APPETITE DISORDERS

Appetite Disorders: From Binge Eating to Anorexia Nervosa

Introduction ‘Eating disorder’ is a term that covers a variety of psychological illnesses in which individuals express abnormal eating behaviors often resulting in either insufficient or excessive food intake. Eating disorders include anorexia nervosa,bulimia nervosa, and binge eating (which is commonly associated with obesity). Food intake, like drinking or reproduction, is a complex motivated behavior essential for the survival of the organism, and the challenge is to understand how it can become aberrant to the extent that it can become life threatening. In this chapter we first detail the main symptoms of the three major eating disorders and then we describe the rodent models that are commonly used to explore them.

While a finely tuned physiological system that controls the balance between energy expenditure and energy intake has evolved, in mammals, there is no clear evidence to support the idea that eating is simply an automatic response to an acute energy demand. In fact, food intake involves endocrine and central effectors directly linked to energy homeostasis, but also to neuronal circuitries regulating stress, emotions, reward, biological rhythms, learning, individual experience with food (anticipatory adaptation) and, in humans, socio‐cultural factors. Eating depends on a concomitant functioning of a hard‐wired homeostatic circuitry that is almost entirely identical between mammals, resulting from evolutionary selection, together with a more flexible non‐homeostatic circuitry whose functions vary according to individuals’ experiences and/or epigenetic variations.

Etiology of eating disorders

The increased prevalence of eating disorders raises the question of whether deregulation of brain systems occurs due to subtle changes in the way we consume food in our modern societies. Amongst the various advanced hypotheses, one can consider a neurodevelopmental problem (such as an altered leptin signal during a critical developmental window, that has been linked, for example, to obesity in rodent models). Alternatively, there could be an alteration in the integration of peripheral metabolic messages at the hypothalamic or mesolimbic levels. There could even be epigenetic alterations that arise, for example, as a consequence of exposure to endocrine disruptors. Whatever the causes that trigger the occurrence of these diseases, whose etiology remains extremely complex and difficult for psychiatrists to precisely define, fundamental research using relevant animal models of eating disorders is critical to better decipher the mechanisms that govern the finely tuned and reciprocal organization between the brain systems and the peripheral signals.

Eating disorders: subtypes and diagnosis

The diagnosis ‘eating disorder’ covers a diverse spectrum of conditions with many phenotypic differences, which are apparent between and within eating disorders cohorts. Here we compare three different disorders at both ends of the spectrum of body weight regulation: eating too much or too little. According to the Diagnostic and Statistical Manual of Mental Disorders (DSM‐5), there are three subtypes of eating disorders: anorexia nervosa (AN), bulimia nervosa (BN), and binge eating disorder (BED). The AN subtype is mainly characterized by distorted body image and excessive dieting. In the BN subtype, the individual is engaged in recurrent episodes of binge eating followed by inappropriate purging behaviors, such as self‐induced vomiting, episodes which do not occur in the AN subtype. The BED subtype was recently included in DSM‐5 as its own category of eating disorder. It is characterized by ‘recurring episodes of eating significantly more food in a short period of time than most people would eat under similar circumstances, with episodes marked by feelings of lack of control’ (DSM‐5).

As is usually indicated by psychiatrists, the subtype determination at the time of diagnosis should be considered carefully since one study found that the majority of women with AN crossed over between the other subtypes, assessed during a seven year period (Eddy et al., 2008). Interestingly, despite divergence in behaviors associated with the AN, BN, and BED subtypes, they share numerous physiological and endocrine disturbances, such as alterations in ghrelin, cortisol, and leptin as well as endocrine‐linked diseases such as osteoporosis and osteopenia.

Anorexia nervosa

Anorexia nervosa is a severe psychiatric disorder, which primarily affects women aged up to 25 years of age. AN has one of the highest mortality rates of all psychiatric diseases (Weiselberg et al., 2011). In a 21‐year follow‐up study, Löwe et al. (2001) showed that 16% of AN patients died due to the consequences of the illness: of these, about 50% died because of somatic complications (renal, cardiac, bone, and digestive pathologies) leading to heart attacks and the other 50% committed suicide. In fact, the course of AN is extremely variable, with 50–60% of individuals with AN recovering, 20–30% partially recovering, and 10–20% remaining chronically ill (Löwe et al., 2001). AN is often associated with psychiatric comorbidities including depression, anxiety, obsessive compulsive or personality disorders and chemical drug abuse (Erdur et al., 2012).

