWG, Pan YX, Lezmi S. Developmental bisphenol A(BPA) exposure leads to sex-specific modification ofhepatic gene expression and epigenome at birth thatmay exacerbate high-fat diet-induced hepatic steatosis.Toxicol Appl Pharmacol. 2015;284:101–112.

209. Howdeshell KL, Hotchkiss AK, Thayer KA, VandenberghJG, vom Saal FS. Exposure to bisphenol A advancespuberty. Nature. 1999;401:763–764.

210. Rubin BS, Murray MK, Damassa DA, King JC, Soto AM.Perinatal exposure to low doses of bisphenol A affectsbody weight, patterns of estrous cyclicity, and plasmaLH levels. Environ Health Perspect. 2001;109:675–680.

211. Miyakawa K, Ryo A, Murakami T, Ohba K, Yamaoka S,Fukuda M, Guatelli J, Yamamoto N. BCA2/Rabring7promotes tetherin-dependent HIV-1 restriction. PLoSPathog. 2009;5:e1000700.

212. Akingbemi BT, Sottas CM, Koulova AI, Klinefelter GR,Hardy MP. Inhibition of testicular steroidogenesis by thexenoestrogen bisphenol A is associated with reducedpituitary luteinizing hormone secretion and decreasedsteroidogenic enzyme gene expression in rat Leydigcells. Endocrinology. 2004;145:592–603.

213. Moriyama K, Tagami T, Akamizu T, Usui T, Saijo M,Kanamoto N, Hataya Y, Shimatsu A, Kuzuya H, Nakao K.Thyroid hormone action is disrupted by bisphenol A asan antagonist. J Clin Endocrinol Metab. 2002;87:5185–5190.

214. Kubo K, Arai O, Omura M, Watanabe R, Ogata R, Aou S.Low dose effects of bisphenol A on sexual differenti-ation of the brain and behavior in rats. Neurosci Res.2003;45:345–356.

215. Durando M, Kass L, Piva J, Sonnenschein C, Soto AM,Luque EH, Munoz-de-Toro M. Prenatal bisphenol Aexposure induces preneoplastic lesions in the mammarygland in Wistar rats. Environ Health Perspect. 2007;115:80–86.

216. Honma S, Suzuki A, Buchanan DL, Katsu Y, Watanabe H,Iguchi T. Low dose effect of in utero exposure tobisphenol A and diethylstilbestrol on female mousereproduction. Reprod Toxicol. 2002;16:117–122.

217. Stel J, Legler J. The role of epigenetics in the latent effectsof early life exposure to obesogenic endocrine dis-rupting chemicals. Endocrinology. 2015;156:3466–3472.

218. Vandenberg LN, Colborn T, Hayes TB, Heindel JJ, JacobsDR, Jr, Lee DH, Shioda T, Soto AM, vom Saal FS,Welshons WV, Zoeller RT, Myers JP. Hormones andendocrine-disrupting chemicals: low-dose effects andnonmonotonic dose responses. Endocr Rev. 2012;33:378–455.

219. Zoeller RT, Brown TR, Doan LL, Gore AC, SkakkebaekNE, Soto AM, Woodruff TJ, Vom Saal FS. Endocrine-disrupting chemicals and public health protection:a statement of principles from the Endocrine Society.Endocrinology. 2012;153:4097–4110.

220. Wei J, Lin Y, Li Y, Ying C, Chen J, Song L, Zhou Z, Lv Z, XiaW, Chen X, Xu S. Perinatal exposure to bisphenol A atreference dose predisposes offspring to metabolicsyndrome in adult rats on a high-fat diet. Endocrinology.2011;152:3049–3061.

221. Angle BM, Do RP, Ponzi D, Stahlhut RW, Drury BE, NagelSC, Welshons WV, Besch-Williford CL, Palanza P, Par-migiani S, vom Saal FS, Taylor JA. Metabolic disruptionin male mice due to fetal exposure to low but not highdoses of bisphenol A (BPA): evidence for effects on



http://dx.doi.org/10.1210/er.2017-00111


body weight, food intake, adipocytes, leptin, adipo-nectin, insulin and glucose regulation. Reprod Toxicol.2013;42:256–268.

222. Marmugi A, Ducheix S, Lasserre F, Polizzi A, Paris A,Priymenko N, Bertrand-Michel J, Pineau T, Guillou H,Martin PG, Mselli-Lakhal L. Low doses of bisphenol Ainduce gene expression related to lipid synthesis andtrigger triglyceride accumulation in adult mouse liver.Hepatology. 2012;55:395–407.

