Metabolic factors determining the susceptibility to weight gain: current evidence

Abstract

Purpose of Review

There is substantial inter-individual variability in body weight change, which is not fully accounted by differences in daily energy intake and physical activity levels. The metabolic responses to short-term perturbations in energy intake can explain part of this variability by quantifying the degree of metabolic “thriftiness” that confers more susceptibility to weight gain and more resistance to weight loss. It is unclear which metabolic factors and pathways determine this human “thrifty” phenotype. This review will investigate and summarize emerging research in the field of energy metabolism and highlight important metabolic mechanisms implicated in body weight regulation in humans.

Recent findings

Dysfunctional adipose tissue lipolysis, reduced brown adipose tissue activity, blunted fibroblast growth factor 21 secretion in response to low-protein hypercaloric diets, and impaired sympathetic nervous system activity might constitute important metabolic factors characterizing “thriftiness” and favoring weight gain in humans.

Summary

The individual propensity to weight gain in the current obesogenic environment could be ascertained by measuring specific metabolic factors which might open up new pathways to prevent and treat human obesity.

Introduction

The widespread obesity epidemic is growing constantly: among US adults aged 20 years and older, the prevalence of obesity (body mass index [BMI] > 30 kg/m²) increased from 33.7% in 2008 to 39.6% in 2016, while the prevalence of severe or morbid obesity (BMI > 40 kg/m²) increased from 5.7% to 7.7% at the same time(1). In the United States, obesity is associated with nearly one out of five deaths(2), therefore it is important to identify individuals at greater risk for excess weight gain, so interventions can start before the onset of obesity. Carefully conducted overfeeding studies controlling for energy intake and physical activity levels demonstrated that there are considerable differences in weight gain among individuals – with some gaining up to 2x more weight than others despite similar conditions(3–5). These findings indicate that additional metabolic factors might determine the extent of weight gain besides the usual suspects: unbalanced (e.g., high-fat or low-protein) hypercaloric diet and sedentary lifestyle. As it is beyond dispute that weight gain is a result of sustained positive energy balance, i.e. where caloric intake chronically exceeds energy expenditure, metabolic factors must influence either one of these components (intake and/or expenditure) and be responsible for the degree of energy imbalance. Recent studies found that inter-individual differences in energy expenditure during acute dietary interventions such as 24 hours of fasting or 200% overfeeding can identify an energy-saving or “thrifty” phenotype that is more susceptible to weight gain and more resistant to weight loss compared to a “spendthrift” counterpart(6). As the thrifty phenotype is defined by whole-body energy expenditure, it must encompass underlying physiological and biological mechanisms, which in concert determine the individual susceptibility to weight gain/loss. The following systems and organs regulating energy utilization and dissipation may therefore constitute important physiological traits of the thrifty phenotype: I) white adipose tissue, as the main organ involved in lipid storage and adipokine production; II) brown adipose tissue, as calorie-burning, “heating-organ”, which also secretes thermogenic hormones; III) the liver, as an organ with high metabolic activity which also secretes hepatokines involved in the regulation of energy homeostasis, IV) the activity of the sympathetic nervous system, and V) the stomach-derived hormones including ghrelin and its opposing effects on energy intake and expenditure. This narrative review focuses on the significance of these parameters with regard to their role in defining the thrifty phenotype and susceptibility to future weight gain. It will extend the concepts previously discussed in other reviews focusing on brown adipose tissue activity(7, 8), adaptive thermogenesis(9, 10), and macronutrient balance(11) with regard to the characterization of weight gain susceptibility.

The human thrifty phenotype concept

The human thrifty phenotype hypothesis presupposes that lower 24-hour energy expenditure (24hEE) during famine preserves body mass and promotes survival. The hypothesis has evolved from an original theory ascribing the high prevalence of type 2 diabetes in today’s modern societies to thrifty genes promoting hyperinsulinemia(12, 13) and to acquired thriftiness through fetal malnutrition leading to impaired pancreatic function(13) to a concept of energy conservation during famine and overnutrition(4, 14–16). While it was advantageous for body weight preservation to be thriftier in historical hunter-gatherer cultures, it may be disadvantageous in today’s food rich environment as thriftier individuals presumably have a smaller increase in energy expenditure in response to increased energy intake. Therefore, during sustained periods of food abundance, thrifty subjects may experience a higher degree of positive energy balance, leading to increased body fat accumulation and ultimately resulting in greater adiposity and metabolic dysregulation.

Previous studies found evidence supporting the thrifty phenotype hypothesis by showing that the degrees of metabolic adaptation to 24h fasting and to 24h overfeeding (200% of eucaloric needs) are highly variable among individuals and that those subjects who decrease their 24hEE more during 24h fasting (thriftier individuals) are those who increase 24hEE less during 24h overfeeding(14, 16). The identification and clinical characterization of human thrifty phenotypes identified by acute dietary interventions is discussed in detail elsewhere(6, 17).