AN diagnosis requires the fulfillment of three major criteria (DSM‐5, American Psychiatric Association, 2013). The first criterion concerns a severe and persistent restriction of energy intake leading to significantly low body weight within the context of what is minimally expected for age, sex, developmental trajectory, and physical health. The second criterion is the intense fear of gaining weight or of becoming fat, or persistent behavior that interferes with weight gain. The third criterion is related to the disturbance in the way AN patients perceive their body weight or shape (dysmorphophobia), with undue influence of body shape and weight on self‐evaluation associated with persistent lack of recognition of the seriousness of their current low body weight. Another criterion removed in the last DSM version concerns amenorrhea, or the absence of at least three menstrual cycles. This criterion was deleted since it cannot be applied to males (AN occurs at a ratio of 1 male to 13 females), pre‐menarchal females, females taking oral contraceptives, and post‐menopausal females.

Moreover, some reports describe individuals who exhibit all other symptoms and signs of AN but still report some menstrual activity. Restrictive AN induces dramatic physiological and psychological consequences on health and may lead to central and/or peripheral reprogramming that permits the organism to endure the reduced energy supply. AN also includes other symptoms, such as behavioral hyperactivity, observed in 31–80% of the cases (Hebebrand et al., 2004a), and metabolic disturbances. In particular, metabolic hormone levels (such as ghrelin or cortisol) are increased and the endocrine function of adipose tissue is modified (resulting in decreased concentrations of leptin). Additionally, osteoporosis (another main complication of AN affecting 20–50% of cases) occurs with bone fracturing and osteopenia appearing in 92% of AN patients (for review see Méquinion et al., 2013).

AN is increasingly recognized as an addictive behavior disorder linked to food deprivation, weight loss or physical activity. Dysfunction of the central dopaminergic system (that is involved in reward behaviors) has been observed: AN patients showed decreased dopaminergic metabolite levels in cerebro‐spinal fluid as well as increased D2/D3 dopamine receptor density (Box 10.1: Kaye, 1999; Bailer et al., 2013). Finally, the etiology of AN is likely to involve interactions between psychosocial and genetic risk factors (risk  3.6%), as recently reviewed in Raevuori et al. (2014).

Bulimia nervosa

Although described as an ominous variant of AN in the 1970s, bulimia nervosa (BN) is now recognized as a distinct syndrome. It is a disorder defined by: (i) binge episodes, during which an individual consumes an unusually large amount of food, (ii) experiences a sense of loss of control over his or her eating, (iii) the frequency of binge episodes should be at least once per week for three or more months, and (iv) the binge episodes should be followed by a calorie‐sparing or inappropriate compensatory behavior (i.e., purging, laxatives, fasting, etc.). Many of the same medical complications described for AN are also true for BN patients. In addition, excessive vomiting may result in enamel erosion, calluses on the fingers, and sialadenosis (a swelling of the salivary glands). Unlike AN, BN patients generally do not have bone disturbances. Typically, BN is female gender‐specific and the onset occurs at a young age with a prevalence of about 1% among young females (Smink et al., 2014). Although research on mortality and treatment success in BN is sparse, the mortality rate for BN is estimated to be a0–3%, and about 50% of patients are free of symptoms after 5 years of treatment (Keel and Mitchell, 1997). Although psychosocial factors are considered to be involved in the development of BN, it has been shown that >50% of the risk of developing BN is accounted for by genetic heritability (Bulik et al., 1998). Different personality traits and temperament, sometimes already apparent before the onset of disease, have been linked to BN as well as AN‐R, such as anxiety, harm avoidance, poor interoceptive awareness, ineffectiveness, and self‐directedness. Moreover, BN patients tend to have poor impulse control, engage in greater novelty‐, pleasure‐, and stimulus‐seeking behavior and are less paralyzed by concerns with future consequences (the last of which also occurs in AN) (Villarejo et al., 2014).

Binge eating typically occurs in the absence of hunger or when deprived of food, and is specifically characterized by overeating sweet and fatty foods in a relative short time period. It can be triggered by negative emotions, mood lability and stress; the subsequent binge–purge cycle reduces dysphoria and/or anxiety (Smyth et al., 2007). It has also been shown that BN patients are more sensitive to reward. Kaye et al. recently reviewed what is known about the neurobiology that is shared between pathways involved in drugs of abuse and the extreme food ingestion observed in BN (Kaye et al., 2013). In addition to the psychological characteristics, physiology is also disrupted as was described for AN earlier, such as increased cortisol and ghrelin levels and decreased leptin levels (Monteleone and Maj, 2013)