223. van Esterik JC, Dolle ME, Lamoree MH, van Leeuwen SP,Hamers T, Legler J, van der Ven LT. Programming ofmetabolic effects in C57BL/6J3FVB mice by exposureto bisphenol A during gestation and lactation. Toxi-cology. 2014;321:40–52.

224. Trasande L, Attina TM, Blustein J. Association betweenurinary bisphenol A concentration and obesity preva-lence in children and adolescents. JAMA. 2012;308:1113–1121.

225. Harley KG, Aguilar Schall R, Chevrier J, Tyler K, Aguirre H,Bradman A, Holland NT, Lustig RH, Calafat AM, EskenaziB. Prenatal and postnatal bisphenol A exposure andbody mass index in childhood in the CHAMACOScohort. Environ Health Perspect. 2013;121:514–520.

226. Vafeiadi M, Roumeliotaki T, Myridakis A, Chalkiadaki G,Fthenou E, Dermitzaki E, Karachaliou M, Sarri K, VassilakiM, Stephanou EG, Kogevinas M, Chatzi L. Association ofearly life exposure to bisphenol A with obesity andcardiometabolic traits in childhood. Environ Res. 2016;146:379–387.

227. Veiga-Lopez A, Kannan K, Liao C, Ye W, Domino SE,Padmanabhan V. Gender-specific effects on gestationallength and birth weight by early pregnancy BPA ex-posure. J Clin Endocrinol Metab. 2015;100:E1394–E1403.

228. Braun JM, Lanphear BP, Calafat AM, Deria S, Khoury J,Howe CJ, Venners SA. Early-life bisphenol a exposureand child body mass index: a prospective cohort study.Environ Health Perspect. 2014;122:1239–1245.

229. Hoepner LA, Whyatt RM, Widen EM, Hassoun A,Oberfield SE, Mueller NT, Diaz D, Calafat AM, Perera FP,Rundle AG. Bisphenol A and adiposity in an inner-citybirth cohort. Environ Health Perspect. 2016;124:1644–1650.

230. Chamorro-Garcıa R, Sahu M, Abbey RJ, Laude J, Pham N,Blumberg B. Transgenerational inheritance of increasedfat depot size, stem cell reprogramming, and hepaticsteatosis elicited by prenatal exposure to the obesogentributyltin in mice. Environ Health Perspect. 2013;121:359–366.

231. Skinner MK, Manikkam M, Tracey R, Guerrero-BosagnaC, Haque M, Nilsson EE. Ancestral dichlorodiphenyl-trichloroethane (DDT) exposure promotes epigenetictransgenerational inheritance of obesity. BMC Med.2013;11:228.

232. ManikkamM, Haque MM, Guerrero-Bosagna C, NilssonEE, Skinner MK. Pesticide methoxychlor promotes theepigenetic transgenerational inheritance of adult-onsetdisease through the female germline. PLoS One. 2014;9:e102091.

233. Skinner MK. Endocrine disruptor induction of epige-netic transgenerational inheritance of disease. Mol CellEndocrinol. 2014;398:4–12.

234. Scheen AJ, Letiexhe M, Rorive M, De Flines J, Luyckx FH,Desaive C. Bariatric surgery: 10-year results of theSwedish Obese Subjects Study [in French]. Rev MedLiege. 2005;60:121–125.

235. Seeley RJ, Chambers AP, Sandoval DA. The role of gutadaptation in the potent effects of multiple bariatricsurgeries on obesity and diabetes. Cell Metab. 2015;21:369–378.

236. Hao Z, Mumphrey MB, Townsend RL, Morrison CD,Munzberg H, Ye J, Berthoud HR. Reprogramming ofdefended body weight after Roux-En-Y gastric bypasssurgery in diet-induced obese mice. Obesity (SilverSpring). 2016;24:654–660.

237. Stefater MA, Perez-Tilve D, Chambers AP, Wilson-PerezHE, Sandoval DA, Berger J, Toure M, Tschop M, WoodsSC, Seeley RJ. Sleeve gastrectomy induces loss of weightand fat mass in obese rats, but does not affect leptinsensitivity. Gastroenterology 2010;138(7):2426–2436.

238. Sommer F, Backhed F. The gut microbiota—masters ofhost development and physiology. Nat Rev Microbiol.2013;11:227–238.