The role of the human thrifty phenotype in future weight change was recently demonstrated in different clinical trials including both inpatient and outpatient studies, where it was shown that individuals with a thriftier metabolism (defined by a greater decrease in 24hEE during 24h fasting) lose less weight during 6 weeks of highly controlled, 50% caloric restriction(15), spontaneously gain more weight over time in free-living conditions(16), and gain more weight during 6 weeks of daily low-protein overfeeding(4), the latter being a dietary tool that best characterizes the individual susceptibility to weight gain(16, 18, 19) (Table 1). In line with these findings for energy expenditure phenotypes defined by short-term changes in 24hEE during acute fasting, another study found that thriftiness defined by a relatively lower sleeping metabolic rate was associated with greater fat mass retained after 6 months in subjects who underwent 8 weeks of sustained overfeeding(20). Further, in the same study, relatively higher 24hEE following 8-week overfeeding (indicative of a more spendthrift phenotype) was associated with more body fat loss in the subsequent 6 months(20).

Table 1 –

Recently discovered metabolic factors associated with weight gain in humans.

Metabolic factor	Effect on future weight gain	Clinical studies
Metabolic adaptation to short-term fasting	A greater decrease in 24hEE during 24h fasting (i.e., thriftier phenotype) predicted greater weight gain after 6 months in free-living conditions and after 6 weeks of highly controlled, daily low-protein overfeeding, as well as less weight loss during 6 weeks of highly-controlled 50% caloric restriction.	Schlögl et al. Energy Expenditure Responses to Fasting and Overfeeding Identify Phenotypes Associated with Weight Change. Diabetes, 2015 (16). Hollstein et al. Metabolic response to fasting predicts weight gain during low-protein overfeeding in lean men: further evidence for spendthrift and thrifty metabolic phenotypes. AJCN, 2019 (4). Reinhardt et al. A Human Thrifty Phenotype Associated With Less Weight Loss During Caloric Restriction. Diabetes, 2015 (15).
Impaired metabolic flexibility to overfeeding	A smaller decrease in RQ after 24h high-fat overfeeding (200% eucaloric requirements), reflecting a smaller increase in lipid oxidation rate, predicted greater weight gain after 6 and 12 months	Begaye et al. Impaired Metabolic Flexibility to High-fat Overfeeding Predicts Future Weight Gain. Diabetes, 2019 (24).
	Impaired fuel selection and lower nocturnal nonesterified fatty acid concentrations during overfeeding predicted greater weight gain in obesity-prone individuals after 4 years of follow-up.	Rynders et al. Associations among dietary fat oxidation responses to overfeeding and weight gain in obesity-prone and resistant adults. Obesity, 2018 (25).
Impaired adipocyte lipolysis	A greater in vitro basal lipolysis was associated with less weight gain after 8 years of follow-up	Frankl. et al. In Vitro Lipolysis is Associated with Whole-Body Lipid Oxidation and Weight Gain in Humans (35).
	Women with high basal lipolysis but low isoprenaline-stimulated adipocyte lipolysis in vitro, reflective of impaired fat cell lipolysis, had a 7.4-fold higher risk to gain weight during 13 years of follow-up.	Arner et al. Weight Gain and Impaired Glucose Metabolism in Women Are Predicted by Inefficient Subcutaneous Fat Cell Lipolysis. Cell Metabolism, 2018 (34).
Impaired brown adipose tissue activity	Low brown adipose tissue activity after 2 hours of mild cold exposure (16°C) was associated with greater fat mass gain after 6 months	Schlögl et al. Overfeeding over 24 hours does not activate brown adipose tissue in humans. JCEM, 2013 (84). Begaye et al. Norepinephrine and T4 Are Predictors of Fat Mass Gain in Humans With Cold-Induced Brown Adipose Tissue Activation. JCEM, 2018 (106).
FGF21 response to 24h low-protein overfeeding	A blunted increase in plasma FGF21 following 24h low-protein overfeeding (200% eucaloric requirements) predicted greater free-living weight gain after 6 months.	Vinales et al. FGF21 is a Hormonal Mediator of the Human “Thrifty” Metabolic Phenotype. Diabetes, 2018 (128).
Central sympathetic	A greater plasma norepinephrine concentration at	Masuo et al. Serum uric acid and plasma
overactivity (≙ high norepinephrine secretion)	baseline predicted future weight gain after 5 and 20 years of follow-up	norepinephrine concentrations predict subsequent weight gain and blood pressure elevation. Hypertension, 2003 (182). Gudmundsdottir et al. Arterial noradrenaline predicts rise in body mass index in a 20-year follow-up of lean normotensive and hypertensive men. Journal of Hypertension, 2008 (181).
	Greater urinary norepinephrine excretion over 24 hours during energy balance predicted greater fat mass gain after 6 months.	Begaye et al. Norepinephrine and T4 Are Predictors of Fat Mass Gain in Humans With Cold-Induced Brown Adipose Tissue Activation. JCEM, 2018 (106).
	A greater increase in plasma norepinephrine concentration in response to mental stress predicted greater central adiposity after 18 years of follow-up	Flaa et al. Does sympathoadrenal activity predict changes in body fat? An 18-y follow-up study. AJCN, 2008 (183).
Reduced central sympathetic activity (≙ low norepinephrine secretion)	Lower urinary norepinephrine excretion over 24 hours predicted greater weight gain after 3 years of follow-up.	Tataranni et al. A low sympathoadrenal activity is associated with body weight gain and development of central adiposity in Pima Indian men. Obesity Research, 1997 (162).
Reduced sympathoadrenal activity (≙ high epinephrine secretion)	Lower urinary epinephrine excretion over 24 hours was predictive for a greater waist-thighratio after 3 years of follow-up.	Tataranni et al. A low sympathoadrenal activity is associated with body weight gain and development of central adiposity in Pima Indian men. Obesity Research, 1997 (162).
	A reduced plasma epinephrine response to mental stress predicted greater BMI and central adiposity after 18 year of follow-up	Flaa et al. Does sympathoadrenal activity predict changes in body fat? An 18-y follow-up study. AJCN, 2008 (183).