Binge eating disorder In the early 1990s, clinicians recognized a separate group of eating disorder patients with similar compulsive‐overeating problems as those seen with BN, but with the difference that the binge was not followed by inappropriate compensatory behavior (DSM‐5). This newly identified syndrome, named binge eating disorder (BED), further differs from AN and BN in terms of age at onset, gender and racial distribution, psychiatric comorbidity, and association with obesity (Spitzer, 1991), and occurs more frequently. BED is often seen in obese individuals, but is distinct from obesity per se regarding levels of psychopathology, weight and shape concerns, and quality of life. The physiological consequences are not widely studied. However, the prevalence of obesity occurring in BED patients is high and with that many additional metabolic disorders, such as type‐2 diabetes, occur. In contrast to what is known for AN and BN patients, BED patients have high leptin levels; however, as obese patients have high leptin levels, this endocrine disturbance is most likely secondary to the obesity rather than a causal factor for BED. BED patients have elevated basal levels of the stress hormone, cortisol, and a blunted response of the hypothalamo–pituitary–adrenal (HPA) axis to a psychological stress test.

Cortisol is positively correlated with food craving in these patients. These effects on the HPA axis are specific to BED patients as weight‐matched obese controls do not show this HPA axis disturbance (Rosenberg et al., 2013). The high proneness to stress is thus a trigger for binge eating, not only in BED patients but also for BN patients. It has also been suggested that binge eating can be viewed as an addiction‐like behavior towards foods rich in fat and sugar; indeed, it has recently been proposed that the term ‘eating addiction’ may serve to describe this kind of behavior (Hebebrand et al., 2014), a behavior characteristic of BED but perhaps less so for BN. These similarities and differences in the endocrine and behavioral parameters characteristic of BED and BN suggest that common, but also divergent, neural signaling mechanisms are involved.

The unknown etiology of AN, BN, and BED renders these complex psychiatric diseases difficult to treat. Indeed, pharmacological treatments seem to have little efficacy during the acute phase of the illness or in preventing a relapse (Barbarich‐Marsteller, 2007). We need a much better understanding of the physiological mechanisms that sustain these diseases, which likely include endocrine/neuroendocrine alterations, adaptation/modification of the energy metabolism signals at central and peripheral levels, and possibly even changes in the gut microbiome. Recently, Million et al. (2013) showed variations in the bacterial load of Lactobacillus species with body mass index (BMI) with, in particular, an increased occurrence of Lactobacillus reuteri with increasing BMI:7, 8, 34, and 22% for anorexic, lean, overweight, and obese individuals, respectively. These data clearly suggest a ‘dose‐dependent’ relationship between some species of bacteria present in the human gut and BMI.

Using pertinent animal models of eating disorders may facilitate assessment of different aspects of the disease, exploring central and peripheral mechanisms as well as short‐ and long‐term components. For this purpose, the choice of the model might need to fulfill most of the validity criteria, face, predictive and construct validity, described by Willner (1984).

Animal models of eating disorders

Animal models can provide invaluable insight for psychiatric disorder research, especially when the etiology is well characterized or when genome wide association study (GWAS) enlightens potential human risk genes for which homologous genes can be easily mutated or deleted in rodent models. Unfortunately, as mentioned later, the lack of such information and the complex nature of AN, BE, and BED have hampered the development of appropriate animal models. The current models described can only provide a few characteristic traits of the human psychiatric disease. In particular, even if we can assess face validity and some aspects of construct validity, predictive validity is impossible to obtain in rodent models since the current treatments used in humans do not give satisfying results for all the patients treated for eating disorders.

Animal models of ‘binge eating’

Genetic models

Although genetic traits have been proposed to play a role in BED and BN (Trace et al., 2013), we lack a genetic animal model that sufficiently replicates elements of these disorders. Although initially mutations in the melanocortin 4 receptor (MC4R), a receptor expressed in the brain that is downstream from leptin signaling and important for the regulation of energy balance, were reported to be associated with binge eating syndrome (Branson et al., 2003), subsequent studies did not confirm this initial observation (Hebebrand et al., 2004b), thus making it unlikely that MC4R knockout animals will be relevant for the study of binge eating.

Environmental models linked to stress and food availability

Binge episodes (which constitute a large proportion of the total dietary intake) consist primarily of carbohydrates and fats, with relatively little protein (Van der Ster Wallin et al., 1994), often in the form of dessert and snack foods (Gross et al., 1986). Both the fat and sugar intake characteristic of binge episodes, and the best‐known trigger of binge episodes, namely stress, have been represented in animal models of binge eating. In addition, food restriction is a key factor in all animal models of binging, whether spontaneous or forced.