239. Turnbaugh PJ, Gordon JI. The core gut microbiome,energy balance and obesity. J Physiol. 2009;587:4153–4158.

240. Rosenbaum M, Knight R, Leibel RL. The gut microbiotain human energy homeostasis and obesity. TrendsEndocrinol Metab. 2015;26:493–501.

241. Sze MA, Schloss PD. Looking for a signal in the noise:revisiting obesity and the microbiome. MBio. 2016;7(4):e01018-16.

242. Mikkelsen KH, Frost M, Bahl MI, Licht TR, Jensen US,Rosenberg J, Pedersen O, Hansen T, Rehfeld JF, Holst JJ,Vilsboll T, Knop FK. Effect of antibiotics on gutmicrobiota, gut hormones and glucose metabolism.PLoS One. 2015;10:e0142352.

243. Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, MardisER, Gordon JI. An obesity-associated gut microbiomewith increased capacity for energy harvest. Nature. 2006;444:1027–1031.

244. Sobal J, Stunkard AJ. Socioeconomic status and obesity:a review of the literature. Psychol Bull. 1989;105:260–275.

245. McLaren L. Socioeconomic status and obesity. EpidemiolRev. 2007;29:29–48.

246. Drewnowski A. Obesity, diets, and social inequalities.Nutr Rev. 2009;67(Suppl 1):S36–S39.

247. Ogden CL, Lamb MM, Carroll MD, Flegal KM. Obesityand socioeconomic status in adults: United States,2005–2008. NCHS Data Brief. 2010;(50):1–8.

248. Lebel A, Kestens Y, Clary C, Bisset S, Subramanian SV.Geographic variability in the association between so-cioeconomic status and BMI in the USA and Canada.PLoS One. 2014;9:e99158.

249. Michimi A, Wimberly MC. Associations of supermarketaccessibility with obesity and fruit and vegetableconsumption in the conterminous United States. Int JHealth Geogr. 2010;9:49.

250. Gustafson AA, Sharkey J, Samuel-Hodge CD, Jones-Smith J, Folds MC, Cai J, Ammerman AS. Perceived andobjective measures of the food store environment andthe association with weight and diet among low-income women in North Carolina. Public Health Nutr.2011;14:1032–1038.

251. Jennings A, Welch A, Jones AP, Harrison F, Bentham G,van Sluijs EMF, Griffin SJ, Cassidy A. Availability of localfood outlets is associated with weight status and dietaryintake in 9–10 year olds. Am J Prev Med. 2011;40:405–410.

252. Caspi CE, Sorensen G, Subramanian SV, Kawachi I. Thelocal food environment and diet: a systematic review.Health Place. 2012;18:1172–1187.

253. Drewnowski A, Aggarwal A, Cook A, Stewart O,Moudon AV. Geographic disparities in Healthy EatingIndex scores (HEI–2005 and 2010) by residentialproperty values: findings from Seattle Obesity Study(SOS). Prev Med. 2016;83:46–55.

254. Drewnowski A, Rehm CD, Arterburn D. The geographicdistribution of obesity by census tract among 59 767insured adults in King County, WA. Int J Obes. 2005;2014(38):833–839.

255. Drewnowski A, Specter S. Poverty and obesity: the roleof energy density and energy costs. Am J Clin Nutr. 2004;79:6–16.

256. Darmon N, Drewnowski A. Does social class predict dietquality? Am J Clin Nutr. 2008;87:1107–1117.

257. Drewnowski A. The carbohydrate-fat problem: can weconstruct a healthy diet based on dietary guidelines?

Advances in Nutrition: An International Review Journal.2015;6:318S–325S.

258. Malik VS, Schulze MB, Hu FB. Intake of sugar-sweetenedbeverages and weight gain: a systematic review. Am JClin Nutr. 2006;84:274–288.

259. Drewnowski A. The real contribution of added sugarsand fats to obesity. Epidemiol Rev. 2007;29:160–171.

260. Oliver G, Wardle J, Gibson EL. Stress and food choice:a laboratory study. Psychosom Med. 2000;62:853–865.

261. Zellner DA, Loaiza S, Gonzalez Z, Pita J, Morales J, PecoraD, Wolf A. Food selection changes under stress. PhysiolBehav. 2006;87:789–793.

262. Torres SJ, Nowson CA. Relationship between stress,eating behavior, and obesity. Nutrition. 2007;23:887–894.

263. Yau YHC, Potenza MN. Stress and eating behaviors.Minerva Endocrinol. 2013;38:255–267.

264. Duncan MJ, Spence JC, Mummery WK. Perceived en-vironment and physical activity: a meta-analysis ofselected environmental characteristics. Int J Behav NutrPhys Act. 2005;2:1–9.