24hEE, 24-hour energy expenditure; RQ, respiratory quotient

In addition to the extent of metabolic adaptation to 24h fasting and overfeeding, the “shift” in fuel preference in response to these acute dietary challenges with different macronutrient content (i.e., metabolic flexibility(11, 21–23)) may constitute a further feature characterizing the human thrifty phenotype. For instance, the degree of metabolic inflexibility during acute (24h) high-fat (60%) overfeeding diet – that is, the inability to switch to fat as source for oxidation in a context of dietary fat abundance – was associated with greater free-living weight gain after 1 year of follow-up(24). This finding for metabolic flexibility to acute overfeeding is in line with another independent study which reported that a lower rate of nocturnal fat oxidation during 3-day overfeeding is associated with greater long-term weight gain(25, 26).

In summary, thriftier and metabolically inflexible individuals might be prone to gain weight over time as they are unable to increase energy expenditure and/or fat oxidation during conditions of energy surplus and, thus, have less potential to dissipate excess energy as heat when overeating, ultimately leading to greater body fat accumulation during chronic states of positive energy balance. The underlying biological mechanisms mediating the metabolic differences between individuals and characterizing the susceptibility to weight gain are not fully elucidated but may encompass: I) white adipose tissue function, II) brown adipose tissue activity, III) the secretion of liver-derived hepatokines involved in energy homeostasis, IV) sympathetic nervous system activity, and V) the stomach-derived hormone ghrelin, all of which will be separately discussed in the following sections.

I). White adipose tissue

White adipose tissue (WAT) is the most abundant adipose tissue in the human body with primary locations in visceral and subcutaneous depots and ranging from 10% of total body mass in lean subjects to ~50% in individuals with severe obesity(27). Importantly, for the same body mass index, females generally possess around 10% higher body fat than males which, in women, is more abundant around the hips (“pear shape”) compared to males who have relatively more abdominal fat (“apple shape”)(28). During weight gain, WAT adipocytes can grow either in size or in number to increase fat storage capacity(29, 30). Adipocyte hypertrophy (i.e. the increase in adipocyte size) during caloric excess has been associated with metabolically “unhealthy” obesity, i.e. insulin resistance, non-alcoholic fatty liver disease, inflammation, and altered blood lipid profiles(29, 31, 32), whereas adipogeneses (i.e. the increase in adipocyte number) seems to favor “healthy obesity”, i.e. high insulin sensitivity and immune tolerance(32).

It has been hypothesized that these structural differences in WAT (i.e., smaller vs bigger cells) might be a determinant of future weight gain; however, no associations have been found(29, 33, 34), suggesting that WAT morphology is likely to be a consequence rather than a cause of future weight gain. Conversely, recent findings suggest that adipocyte function might be more relevant than adipocyte morphology in characterizing the susceptibility to weight gain. For instance, impaired lipolysis in abdominal subcutaneous WAT predicted greater long-term weight gain in an ethnically diverse cohort(35) and in a cohort of premenopausal women of Caucasian descent(34) (Table 1), presumably due to the effect of lipolysis on whole-body lipid oxidation rate(35, 36). Accordingly, in other separate studies, reduced 24h lipid oxidation rate during short-term high-fat overfeeding (24h with 200% of eucaloric requirements)(37) as well as reduced nocturnal lipid oxidation rate during short-term balanced overfeeding (3d with 140% of energy requirements)(25) have also been associated with greater free-living weight gain. In line with these observations, the assessment of carbon isotope ratios of stored triglycerides in abdominal subcutaneous adipocytes revealed that overweight subjects have an increased triglyceride age in their adipocytes compared to lean subjects which is indicative for a decreased lipid turnover(38). Interestingly, lipid turnover rate also decreases with age(39) which may favor age-related adiposity(39). Conversely, in patients undergoing bariatric surgery, a reduced lipid turnover rate assessed before surgery predicted better post-surgical weight loss and better weight loss maintenance(39). According to the authors, subjects with obesity and low lipid turnover might have more capacity to improve their lipid turnover than those with a higher initial lipid turnover(39). Hence, in that study, subjects with better weight loss outcomes after bariatric surgery also increased their lipid turnover(39).