(a) Stress models

In both BN and BED patients, stress is a major trigger for binge eating and this aspect is specifically addressed by the stress‐induced hyperphagia model in which binge eating is provoked by subjecting female rats to several cycles of restriction and unlimited access in combination with acute stress (Hagan et al., 2002). When the model was first introduced by Hagan and Moss (1997), rats were presented with a high‐fat food together with less preferred normal chow; during the ‘binge meal’ the rats mainly consumed the high‐fat food. Later on, however, the group of Hagan also showed that only a small quantity of palatable food (i.e., priming with palatable food) is sufficient to provoke a binge in rats on preferred chow. They also provided evidence that animals are not binging for metabolic need, but for reward as a hunger state was not necessary to provoke a binge. Interestingly, foot shock is not the only stressor with a capacity to evoke binges: female mice with a history of food restriction display binge behavior for palatable high‐fat food during a chronic variable stress paradigm (Pankevich et al., 2010). Both of these stress‐linked models of binge eating are dependent on restriction cycles prior to stress‐induced binging and as such mimic binge eating observed in BN patients, and only simulate a subgroup of BED patients that have a history of dieting.

(b) Palatable food paradigms

Although not every BED patient has a history of dieting, restricting food intake before and after binge eating has been described to occur frequently (de Zwaan et al., 1994). In rats, spontaneous reductions in food intake have been observed when animals are exposed to a limited high palatable food access paradigm. In 1998, Corwin et al. described a model in which fat was given three times a week at the end of the light period for 2 hours, which resulted over time in clear binges on the fat provided, but also in compensatory behavior with clear reductions in food intake in between the days of the 2 hour fat exposure (Corwin et al., 1998, 2004). This was also used by other research groups (Berner et al., 2008; Lardeux et al., 2013). Rats are exposed to palatable food for 30, 60, 90 min or 2 hours at the end of the light period, while receiving ad libitum chow and water throughout the experiment. The palatable food always includes a fat source (crisco, sweetened fat, margarine, high fat chow diets); the intake increases over the subsequent episodes, resulting in stable high kilocalorie intake after 4–5 sessions. Interestingly, in these binge models, the animals show a reflex adjustment in their intake of regular chow, which is why the animals do not usually increase their body weight. It has been postulated that in BED and BN patients dietary restriction predicts binge eating (Zunker et al., 2011) but both rat data provided by Lardeux et al. (2013) and Davis et al. (2007) point to the spontaneous food restriction being compensatory rather than a form of anticipatory behavior.

With respect to anticipatory restriction to a binge episode, a slightly adjusted model was described in which 30 min chow access is followed by 30 min palatable food access after animals have been deprived of food for 2 hours at the beginning of the dark period (Cottone et al., 2008). The food deprivation will increase the motivation to eat, especially at a time of day when food intake is highest, and after several sessions animals will anticipate the palatable 30 min period by reducing intake of the standard chow diet that precedes the availability of palatable food.

In addition to providing rats on a regular chow diet with limited access to palatable food for several hours on three days of the week, other models allow access for two days per week. Also in this paradigm, animals binge when switched from chow to palatable food, consume more in the two days on palatable food, and restrict intake when less‐preferred chow is provided during the five days in between (Rossetti et al., 2013). Interestingly, Rossetti et al. (2013) showed that this palatable food intake regimen resulted in increased perseverance to get a fat pellet despite a foot shock. This finding is in line with the data from Hagan et al. (2002), although in these experiments the food restriction was spontaneous whereas for Hagan et al. (2002), it was scheduled.

As mentioned earlier, binge episodes in BN and BED patients are characterized by increased intake of both fat and sugar. The earlier rodent models have the fat component in common, and it is clear that restricted access to fat specifically provokes binge eating. Sugar binging, however, has also been described, albeit with specific timing of providing the food and periods of food restriction. Colantuoni et al. (2001) showed that male or female rats, when provided with a 25% glucose solution 4 hours into the dark period after 12 hours fast, drank large amounts of the glucose solution. Later studies demonstrated similar effects with a 10% sucrose solution (Avena and Hoebel, 2003). A similar escalating sugar intake with intermittent sugar access was also shown in the model described by Bello et al. (2009); in this case, the restriction was more extreme as animals were only fed chow for 2 hours, after which the animals were provided with sugar solution. These models are difficult to match with the characteristics of binge eating in BED and BN subjects, as restricting food for 12 hours until 4 hours into the dark period is rather extreme and binges are rarely only on sugar solutions and thus do not represent patterns characteristic of eating disorders.