265. Ding D, Gebel K. Built environment, physical activity,and obesity: what have we learned from reviewing theliterature? Health Place. 2012;18:100–105.

266. Rundle A, Quinn J, Lovasi G, Bader M, Yousefzadeh P,Weiss C, Neckerman K. Associations between bodymass index and park proximity, size, cleanliness andrecreational facilities. Am J Health Promot. 2013;27:262–269.

267. Bell JF, Wilson JS, Liu GC. Neighborhood greenness and2-year changes in body mass index of children andyouth. Am J Prev Med. 2008;35:547–553.

268. Lovasi GS, Bader MD, Quinn J, Neckerman K, Weiss C,Rundle A. Body mass index, safety hazards, andneighborhood attractiveness. Am J Prev Med. 2012;43:378–384.

269. Ellaway A, Macintyre S, Bonnefoy X. Graffiti, greenery,and obesity in adults: secondary analysis of Europeancross sectional survey. BMJ. 2005;331:611–612.

270. Boehmer TK, Hoehner CM, Deshpande AD, BrennanRamirez LK, Brownson RC. Perceived and observedneighborhood indicators of obesity among urbanadults. Int J Obes. 2007;31:968–977.

271. Burdette AM, Hill TD. An examination of processeslinking perceived neighborhood disorder and obesity.Soc Sci Med. 2008;67:38–46.

272. Duncan DT, Castro MC, Gortmaker SL, Aldstadt J, MellySJ, Bennett GG. Racial differences in the builtenvironment–body mass index relationship? A geo-spatial analysis of adolescents in urban neighborhoods.Int J Health Geogr. 2012;11:11.

273. Fiechtner L, Block J, Duncan DT, Gillman MW, Gort-maker SL, Melly SJ, Rifas-Shiman SL, Taveras EM.Proximity to supermarkets associated with higher bodymass index among overweight and obese preschool-agechildren. Prev Med. 2013;56:218–221.

274. Hattori A, An R, Sturm R. Neighborhood food outlets,diet, and obesity among California adults, 2007 and2009. Prev Chronic Dis. 2013;10:1223.

275. Fone DL, Dunstan F, Christie S, Jones A, West J, WebberM, Lester N, Watkins J. Council tax valuation bands,socio-economic status and health outcome: a cross-sectional analysis from the Caerphilly Health and SocialNeeds Study. BMC Public Health. 2006;6:115.

276. Connolly S, O’Reilly D, Rosato M. House value as anindicator of cumulative wealth is strongly related tomorbidity and mortality risk in older people: a census-based cross-sectional and longitudinal study. Int J Epi-demiol. 2010;39:383–391.

277. Drewnowski A, Aggarwal A, Rehm CD, Cohen-Cline H,Hurvitz PM, Moudon AV. Environments perceived asobesogenic have lower residential property values. Am JPrev Med. 2014;47:260–274.

278. Khan LK, Sobush K, Keener D, Goodman K, Lowry A,Kakietek J, Zaro S. Recommended community strategies



and measurements to prevent obesity in the United

States. MMWR Recomm Rep. 2009;58:1–26.279. Powell LM, Auld MC, Chaloupka FJ, O’Malley PM,

Johnston LD. Associations between access to food

stores and adolescent body mass index. Am J Prev Med.

2007;33:S301–S307.280. Larson NI, Story MT, Nelson MC. Neighborhood en-

vironments: disparities in access to healthy foods in the

U.S. Am J Prev Med. 2009;36:74–81.281. Drewnowski A, Aggarwal A, Hurvitz PM, Monsivais P,

Moudon AV. Obesity and supermarket access: prox-

imity or price? Am J Public Health. 2012;102:e74–e80.282. Moore LV, Diez Roux AV, Nettleton JA, Jacobs DR.

Associations of the local food environment with diet

quality—a comparison of assessments based on surveys

and geographic information systems: the multi-ethnic

study of atherosclerosis. Am J Epidemiol. 2008;167:917–924.