In summary, the defects in adipocyte lipolysis and lipid turnover may constitute relevant factors underlying the changes in energy expenditure and substrate oxidation under dynamic feeding, and, ultimately, weight gain susceptibility. As catecholamines are the major stimuli for lipolysis(40, 41), impaired adipocyte function might also be considered as a downstream effect of impaired catecholamine signaling. This will be discussed in detail in section IV (The sympathetic nervous system).

Besides its fat storage (i.e. lipogenesis) and utilization (lipolysis) capabilities, WAT it is also an active endocrine organ secreting a variety of hormones termed adipokines(27). The adipokines released by WAT constitute important factors in the pathophysiology of human obesity as they play a major role on energy homeostasis(42–44). The most relevant adipokines with regard to energy metabolism and weight gain are leptin and adiponectin, although the number of adipocyte-derived hormones is in the hundreds(45).

Leptin circulates in proportion to body fat mass and informs the brain about the size of energy stores, as several studies in rodents have indicated(46–48). In mice, leptin acts on the hypothalamus to reduce food intake and to increase energy expenditure(49, 50). In humans, the large majority of studies indicate that leptin has little or no effect on energy expenditure(46, 51). However, one landmark study showed that low-dose leptin treatment restored the decline in energy expenditure after caloric restriction and weight loss(52) and, recently, another study demonstrated that leptin is a determinant of resting metabolic rate in lean subjects independent of fat mass(53). Although the role of leptin on human energy expenditure is not fully elucidated, other studies indicate that leptin influences food intake as leptin deficiency causes severe hyperphagia and excessive weight gain(54) and leptin treatment can diminish perception of food reward in humans, leading to greater satiety(55). As early as leptin was discovered in humans, it was hypothesized that lower leptin concentrations relative to a given body fat mass may predispose individuals to future weight gain. However, apart from one single study in overweight Native Americans(56), this association could not be confirmed in other ethnicities(57, 58). In fact, an inverse association was reported, that is, a higher leptin concentration relative to body fat mass was associated with greater weight gain(59, 60). Additionally, obesity, as a result of increased body fat mass, is associated with hyperleptinemia and, presumably, concomitant leptin resistance(61), possibly due to a defect in intracellular signaling associated with the leptin receptor or decreases in leptin transport across the blood–brain barrier as rodent studies have indicated(62). Thus, treating obesity by administering leptin might fail due to increased leptin resistance. This hypothesis is supported by the results of a large randomized clinical trial showing that leptin administration in subjects with obesity did not contribute to further weight loss during sustained caloric restriction for 24 weeks(63). In this same study, clinically significant losses in body weight and fat mass were only achieved at high-dose leptin (0.30 mg/kg per day) but not at lower doses (0.01–0.10 mg/kg per day) and only in a subset of participants, possibly by overcoming the putative greater degree of leptin resistance(63).

Adiponectin was discovered one year later than leptin and is – in contrast to leptin – exclusively produced by adipose tissue(64). Adiponectin is considered a “healthy” adipokine as it improves insulin sensitivity(65), reduces inflammation(66), and has vascular-protective effects(67). Accordingly, the concentration of adiponectin is increased in lean subjects while it is decreased in subjects with obesity(68, 69), and loss-of-function polymorphisms in the adiponectin gene or its receptor genes are associated with impaired lipid oxidation and central adiposity suggesting a role of adiponectin in obesity and metabolic flexibility(70). Although these metabolic characteristics indicate that lower adiponectin concentrations might be a causal factor for weight gain, this could not be confirmed in longitudinal studies(58, 71) except for one conducted in women(72). Also, higher adiponectin concentration was not associated with greater lipid oxidation in another human study(73).

In summary, despite their abundant secretion by adipose tissue, both leptin and adiponectin do not appear to play a major role in characterizing inter-individual differences in energy metabolism and future weight gain in humans. However, as many other adipokines are still unknown or less investigated (i.e. angiopoietins, bone morphogenic proteins, chemerin, endotrophin, lipocalin 2, and neuregulin 4(43)), future research in this area is warranted to identify and metabolically characterize adipokines that might be implicated in the regulation of body weight.

II). Brown adipose tissue

In contrast to WAT, brown adipose tissue (BAT) volume is much lower in the human body, ranging from 0 to 600 g and mainly depending on sex, ethnicity, and body size(74–76) and being predominantly localized in the cervical-supraclavicular region(76–78). During thermoneutral conditions, BAT is only detected in 7.5% of females and 3.1% of males(76) and its major metabolic differences compared to WAT are the higher numbers of mitochondria (which gives the “brownish color”) and the elevated expression of uncoupling protein 1 (UCP1), allowing BAT to be the active site for thermogenesis and heat production(76). BAT activity is mainly stimulated via cold exposure(79–81). Additionally, BAT appears to be activated in some individuals by high-carbohydrate diets(82, 83) but not after high-fat diet overfeeding.(84, 85). As BAT adipocytes express β3- adrenoceptors on their cell surface, they can also be activated via β3- adrenoceptor agonists like mirabegron(86–88) leading to greater BAT activity in humans(89).