(c) Depression/anxiety‐like behavior

In addition to dietary history and dietary restraint being important associations with binge eating, which have been described extensively earlier, pathways of negative affect have been hypothesized as being involved in both BED and BN. It has been postulated that body dissatisfaction, in addition to dietary restraint, may result in negative emotions, such as depression, and this, in turn, causes binge eating behavior because overeating can distract people from feelingsof aversion (Heatherton and Baumeister, 1991). For all of the afore-mentioned models, tests have been performed to validate whether depression-like and anxiety‐like behavior occur. Indeed for stress‐induced hyperphagia, for restricted palatable access as well as for sugar binging, depression-like and anxiety‐like symptoms have been described (Cottone et al., 2009). In detail, Cottone et al. (2009) showed that rats subjected to a five‐day chow–two day palatable diet eating pattern exhibited anxiety‐like behavior (measured in the open field test), but, interestingly, only when tested in the chow period, and not when tested in the palatable feeding period. These results validate the anxiety described by BED patients and the subsequent reduction of anxiety after the binge. In addition, the observed blunted response of the HPA axis in BED patients is also observed in the rat models. Both Rossetti et al. (2013) (using a two‐day palatable food–five‐day chow paradigm) and Bello et al. (2014) (using a three day a week 2 hour exposure to sugar–fat mixtures), showed a blunted response to restraint stress. Moreover, Bello et al. (2014) also showed more palatable intake after the stressor, further supporting the idea that stress is an important trigger for binging behavior.

Animal models of ‘anorexia nervosa’

Genetic models

(a) Spontaneous mutation: anx/anx mice

To our knowledge, only one mouse model of spontaneous autosomal recessive lethal mutation related to AN has been described, with a mutation located on chromosome 2 (Maltais et al., 1984). This mutant anx/anx mouse arose spontaneously at the Jackson Laboratory in 1976. The prominent phenotype of these mice is an emaciated appearance, a reduced food intake, and death by 3–5 weeks of age. From at least postnatal day (P) 5, the anx/anx mouse eats less than normal littermates, despite free access to the mother (Maltais et al., 1984).

The anx/anx mouse begins to deviate significantly from the normal growth curve from around P9 and weighs half as much as normal littermates by P21. Recently, Lindfors et al. (2011) mapped the anx mutation on chromosome 2, which concerned genes encoding one of several proteins important for the proper assembly of the mitochondrial complex I. These mice also showed several deviations in the hypothalamic neuropeptidergic and neurotransmitter systems involved in the regulation of food intake and energy metabolism, which were associated with mitochondrial dysfunction and neurodegeneration/neuroinflammation processes (see review by Nilsson et al., 2013). These data were corroborated by the data obtained by Lachuer et al. (2005), who were the first to show an overexpression of genes involved in the inflammatory process in the hypothalamus of anx/anx mice. This natural genetic model of anorexia represents a perfect model to dissect mechanisms that lead to physiological dysfunctions observed in AN, especially anorexia‐cachexia. In fact, neurodegenerative processes have been described in restrictive AN patients, with a global reduction in white matter, focal reductions in gray matter in the hypothalamus and some other brain regions. However, the main limitations of this genetic anorexia model are: (i) the premature death of the mice before reaching puberty and (ii) effects on both male and female mice. In humans, prepubertal AN is rare and the incidence of this eating disorder concerns mainly teenage and young adult female individuals.

(b) Genetic deficient mouse models

A remarkable number of genetically deficient mouse models for one or multiple genes involved in the regulation of feeding behavior/reward/energy metabolism/neuroendocrine and immune systems have been developed (for review see Méquinion et al., 2015a). Moreover, in a recent review, Rask‐Andersen et al. (2010) listed putative genes involved in the control of food intake and body weight through human genetic association studies on AN patients. Among the most prevalent genes implicated in AN, those coding for the monoamine system have been well investigated. Of these, the dopaminergic and serotoninergic systems have been studied most extensively, as brain imaging studies have revealed altered activity of these pathways in AN (Bailer et al., 2013). Szczypka et al.
(1999) used a gene‐targeting strategy to inactivate specifically the tyrosine hydroxylase gene in dopaminergic neurons, sparing the production of dopamine as a precursor for adrenaline and noradrenaline. These mice, called ‘dopamine deficient mice’ became hypophagic and died from starvation at 34 days because they showed locomotor deficiencies. Routine treatment with the dopaminergic precursor L‐DOPA restored a level of food intake similar to wild‐type mice.