283. Wang MC, Cubbin C, Ahn D, Winkleby MA. Changes in

neighbourhood food store environment, food behav-

iour and body mass index, 1981–1990. Public Health

Nutr. 2008;11:963–970.284. Jago R, Baranowski T, Baranowski JC, Cullen KW,

Thompson D. Distance to food stores & adolescent

male fruit and vegetable consumption: mediation ef-

fects. Int J Behav Nutr Phys Act. 2007;4:35.285. Laska MN, Hearst MO, Forsyth A, Pasch KE, Lytle L.

Neighbourhood food environments: are they associated

with adolescent dietary intake, food purchases and

weight status? Public Health Nutr. 2010;13:1757–1763.286. Aggarwal A, Monsivais P, Cook AJ, Drewnowski A.

Positive attitude toward healthy eating predicts higher

diet quality at all cost levels of supermarkets. J Acad Nutr

Diet. 2014;114:266–272.287. Rao M, Afshin A, Singh G, Mozaffarian D. Do healthier

foods and diet patterns cost more than less healthy

options? A systematic review and meta-analysis. BMJ

Open. 2013;3:e004277.288. Northstone K, Emmett PM. Dietary patterns of men in

ALSPAC: associations with socio-demographic and

lifestyle characteristics, nutrient intake and comparison

with women’s dietary patterns. Eur J Clin Nutr. 2010;64:978–986.

289. Estaquio C, Druesne-Pecollo N, Latino-Martel P, Dau-

chet L, Hercberg S, Bertrais S. Socioeconomic differences

in fruit and vegetable consumption among middle-

aged French adults: adherence to the 5 A Day rec-

ommendation. J Am Diet Assoc. 2008;108:2021–2030.290. Malon A, Deschamps V, Salanave B, Vernay M, Szego E,

Estaquio C, Kesse-Guyot E, Hercberg S, Castetbon K.

Compliance with French nutrition and health program

recommendations is strongly associated with socio-

economic characteristics in the general adult pop-

ulation. J Am Diet Assoc. 2010;110:848–856.291. Lallukka T, Laaksonen M, Rahkonen O, Roos E, Lahelma

E. Multiple socio-economic circumstances and healthy

food habits. Eur J Clin Nutr. 2007;61:701–710.292. Mullie P, Clarys P, Hulens M, Vansant G. Dietary pat-

terns and socioeconomic position. Eur J Clin Nutr. 2010;

64:231–238.293. Harrington J, Fitzgerald AP, Layte R, Lutomski J, Molcho

M, Perry IJ. Sociodemographic, health and lifestyle

predictors of poor diets. Public Health Nutr. 2011;14:2166–2175.

294. McNaughton SA, Ball K, Crawford D, Mishra GD. An

index of diet and eating patterns is a valid measure of

diet quality in an Australian population. J Nutr. 2008;

138:86–93.295. Darmon N, Ferguson EL, Briend A. A cost constraint

alone has adverse effects on food selection and nutrient

density: an analysis of human diets by linear pro-

gramming. J Nutr. 2002;132:3764–3771.

296. Darmon N, Ferguson E, Briend A. Do economic con-straints encourage the selection of energy dense diets?Appetite. 2003;41:315–322.

297. Butler P. Britain in nutrition recession as food prices riseand incomes shrink. Available at: http://www.the-guardian.com/society/2012/nov/18/breadline-britain-nutritional-recession-austerity. Accessed 15 March2017.

298. Diabetes Prevention Program Research Group, KnowlerWC, Fowler SE, Hamman RF, Christophi CA, HoffmanHJ, Brenneman AT, Brown-Friday JO, Goldberg R,Venditti E, Nathan DM. 10-year follow-up of diabetesincidence and weight loss in the Diabetes PreventionProgram Outcomes Study. Lancet. 2009;374:1677–1686.

299. Flegal KM, Kruszon-Moran D, Carroll MD, Fryar CD,Ogden CL. Trends in obesity among adults in theUnited States, 2005 to 2014. JAMA. 2016;315:2284–2291.

300. Casazza K, Fontaine KR, Astrup A, Birch LL, Brown AW,Bohan Brown MM, Durant N, Dutton G, Foster EM,Heymsfield SB, McIver K, Mehta T, Menachemi N,Newby PK, Pate R, Rolls BJ, Sen B, Smith DL, Jr, ThomasDM, Allison DB. Myths, presumptions, and facts aboutobesity. N Engl J Med. 2013;368:446–454.

301. Swinburn B, Sacks G, Ravussin E. Increased food energysupply is more than sufficient to explain the US epi-demic of obesity. Am J Clin Nutr. 2009;90:1453–1456.