Rodent studies show that increased BAT activity is protective against diet-induced obesity and diabetes(90) while UCP1-ablation/Ucp1-deficieny increases weight gain in some(91, 92) but not all studies(93), likely depending on ambient temperature and dietary composition(94). In humans, BAT activity is 4-fold higher in lean compared to overweight individuals(79). Likewise, greater BAT volume is associated with higher resting energy expenditure(75) and selective BAT activation via β3- adrenoceptor agonists like mirabegron increases resting energy expenditure(86, 87). These findings suggest that BAT might play a role in weight gain via its effects on metabolic rate. However, it is unclear whether increased energy expenditure itself attenuates weight gain as human studies investigating the relationship between energy expenditure and future weight change have reported mixed results. Some studies showed that a relatively higher metabolic rate is associated with weight loss(95–98) – most of them done in overweight Native Americans – while others could not replicate these findings(99–101) or even reported opposite results, i.e. a higher metabolic rate was related to greater weight gain in a Nigerian population(102). To explain these divergent findings, it has been hypothesized that increased energy expenditure may not lead to negative energy balance as it could trigger counterregulatory effects on food intake to restore energy balance, which could subvert weight loss or even lead to weight gain if the counter effects of increased energy expenditure on food intake are overcompensatory(6, 17). These putative “energy sensing” mechanisms(103) might therefore render therapies focused on increasing metabolic rate less effective to induce weight loss. Focusing on increasing BAT volume (browning) and its activity might still be promising, as recent research in mice shows that greater secretin-mediated BAT activity also improves the other side of the energy balance equation by increasing satiety(104). However, a recent study reported that inter-individual variability both in BAT volume and activity is not associated with measures of ad libitum food intake and appetite-related sensations (105), indicating that BAT may not play a major role in the regulation of energy intake in humans.

To date, apart from one small study showing that reduced cold-induced BAT activity was associated with greater fat mass gain after 6 months in free-living conditions(84, 106) (Table 1), no other human study showed that either small or absent BAT volume or reduced BAT activity are indicative of individual propensity to gain weight. Larger prospective studies are warranted to replicate those findings and demonstrate that the morphologic and metabolic parameters related to BAT physiology are associated with changes in energy expenditure and substrate oxidation under dynamic feeding and can determine the individual susceptibility to weight gain.

Furthermore, it has to be clarified whether pharmacological BAT activation and/or expansion of BAT tissue (browning) can prevent weight gain or induce weight loss in subjects with obesity.

III). Liver and hepatokines

The liver is a major organ with high metabolic rate activity(107) as it has a ~15-fold higher basal energy expenditure than skeletal muscle (200 kcal/kg/day vs 13 kcal/kg/day), most likely due to energy-costly processes like gluconeogenesis and glycogenolysis. Additionally, the liver is an endocrine organ involved in the regulation of whole-body energy expenditure and energy homeostasis through the secretion of hepatokines(108).

Among the hepatokines secreted by the liver that have a systemic role in body metabolism(108, 109), one of the most important hepatokines with regard to energy metabolism is FGF21(110, 111). Besides being secreted from the liver, in mice FGF21 is also secreted by other organs including pancreas, thymus, WAT, and BAT(110). In humans, FGF21 is almost exclusively produced in the liver and, to lesser extent, in the brain and in skeletal muscle(110). In mice, FGF21 improves the metabolic profile by increasing insulin sensitivity(112) and energy expenditure(113), and by modulating appetite and food preference(114). Similarly, in humans, FGF21 was shown to increase energy expenditure(115) and modulate food preference(116). The effects of FGF21 on energy expenditure are mainly regulated via activation of UCP1 in BAT and WAT in both mice and humans(115, 117), although recent research in rodents suggests that some thermogenic effects of FGF21 may be independent of UCP1(118, 119) and may include the central nervous system(120, 121). Paradoxically, FGF21 concentrations are elevated in obesity(122) which led to the hypothesis that obesity may represent an FGF21-resistant state(123, 124). However, this concept remains controversial and requires more investigation(124).

Interestingly, in humans, dietary perturbations like short-term high-carbohydrate overfeeding(125), especially fructose-rich diets(126), and short- and long-term low-protein overfeeding(4, 127, 128) can increase FGF21 secretion which is likely mediated by the periprandial insulin-response(129). Due to its remarkable increase during low-protein diets (up to 44-fold(4)), FGF21 also represents an independent endocrine signal of protein restriction(127). However, in humans this only seems to be true during dietary protein restriction with positive energy balance as short-term fasting decreases FGF21(128). In contrast, famine rapidly induces FGF21 in mice(130).

The acute FGF21 response to 24h of low-protein overfeeding is highly variable among healthy individuals, with an average ~3-fold increase(128). Interestingly, a blunted FGF21 response during this hypercaloric diet was associated with smaller concomitant increase in 24hEE (indicative for a thriftier phenotype) and predicted greater spontaneous weight gain after 6 months(128). These results indicate that FGF21 is one of the major hormonal determinants of the thrifty phenotype and susceptibility to weight gain.