The serotoninergic system is also strongly implicated in feeding and satiation. Usually, pharmacologically 5‐hydroxytryptamine (5‐HT) stimulation inhibits food intake. The central action of 5‐HT is complicated by the diversity of receptors and transporters. As an example, a study on the role of 5‐HT4 receptors concluded that the overexpression of these receptors in the nucleus accumbens upregulated CART (cocaine‐ and amphetamine‐regulated transcript) in this limbic region, provoking anorexia and hyperactivity (Jean et al., 2012). Thus, because of the multifactorial response to 5‐HT, animals with specific or inducible genetic deletion of different 5‐HT receptors or transporters might be good genetic tools to study AN as well as BN/BED. Among the other genes whose actions are described in the literature (Kim, 2012), BDNF, delta opioid receptor (OPRD1), muscarinic receptor, MCH receptor, AgRP, and CRH have been suggested to be associated with AN.

Environmental models

Despite the evident relevance of the aforementioned genetic models, they only provide mechanistic data related to one specific pathway and do not completely reflect the face validity of the disease, that is, most of the symptoms observed in humans (Smith, 1989). Indeed, the use of more ‘environmental models’ that mimic most of the physiological symptoms of AN would be preferable in order to better understand the different physiological and neurobiological aspects of the disease. Initially, the most commonly used animal model, whatever the species, was the chronic food restriction model.

(a) Chronic food restriction models

Among the first experiments using this paradigm, caloric restriction was studied as a means to increase longevity by retardation of the ageing process(es) (Lewis et al., 1985) and amelioration of many pathological genetic changes during aging. Various protocols of caloric restriction are described in the literature in which the percentage of restriction ranges from 30 to 60%, with or without supplementation (in tyrosine, for example) and in different species. However, significant drawbacks to this type of model are that they do not take into account various aspects of the disease, such as self‐starvation, hyperactivity, and chronic stress, commonly described in AN patients (Méquinion et al., 2015b).
)
(b) Activity‐based anorexia (ABA

The rat model of self‐starvation developed by Routtenberg and Kuznesof (1967) addresses two aspects described in AN: hyperactivity and chronic stress. For this purpose, one rat housed in a cage equipped with a running wheel is subjected to a food restriction (1 hour of feeding per day). This model produces a rapid loss of weight and hypophagia, hyperactivity, hypothermia, loss of estrus, and increases in HPA axis activity (Hall and Hanford, 1954; Routtenberg and Kuznesof, 1967; Burden et al., 1993). Moreover, the ABA rats eat less than inactive rats fed with the same schedule, and usually starve themselves to death. In many aspects this model mimics numerous physiological alterations observed in AN. However, as specified by Klenotich and Dulawa (2012), the ABA paradigm is strongly dependent on some factors that can amplify or reduce some parts of the phenotype (Table 10.2), such as the choice of rodent strain (more or less resistant to ABA), the sex of the animal, the age, the temperature (increasing the temperature to 32 °C strongly reduces the ABA behavior; Cerrato et al., 2012). The group of Boakes (Boakes and Juraskova, 2001; Boakes, 2007) demonstrated that the ‘self‐starvation’ observed in ABA rats might reflect both the reduced palatability of the dry chow for a dehydrated animal and satiety signals from a stomach full of water. Thus, giving hydrated food during the 1 hour feeding schedule essentially abolishes the ABA phenotype (rapid weight loss, hyperactivity, etc.). Currently, we have developed an adaptation of the ABA model in female mice that aims to follow the long‐term physiological alterations induced by a combination of physical activity and food restriction.

(c) Chronic stress models

Other environmental models are based on chronic stress associated with, or not with, food deprivation. They include tail pinching, cold swimming, even brain stimulation, and chronic separation. In fact, it has been reported that the endocrine changes induced by a life stressor are frequently involved with some forms of eating disorders. The wide use of chronic stress to mediate eating behavior is justified because it does not require the manipulation of food availability. However, even if such physical stress leads to weight loss and might contribute to a loss of appetite, excessive manipulations can physically harm the animals, which is in contrast with the psychological stress experienced by AN patients.