302. Swinburn BA, Sacks G, Lo SK, Westerterp KR, Rush EC,Rosenbaum M, Luke A, Schoeller DA, DeLany JP, ButteNF, Ravussin E. Estimating the changes in energy fluxthat characterize the rise in obesity prevalence. Am JClin Nutr. 2009;89:1723–1728.

303. Swinburn BA, Jolley D, Kremer PJ, Salbe AD, Ravussin E.Estimating the effects of energy imbalance on changesin body weight in children. Am J Clin Nutr. 2006;83:859–863.

304. Church TS, Thomas DM, Tudor-Locke C, KatzmarzykPT, Earnest CP, Rodarte RQ, Martin CK, Blair SN,Bouchard C. Trends over 5 decades in U.S. occupation-related physical activity and their associations withobesity. PLoS One. 2011;6:e19657.

305. Simpson SJ, Raubenheimer D. Obesity: the protein le-verage hypothesis. Obes Rev. 2005;6:133–142.

306. Brooks RC, Simpson SJ, Raubenheimer D. The price ofprotein: combining evolutionary and economic analysisto understand excessive energy consumption. Obes Rev.2010;11:887–894.

307. Cotton JR, Burley VJ, Weststrate JA, Blundell JE. Dietaryfat and appetite: similarities and differences in the sa-tiating effect of meals supplemented with either fat orcarbohydrate. J Hum Nutr Diet. 2007;20:186–199.

308. Williams RA, Roe LS, Rolls BJ. Assessment of satietydepends on the energy density and portion size of thetest meal. Obesity (Silver Spring). 2014;22:318–324.

309. Lustig RH. Childhood obesity: behavioral aberration orbiochemical drive? Reinterpreting the first law ofthermodynamics. Nat Clin Pract Endocrinol Metab. 2006;2:447–458.

310. Taubes G. The science of obesity: what do we reallyknow about what makes us fat? An essay by GaryTaubes. BMJ. 2013;346:f1050.

311. Astwood EB. The heritage of corpulence. Endocrinology.1962;71:337–341.

312. Pennington AW. A reorientation on obesity. N Engl JMed. 1953;248:959–964.

313. Pennington AW. Obesity. Med Times. 1952;80:389–398.314. Rouhani MH, Haghighatdoost F, Surkan PJ, Azadbakht L.

Associations between dietary energy density andobesity: a systematic review and meta-analysis of ob-servational studies. Nutrition. 2016;32(10):1037–1047.

315. Ebbeling CB, Swain JF, Feldman HA, Wong WW, HacheyDL, Garcia-Lago E, Ludwig DS. Effects of dietary com-position on energy expenditure during weight-lossmaintenance. JAMA. 2012;307:2627–2634.

316. Veldhorst MA, Westerterp-Plantenga MS, WesterterpKR. Gluconeogenesis and energy expenditure aftera high-protein, carbohydrate-free diet. Am J Clin Nutr.2009;90:519–526.

317. Hall KD, Bemis T, Brychta R, Chen KY, Courville A,Crayner EJ, Goodwin S, Guo J, Howard L, Knuth ND,Miller BV III, Prado CM, Siervo M, Skarulis MC,Walter M,Walter PJ, Yannai L. Calorie for calorie, dietary fat re-striction results in more body fat loss than carbohydraterestriction in people with obesity. Cell Metab. 2015;22:427–436.

318. Tobias DK, Chen M, Manson JE, Ludwig DS, Willett W,Hu FB. Effect of low-fat diet interventions versus otherdiet interventions on long-term weight change inadults: a systematic review and meta-analysis. LancetDiabetes Endocrinol. 2015;3:968–979.

319. Ross R, Janssen I. Physical activity, total and regionalobesity: dose-response considerations. Med Sci SportsExerc. 2001;33(6 Suppl):S521–S527.

320. Klem ML, Wing RR, McGuire MT, Seagle HM, Hill JO. Adescriptive study of individuals successful at long-termmaintenance of substantial weight loss. Am J Clin Nutr.1997;66:239–246.

321. RosenbaumM, Goldsmith R, Bloomfield D, Magnano A,Weimer L, Heymsfield S, Gallagher D, Mayer L, MurphyE, Leibel RL. Low-dose leptin reverses skeletal muscle,autonomic, and neuroendocrine adaptations tomaintenance of reduced weight. J Clin Invest. 2005;115:3579–3586.