In this context, it is tempting to speculate that targeted exogenous FGF21 therapy might promote weight loss in some individuals by converting a thriftier into a more spendthrift metabolism. Indeed, previous studies reported that FGF21 treatment induces weight loss in diet-induced obese and ob/ob mice(113) as well as in humans with obesity(131, 132) – although in one human study, pegylated FGF21 treatment did not lead to reduction in body weight(133). While it was shown in animals that the FGF21-induced weight loss is achieved via increased energy expenditure and decreased food intake(113, 132), it is unknown how FGF21 treatment may lead to weight loss in humans. Hence, randomized clinical trials administering FGF21 analogs and including assessments both of energy expenditure and food intake are warranted to elucidate the mechanisms thereby FGF21 may lead to weight loss.

Fetuin-A, also known as α2-HS-glycoprotein, might constitute another important hepatokine involved in energy metabolism regulation and weight gain susceptibility(134). Like FGF21, Fetuin-A is not exclusively produced by the liver but also by adipose tissue(134). Apart from its protective effects against aortic and mitral valve calcification(135), fetuin-A has several adverse properties that increase the risk of developing a metabolic syndrome as it inhibits the insulin receptor in vitro, promotes systemic low-grade inflammation, and impairs adipocyte function – in detail summarized elsewhere(134). Accordingly, recent human studies found that higher fetuin-A concentrations are positively associated with insulin resistance(136, 137) and increased risk of developing type 2 diabetes(136, 138, 139), although predominantly in the elderly and in women(136, 138, 139). Further, they are also associated with higher BMI and visceral adipose tissue accumulation(136, 138, 140), suggesting a potential role for fetuin-A in weight gain. This is supported by rodent studies showing that fetuin-A deficient mice (through knockout of the Ahsg gene) are resistant to weight gain and body fat accumulation while being more insulin sensitive, even on a high-fat diet(141). In humans, bidirectional mendelian randomization studies found that fetuin-A is likely to be causal for higher BMI but not vice versa(142). In accordance with these results, one prospective study reported that higher serum fetuin-A concentrations at baseline predicted greater visceral adipose tissue accumulation in the elderly after 5 years of follow-up(143). Similarly, weight loss interventions consistently reduced fetuin-A concentrations (studies are summarized in detail elsewhere(134)).

In summary, fetuin-A likely is likely to have a causal role in the development of obesity and related metabolic disorders such as type 2 diabetes. However, it remains unclear how fetuin-A may characterize weight gain susceptibility. Therefore, future studies are warranted to clarify whether fetuin-A is associated with decreased energy expenditure or increased food intake. Furthermore, future studies should also investigate whether higher fetuin-A concentrations are a trait of the thrifty phenotype.

IV). The sympathetic nervous system

The sympathetic nervous system (SNS) is involved in the regulation of heart rate, blood pressure, lipolysis, brown adipose tissue activity, and digestion(144, 145). The SNS can be conceptually divided into two branches: the central sympathetic branch, which releases norepinephrine from postganglionic synaptic nerve endings directly into target tissues, and the sympathoadrenal branch represented by the adrenal medulla, which almost exclusively releases epinephrine into the blood stream(146). In the context of energy homeostasis and body weight regulation, both branches can be specifically (and somewhat independently) activated by different stressors, i.e. cold exposure or overfeeding lead to a greater central sympathetic activation (≙ increased norepinephrine secretion)(147, 148) whereas emotional stress and fasting-induced hypoglycemia mainly activate the sympathoadrenal branch (≙ increased epinephrine secretion)(146, 147, 149). Both branches will be discussed separately.

The central sympathetic branch (norepinephrine)

Studies in rodents and humans showed that norepinephrine modulates both thermogenesis(150–156) and food intake(157). Further, in vivo studies showed that norepinephrine has a pronounced effect on stimulating lipolysis(41, 158), thus indicating a role in fuel selection, macronutrient balance, and future body weight and fat mass gain. Despite this, research in humans has largely reported a positive association between higher norepinephrine secretion and adiposity(150, 152, 159–162), although some studies in lean adults and children with obesity also found an inverse association(151, 163). However, it is generally accepted that central sympathetic overactivity is a metabolic trait of obesity in most individuals(164, 165). The question remains why this is the case – given that increased sympathetic outflow should actually favor weight loss due to its thermogenic and appetite-suppressing effects. Two potential explanations will be briefly discussed:

1. Landsberg proposed a theory which links diet-induced hyperinsulinemia to sympathetic overactivation and obesity(166). According to his hypothesis, as individuals gain weight, they develop insulin resistance in peripheral adipose and muscular tissues. The central SNS is remains sensitive to insulin even in a such insulin-resistant physiologic environment(167). Thus, higher insulin concentrations increase sympathetic activity centrally, ultimately leading to increased metabolic rate that may attenuate the extent of diet-induced weight gain. Thus, under this framework, the observed sympathetic overactivity in overweight individuals can be considered a physiological mechanism to limit excessive weight gain(168), fitting the theory of “luxuskonsumption” by Rothwell and Stock(169). This hypothesis is supported by several human studies showing that chronic overfeeding induces central sympathetic overactivation(170–172). However, not all studies found this association(173) and, from an evolutionary perspective, it has been questioned whether the human body would have also developed defensive mechanisms to prevent excess weight gain, as our ancestors were likely more threatened by famine than by excess food intake. This compensatory mechanism also seems to have a physiologic limit(147). If excess dietary intake is substantial and persists over a long period, the hyperinsulinemia-induced central sympathetic overactivity may reach a “threshold” above which it cannot further increase metabolic rate(147). In this scenario, sustained overeating might lead to chronic central sympathetic overactivation and obesity at the same time, which may explain the positive cross-sectional association between norepinephrine (as a marker of sympathetic activity) and adiposity. Furthermore, the greater sympathetic tone would come along with detrimental health effects such as obesity-related hypertension(147, 174).

2. An alternative hypothesis proposes that chronic central sympathetic overactivity (i.e. as a result of overeating) may lead to insensitivity and/or downregulation of β-adrenergic receptors, a phenomenon which could represent a state of “catecholamine resistance”(40, 175) that may reduce lipolysis and blunt the downstream thermogenic effects of catecholamines(164). The “catecholamine resistance” hypothesis is supported both by animal and in vitro studies showing that β-adrenergic agonists(176) and chronic low-grade inflammation(177) – a consequence of increased adiposity and metabolic dysregulation – can induce desensitization of β-adrenergic receptors. In women, decreased β-adrenoceptor density also contributes to the development of abdominal obesity(178). Catecholamine resistance may be in part hereditary, as genetic polymorphisms in the β2-adrenoceptor gene can cause marked differences in the lipolytic sensitivity of this receptor in human adipocytes and are associated with increased adiposity(179). Blunted β-adrenoceptor activity due to genetic polymorphisms is also associated with higher plasma norepinephrine concentrations and greater rebound weight gain after successful weight loss(180), suggesting that catecholamine resistance may lead to an overproduction of catecholamines through a physiological feedback-loop.

Regardless of its underlying biology, central sympathetic overactivity – reflected by greater norepinephrine production – is not only associated with greater adiposity in cross-sectional analyses but has also been linked to future weight gain in two relatively large longitudinal studies with a follow-up of 5 and 20 years(181, 182) (Table 1). Furthermore, one recent study reported that greater urinary norepinephrine excretion rate over 24h during energy balance conditions predicted greater fat mass gain after 6 months in a small group of mixed ethnicities(106). Lastly, in another study, a greater increase in plasma norepinephrine concentration in response to a mental stress test (where subjects were asked to perform repetitive calculations in a noisy environment) predicted increased central adiposity and also tended to predict greater weight gain after 18 years of follow-up(183). All these studies corroborate the hypothesis that central sympathetic overactivity is a metabolic determinant of future weight gain. However, one single study done in Native Americans found an opposite association(162), suggesting that ethnic-specific differences may play an important role in the relationship between SNS activity and future weight gain.

The sympathoadrenal branch (epinephrine)

Similar to norepinephrine, epinephrine is implicated in the regulation of thermogenesis(155, 184) and lipolysis(155, 185, 186) in humans. However, in contrast to the positive cross-sectional association between adiposity and norepinephrine concentration, greater adiposity has been associated with lower epinephrine concentration in the majority(159–161, 187) but not in all cross-sectional studies(162). Importantly, a longitudinal study conducted in Native Americans reported that lower urinary epinephrine excretion over 24 hours of eucaloric conditions is predictive for increased central adiposity after 3 years of follow-up(162) (Table 1). These findings are supported by another study in Caucasians showing that a reduced plasma epinephrine response to mental stress predicted greater weight gain after 18 year of follow-up(183). Low sympathoadrenal activity may favor adiposity by decreased lipolysis(186) and/or decreased thermogenesis(155, 184).

To conclude, central sympathetic overactivity and less sympathoadrenal activity might predict future weight gain partly via their effects on energy metabolism. However, future studies are warranted to further investigate how norepinephrine and epinephrine interrelate with regard to the development of obesity and how their effects on lipolysis and thermogenesis can be augmented pharmacologically. Furthermore, the role of catecholamines as biological mediators of changes in energy expenditure under dynamic feeding should be further investigated as there is preliminary evidence for a physiological connection between epinephrine and metabolic thriftiness defined by a greater decrease in 24hEE during short-term fasting(188).

V). Ghrelin

Ghrelin is a mainly stomach-derived hormone which is also produced in small amounts by other organs like heart, lung, kidney, adrenal glands, testis, ovaries, thyroid gland, and pancreas(189). Ghrelin circulates in a non-acylated (mainly inactive) and in an acylated (mainly active) form(189). The ratio of inactive to active ghrelin in the blood is 9:1(189). Ghrelin concentration fluctuates during the day with an increase before meals and a decrease upon food consumption(190), qualifying ghrelin as a “hunger hormone”. The postprandial decrease in ghrelin appears to be macronutrient specific and is greatest following carbohydrates ingestion, followed by proteins and lipids(191).