(d) Separation‐based anorexia (SBA)

Among the stress models, the ‘separation‐based anorexia’ model , first described by van Leeuwen et al. (1997), permitted study of both the impact of chronic stress and caloric insufficiency. Recently, we further characterized this model (Zgheib et al., 2014) by using eight week old female mice separated and fed with a time‐restricted food access for up to ten weeks. The results obtained showed marked alterations in body weight (20–25% body weight loss), fat mass, lean mass, bone mass acquisition, reproductive function, GH/IGF‐1 axis, and hypoleptinemia. Moreover, mRNA levels of markers of lipogenesis, lipolysis, and the brown‐like adipocyte lineage in subcutaneous adipose tissue are also changed.

The two ‘environmental animal models’ (SBA and modified ABA) are currently among the best models available (and even the most ethological) that allow long‐term studies of the impact of chronic food restriction associated with voluntary physical or chronic stress on physiology and neurophysiology (energy metabolism, reproduction, bone/fat regulation, hypothalamic alterations, etc.). Despite the absence of self‐starvation, which is a cognitive characteristic of AN patients, such models fulfill most of the aspects of construct validity and face validity (same symptoms in the short‐ and long‐term). The predictive validity might be studied with potential current pharmacological treatments used in AN patients. They provide new windows of opportunity to assess the mechanisms responsible for the maintenance of these alterations on different tissues often not available in patients (brain, bones, fat, muscle, liver, intestine, microbiota, etc.). They will also make it possible to determine whether the dramatic outcomes in patients might be related to a specific deregulation of one or many biological factors that can be considered as markers of the disease and its evolution.

Underlying mechanisms and targets for treatment

Over the past 20 years or so, we have learned a great deal about the neurobiology underpinning appetite control and its regulation, including regulation by the endocrine system. There is a general expectation that we will discover that eating disorders represent a malfunction of neuroadaptation of the appetitive networks, but the neurobiological substrates and mechanisms remain completely unknown. A critical question remains: what could trigger an individual at a given moment to suddenly express a complete deregulation of their eating behavior? Among the biological parameters that are specifically altered in AN and BN patients, endocrine, immune, bone, metabolic systems as well as neuromediators regulating appetite and feeding (homeostatic or non‐homeostatic pathways), represent candidate systems that may have a role in these diseases. It has been suggested that these systems first adapt to starvation but often become directly involved in the complications of the disease (Estour et al., 2010). Regarding endocrine control, the changes in metabolic hormones are rather similar for AN and BN patients and may be linked (or even secondary) to the changes in body weight that occur in these disorders. By contrast, in BED, the various physiological alterations are closely related to those observed in obesity. The binge eating behavior is usually attributed to a coping strategy to alleviate a chronic/inescapable stress. Finally, we cannot neglect diet history, the psychological personality and traumas that might drive these patients to adopt such unadapted behavior. Are these alterations a cause or a consequence of the illness? Are endocrine predictors of AN or BN useful to decipher the evolution of the disease?

The physiopathological symptoms of AN and BN are tuned to find adaptive metabolic solutions to preserve energy before reaching a point that leads to exhaustion, mainly due to somatic and psychiatric complications (Beckman et al., 2007; Estour et al., 2010; Roux et al., 2013). The use of animal models is thus useful to decipher the mechanisms involved at different levels, both peripherally and centrally, in order to find more suitable therapeutic treatments than those currently used. Comprehensive data obtained from environmental animal models are scarce and often refer to one particular hormone or one particular neuronal circuit. Moreover, we must keep in mind that the psychiatric aspects of the disease (impulsivity, compulsivity, dysmorphobia, anxiety trait, etc.) cannot be reproduced in totality in rodent models. Among the underlying mechanisms that might participate directly or indirectly in the maintenance of the disease in a vicious circle are the hypothalamo–pituitary–adrenal axis (HPA), the ying–yang endocrine system involved in feeding, namely, leptin and ghrelin, the motivational/reward dopaminergic system, and the serotoninergic system. Nonetheless, we cannot exclude the involvement of other processes, less well documented in the literature for AN, BN, and BED: the opioid system, the endocannabinoid system, the immune system, the impact of environmental factors (epigenetic modifications), and neurodevelopmental alterations (prenatal stress) for example.

The complex etiology of eating disorders and the current crossing from one disease to another during the lifetime make therapeutic approaches difficult to identify for the clinician. The therapeutic armory for eating disorders must be considered as meager, even if various approaches are tested using clinical management, cognitive behavior therapy, interpersonal psychotherapy, art therapy associated or not with psychopharmacological treatment through the use of antidepressants (in particular selective serotonin reuptake inhibitors) or antipsychotics. In light of the low efficacy of these treatments (see Hebebrand and Albayrak, 2012), it is important to focus on other possibilities involving not only appetite modulators controlling the homeostatic part of eating behavior, but also those affecting the non‐homeostatic cognitive, emotional and rewarding components of food intake. The observed physiological changes in AN and BN may represent not only homeostatic adaptations to severe chronic food restriction, but also might participate to the development and/or the maintenance of aberrant non‐homeostatic behaviors (self‐starvation, binge eating, mood disorders, amongst others).