322. Levine JA, Lanningham-Foster LM, McCrady SK, KrizanAC, Olson LR, Kane PH, Jensen MD, Clark MM. In-terindividual variation in posture allocation: possiblerole in human obesity. Science. 2005;307:584–586.

323. Ravussin E, Lillioja S, Anderson TE, Christin L, BogardusC. Determinants of 24-hour energy expenditure in man.Methods and results using a respiratory chamber. J ClinInvest. 1986;78:1568–1578.

324. Bush T, Lovejoy JC, Deprey M, Carpenter KM. The effectof tobacco cessation on weight gain, obesity, and di-abetes risk. Obesity (Silver Spring). 2016;24:1834–1841.

325. Mineur YS, Abizaid A, Rao Y, Salas R, DiLeone RJ,Gundisch D, Diano S, De Biasi M, Horvath TL, Gao XB,Picciotto MR. Nicotine decreases food intake throughactivation of POMC neurons. Science. 2011;332:1330–1332.

326. Dhurandhar NV, Bailey D, Thomas D. Interaction ofobesity and infections. Obes Rev. 2015;16:1017–1029.

327. Schenk S, Saberi M, Olefsky JM. Insulin sensitivity:modulation by nutrients and inflammation. J Clin Invest.2008;118:2992–3002.

328. Thaler JP, Yi CX, Schur EA, Guyenet SJ, Hwang BH,Dietrich MO, Zhao X, Sarruf DA, Izgur V, Maravilla KR,Nguyen HT, Fischer JD, Matsen ME, Wisse BE, MortonGJ, Horvath TL, Baskin DG, Tschop MH, Schwartz MW.Obesity is associated with hypothalamic injury in ro-dents and humans. J Clin Invest. 2012;122:153–162.

329. Valdearcos M, Robblee MM, Benjamin DI, Nomura DK,Xu AW, Koliwad SK. Microglia dictate the impact ofsaturated fat consumption on hypothalamic in-flammation and neuronal function. Cell Reports. 2014;9:2124–2138.

330. Schur EA, Melhorn SJ, Oh SK, Lacy JM, Berkseth KE,Guyenet SJ, Sonnen JA, Tyagi V, Rosalynn M, De Leon B,Webb MF, Gonsalves ZT, Fligner CL, Schwartz MW,Maravilla KR. Radiologic evidence that hypothalamicgliosis is associated with obesity and insulin resistance inhumans. Obesity (Silver Spring). 2015;23:2142–2148.

331. DeBenedictis B, Guan H, Yang K. Prenatal exposure tobisphenol A disrupts mouse fetal liver maturation ina sex-specific manner. J Cell Biochem. 2016;117:344–350.

332. Faulk C, Kim JH, Anderson OS, Nahar MS, Jones TR,Sartor MA, Dolinoy DC. Detection of differential DNAmethylation in repetitive DNA of mice and humans



http://dx.doi.org/10.1210/er.2017-00111


perinatally exposed to bisphenol A. Epigenetics. 2016;11:489–500.

333. Rosenbaum M, Murphy EM, Heymsfield SB, MatthewsDE, Leibel RL. Low dose leptin administration reverseseffects of sustained weight-reduction on energy ex-penditure and circulating concentrations of thyroidhormones. J Clin Endocrinol Metab. 2002;87:2391–2394.

334. Rosenbaum M, Sy M, Pavlovich K, Leibel RL, Hirsch J.Leptin reverses weight loss-induced changes in regionalneural activity responses to visual food stimuli. J ClinInvest. 2008;118:2583–2591.

335. Kissileff HR, Thornton JC, Torres MI, Pavlovich K, MayerLS, Kalari V, Leibel RL, Rosenbaum M. Leptin reversesdeclines in satiation in weight-reduced obese humans.Am J Clin Nutr. 2012;95:309–317.

336. Fothergill E, Guo J, Howard L, Kerns JC, Knuth ND,Brychta R, Chen KY, Skarulis MC, Walter M, Walter PJ,Hall KD. Persistent metabolic adaptation 6 years after“The Biggest Loser” competition. Obesity (Silver Spring).2016;24:1612–1619.

337. Sumithran P, Prendergast LA, Delbridge E, Purcell K,Shulkes A, Kriketos A, Proietto J. Long-term persistenceof hormonal adaptations to weight loss. N Engl J Med.2011;365:1597–1604.

338. RosenbaumM, Leibel RL. Models of energy homeostasisin response to maintenance of reduced body weight.Obesity (Silver Spring). 2016;24:1620–1629.