Ghrelin is implicated in the regulation of weight gain as it modulates both sides of the energy balance equation in opposing directions. On one hand, ghrelin increases food intake, e.g., acute administration of ghrelin was shown to increase appetite and food intake in rodents(192, 193) and humans(194, 195), most likely via activation of homeostatic hypothalamic circuits(193), ultimately leading to increased adiposity in preclinical models(192). On the other hand, ghrelin conserves energy by decreasing metabolic rate as elegantly demonstrated in a rodent study where ghrelin-antibody treatment increased metabolic rate(196). Accordingly, in humans, higher ghrelin concentrations are associated with lower resting metabolic rate(197, 198) and lower postprandial thermogenesis(198). Ghrelin may inhibit energy expenditure through its suppressive effects on sympathetic nervous system activity and brown adipose tissue thermogenesis(199)

Individuals with obesity have lower basal ghrelin levels(200), possibly reflecting metabolic adaptation to a chronic state of positive energy balance(200), and also have an attenuated postprandial ghrelin suppression(201), which may diminish satiety and may drive to further overeating and weight gain(201). Interestingly, altered circadian rhythms due to sleep disorders have been consistently associated with elevated ghrelin concentrations, increased hunger feelings, and greater food intake in cross-sectional studies(202–207). As sleep curtailment due to shift work and other factors is a hallmark of modern societies, these findings may emphasize an important role for ghrelin in today’s obesity pandemic. However, there is a lack of evidence showing that ghrelin may be causal for weight gain in longitudinal studies. Indeed, only one human study reported such association, that is, baseline ghrelin predicted greater weight regain after a period of weight loss(208). However, in another prospective study, baseline ghrelin did not predict future weight gain (but predicted greater food cravings) after a 6-month follow-up period(209). Accordingly, some researchers question the role of ghrelin with regard to its relevance in sleep curtailment-induced obesity(210).

In summary, ghrelin has potent effects on energy metabolism as it suppresses metabolic rate while simultaneously stimulating food intake. However, it remains to be established in future prospective studies whether ghrelin can predict weight gain susceptibility and whether it may constitute a hormonal trait of the thrifty phenotype. In this context, the investigation of ghrelin responses to dietary perturbations such as fasting and overfeeding rather than the measurement of basal (overnight-fasted) ghrelin levels might be of special interest.

Conclusion

New treatment options acting on energy metabolism in concert with dietary restriction paradigms are required to limit weight gain in obesity-prone individuals or to induce weight loss in individuals with obesity. As highly controlled overfeeding studies demonstrated that there are individuals more resistant or more susceptible to the development of obesity, one approach to identify new candidate pathways for clinical interventions might be the phenotypic characterization of subjects, especially those more resistant to weight gain, to identify the metabolic characteristics that differentiate them from subjects more susceptible to obesity. The high-level concept of human energy expenditure phenotypes, which allows to distinguish obesity-prone (thriftier) from obesity-resistant (more spendthrift) individuals via the assessment of short-term dynamic changes in metabolic rate and fuel utilization in response to acute manipulations of energy intake (e.g., fasting or overfeeding), is an important first step. However, the underlying (low-level) biological mechanisms that characterize these phenotypes are not fully elucidated. As this review indicates, further attention should be drawn to the efficiency of adipose tissue lipolysis, brown fat activity, FGF21 secretion during dietary protein restriction, and to the function of both branches of the SNS (Figure 1). In this context, the treatment of dysfunctional lipolysis, i.e. by restoring adipocyte function or by improving catecholamine signaling, the recruitment and expansion of brown adipose tissue volume and its activation, the exogenous administration of FGF21, and the stimulation of adrenomedullary activity, might constitute new pathways to reduce weight gain susceptibility and fight the worldwide obesity pandemics.

Figure 1 – Metabolic factors determining susceptibility or resistance to weight gain.

According to recent research, a greater decrease in 24hEE during 24h fasting (defining a thriftier phenotype) and metabolic inflexibility to high-fat overfeeding quantify the susceptibility to future weight gain (“high-level concept”). The underlying (“low-level”) biological mechanisms mediating the metabolic differences between individuals and conferring greater susceptibility to weight gain may encompass impaired lipolysis, reduced brown adipose tissue activity, a blunted FGF21 secretion in response to protein-restricted hypercaloric diets, central sympathetic overactivity (≙ high norepinephrine secretion), and low sympathoadrenal activity (≙ low epinephrine secretion).

24hEE, 24-hour energy expenditure

Abbreviations

24hEE

24-hour energy expenditure

BAT

brown adipose tissue

BMI

body mass index

FGF21

fibroblast growth factor 21

RQ

respiratory quotient

SNS

sympathetic nervous system

UCP1

uncoupling protein 1

WAT

white adipose tissue