Glossary

1. animal model: A valuable animal model fulfills three main criteria: the construct validity concerns the accuracy with which the model measures what it is intended to measure according to the theoretical rationale; the face validity concerns the attempt to mimic diagnostic criteria of the psychiatric conditions; and the predictive validity concerns the success of predictions made from the model and the ability to make consistent predictions about a criteria of interest (anxiety…), similarity of pharmacological responses.

2. anorexia nervosa (AN): An eating disorder mainly characterized by distorted body image and excessive dieting leading to severe weight loss with a pathological fear of becoming fat.

3. bulimia nervosa (BN): An eating disorder where an individual is engaged in recurrent episodes of binge eating followed by inappropriate purging behaviors, such as self‐induced vomiting, episodes which do not occur in the AN subtype.

4. binge eating disorder (BED): An eating disorder recently included in DSM‐5 and characterized by ‘recurring episodes of eating significantly more food in a short period of time than most people would eat under similar circumstances, with episodes marked by feelings of lack of control.’

5. cachexia: General weight loss and wasting occurring in the course of a chronic disease or emotional disturbance or malnutrition.

6. Diagnostic and Statistical Manual of Mental Disorders (DSM‐5): ‘This manual is the standard classification of mental disorders used by mental health professionals in the United States. It is intended to be used in all clinical settings by clinicians of different theoretical orientations. It can be used by mental health and other health professionals, including psychiatrists and other physicians, psychologists, social workers, nurses, occupational and rehabilitation therapists, and counselors. DSM‐5 can also be used for research in clinical and community populations. It is also a necessary tool for collecting and communicating accurate public‐health statistics.’ (http://psychiatry.org/psychiatrists/practice/dsm).

7. dysphoria: A mood of general dissatisfaction, restlessness, depression, and anxiety; a feeling of unpleasantness or discomfort.

8. epigenetic alterations: ‘epigenetic’ refers to heritable changes in gene expression (active versus inactive genes) that does not involve changes to the underlying DNA sequence. Thus, a change in phenotype without a change in genotype. Epigenetic change is a regular and natural occurrence but can also be influenced by several factors including age, the environment/lifestyle, and disease state. Epigenetic modifications can manifest as commonly as the manner in which cells terminally differentiate to end up as skin cells, liver cells, brain cells, etc. Or, epigenetic change can have more damaging effects that can result in diseases such as cancer. At least three systems including DNA methylation, histone modification and non‐coding RNA (ncRNA)‐associated gene silencing are currently considered to initiate and sustain epigenetic change.

9. Genome Wide Association Study: Genome‐wide association studies are a way to identify genes involved in human diseases. This method searches the genome for small variations, called single nucleotide polymorphisms or SNPs (pronounced ‘snips’), which occur more frequently in people with a particular disease than in people without the disease. Each study can look at hundreds or thousands of SNPs at the same time. Researchers use data from this type of study to pinpoint genes that may contribute to a person’s risk of developing a certain disease.

10. gut microbiome (gut microbiota): Formerly called gut flora, it is the name given today to the microbe population living in our gut/intestine. It contains more than tens of trillions of microorganisms, including at least 1000 different species of known bacteria with more than 3 million genes (150 times more than human genes). Microbiota can, in total, weigh up to 2 kg. One third of our gut microbiota is common to most people, while two thirds are specific to each one of us. In other words, the microbiota in ones’ intestine is like an individual identity card.

11. negative affect (psychology definition): This is an internal feeling or emotion which is typically experienced after one has failed to complete a task or goal, or where they have completed the task but at a lower than required standard.

12. stimulus/sensation‐seeking behavior: This is the tendency to pursue sensory pleasure and excitement. It is a trait of individuals (human and other animal species) who go after novelty, complexity, and intense sensations, who may take risks in the pursuit of such experience.

By Mathieu Méquinion, Susanne la Fleur, and Odile Viltart in "Neuroendocrinology of Appetite", first edition, edited by Suzanne L. Dickson and Julian G. Mercer, John Wiley & Sons, UK, 2016, excerpts pp.201-219. Adapted and illustrated to be posted by Leopoldo Costa.

Share this post