339. Subramanian SV, Perkins JM,Ozaltin E, Davey SmithG.Weightof nations: a socioeconomic analysis of women in low- tomiddle-income countries. Am J Clin Nutr. 2011;93:413–421.

340. Bellissimo N, Akhavan T. Effect of macronutrientcomposition on short-term food intake and weight loss.Adv Nutr. 2015;6:302S–308S.

AcknowledgmentsWe acknowledge science writer Eric Vohr for his contributionto this scientific statement, and Drs. William Heisel and

Ashkan Afshin of the Institute for Health Metrics andEvaluation at the University of Washington for their helpfulcomments.Address all correspondence and requests for reprints to:

Michael W. Schwartz, MD, Department of Medicine, Uni-versity of Washington, UW Medicine at South Lake Union,850 Republican Street, Box 358055, Seattle, Washington98109. E-mail: [email protected] work was supported by National Institutes of Health

Grants DK083042, DK090320, and DK101997 and by theNational Institute of Diabetic and Digestive and KidneyDisease–funded Nutrition Obesity Research Center at theUniversity of Washington (to M.W.S.), R01 DK089038 (toL.M.Z.), DK52431 (to R.L.L.), DK076608 (to A.D.), DK072476(to L.M.R. and E.R.), and the Russell Berrie Foundation (toR.L.L.).Disclosure Summary: M.W.S. receives research funding

from and serves as a consultant to Novo Nordisk. R.J.S. re-ceives research support from Ethicon/Johnson & Johnson,Novo Nordisk, Janssen/Johnson & Johnson, MedImmune,Sanofi, and Boehringer-Ingelheim, and serves as a consultant/advisor to Ethicon/Johnson & Johnson, Novo Nordisk,Orexigen, Daiichi Sankyo, Janssen/Johnson & Johnson,Novartis, Paul Hastings Law Firm, Zafgen, Takeda, andBoehringer-Ingelheim. A.D. has received grants, contracts,honoraria, and consulting fees from numerous food andbeverage companies and other commercial and nonprofitentities with interests in diet quality and health. E.R. serves onthe Scientific Advisory Board to the Nutrilite Health Institutewith Amway, has a consultant contract with Janssen, serveson the Scientific Advisory Board for the Institute of Car-diometabolism and Nutrition in Paris, France, gives lectures atthe Open Academy in Venice, and is an advisor for theCenter for Medical Weight Loss. He has received researchgrants or unrestricted gifts from Amway, Nestle, the Nu-trition Science Initiative, and Ethicon Surgery. L.M.R. serves onthe Scientific and Medical Board Advisor to AdvoCare In-ternational, LP, and receives royalties for SmartLoss® and

SmartMoms®, which are intensive weight management in-

terventions delivered via mHealth technology. The remaining

authors have nothing to disclose.Disclaimer Statement: The Endocrine Society develops

Scientific Statements to assist clinicians by providing guid-

ance and recommendations for particular areas of practice.

Clinicians and other users should not consider this Scientific

Statement inclusive of all proper approaches or methods or

exclusive of others. It cannot guarantee any specific outcome,

nor does it establish a standard of care. It is not intended to

dictate the treatment of a particular patient. The in-

dependent judgment of health care providers and each

patient’s individual circumstances must determine treat-

ment decisions. The Endocrine Society makes no warranty,

express or implied, regarding this Scientific Statement and

specifically excludes any warranties of merchantability and

fitness for a particular use or purpose. The Endocrine Society

shall not be liable for direct, indirect, special, incidental, or

consequential damages related to the use of the information

contained herein.

AbbreviationsAgRP, agouti-related protein; ARC, arcuate nucleus; BAT,brown adipose tissue; BMI, body mass index; BPA, bisphenol

A; CGRPPBN, calcium gene-related peptide expressing neu-

rons in the parabrachial nucleus; EDC, endocrine-disrupting

chemical; GABA, g-aminobutyric acid; GI, gastrointestinal;

GWAS, genome-wide association study; HFD, high-fat diet;

iBAT, interscapular brown adipose tissue; LPL, lipoprotein

lipase; NEAT, nonexercise activity thermogenesis; NPY,

neuropeptide Y; PFC, perfluorinated chemical; POMC, pro-

opiomelanocortin; SES, socioeconomic status.



mailto:[email protected]

Documents

SCIENTIFIC STATEMENT€¦ · underlying this “upward setting” or “resetting” of the defended level of body-fat mass, whether inherited or acquired. The ongoing study of how