The emotional cost of Anorexia Nervosa: evidence of stress response dysfunctions

Anorexia Nervosa (AN) is a severe psychiatric disorder which mainly affects adolescent women. Characterized by an intense fear of gaining weight coupled with body-dysmorphia and an absolute control over food, individuals with AN engage in restrictive dieting regimens and rigorous physical activity to pursue an unhealthy ideal of thinness, ultimately reaching a weight that is lower than normal. These behaviors are sustained and exacerbated by psychological traits, such as perfectionism and low self-esteem, psychiatric comorbidities and a socio-cultural environment that promotes specific standards of body images, thus worsening the condition. AN is associated with high relapse and mortality rates; nevertheless, its etiopathogenesis remains poorly understood, and available pharmacological treatments are largely symptomatic, failing to address the underlying causes of the disorder. Subjects with AN display hypercortisolemia and an altered emotional state, which suggest a hyperactivation of the Hypothalamic-Pituitary-Adrenal (HPA) axis, the neuroendocrine stress response system, as potential driver of AN. Glucocorticoids, the main molecular mediators of the HPA axis, are known to modulate cognitive and emotional processes, which are dysfunctional in patients with AN. An impaired ability of the HPA axis stress responsivity to properly cope with stressful conditions may compromise brain and body homeostasis, ultimately sustaining the maintenance of the anorexic phenotype and preventing full recovery. This project aimed at investigating the neurobiological bases of AN focusing on the involvement of the HPA axis through a translational approach in different phases of the pathology, from the acute state of the disease until weight recovery. To this end, we first characterized a cohort of patients diagnosed with AN in terms of body weight, psychological attitude toward physical exercise and stress response-dependent markers, such as circulating levels of cortisol and cholesterol, at the first evaluation and after 1-year follow up, in the attempt to define potential stage-dependent diagnostic biomarkers. At the preclinical level, we employed the well-established activity-based anorexia (ABA) model on adolescent female rats, which recapitulated the self-induced weight loss and hyperactivity typical of human AN by combining food restriction and physical exercise. To unravel alterations of the HPA axis in the ABA phenotype, we measured peripheral component of the axis, such as circulating levels of corticosterone and cholesterol and gene and protein expression of corticosterone synthetic pathway in the adrenal glands. Moreover, we analyzed genomic glucocorticoid signaling at the central level, specifically in the amygdala and the hippocampus, two brain areas highly responsive to glucocorticoids, critical for emotional and cognitive functions, and involved in the HPA axis regulatory feedback. Since glucocorticoids can mediate cognitive functions through fast acting non-genomic mechanisms, we evaluated structural, morphological and molecular determinants in the hippocampus of ABA rats and their cognitive-related outcomes. All the analyses were performed at two time points: at the establishment of the anorexic phenotype at post-natal day (P) 42, corresponding to the acute phase of the disease, and after a short-term weight recovery at P49, to unravel potential molecular and functional scars induced by the ABA phenotype that persist even after weight recovery. Then, to verify the involvement of the HPA axis in sustaining the acute anorexic phenotype, we pharmacologically modulated the glucocorticoid receptor (GR) by treating ABA rats during the induction with RU486, a GR antagonist, in the attempt also to ameliorate the ABA phenotype and central glucocorticoid- mediated genomic signaling. In addition, this project aimed to dissect the relationship between AN and psychiatric comorbidities, since it is well known that a history of AN can lead to the development of such comorbidities, although the underlying molecular mechanisms remain still unclear. To investigate emotional impairments as consequences of AN induction, we evaluated whether ABA induction during adolescence was able to alter the emotional state of rats after recovery and their ability to cope with an acute stressor, with a molecular focus on neuroplastic adaptations mediated by the Brain-Derived Neurotrophic Factor (BDNF) within the prelimbic and infralimbic subregions of the prefrontal cortex. On the other hand, to investigate emotional instability as a risk factor for AN, we induced the ABA phenotype in Serotonin Transporter Knock-out (SERT KO) rats, a preclinical model of anxiety- and depressive-like behaviors, and assessed behavioral differences in body weight loss, food intake and hyperactivity. Notably, in this experiment we also included male animals to uncover potential sex-dependent patterns of ABA phenotype, which may help explaining differences in clinical features between men and women. Overall, our analyses revealed a chronic state of hypercortisolemia in adolescent patients with AN that persisted until a 1-year follow up, regardless of BMI recovery. Turning into the preclinical model, the ABA rats reproduced the hypercortisolemic state of underweight patients during the acute phase of the disease. Circulating corticosterone levels in ABA animals aligned with its synthesis in the adrenal glands: at the acute phase increased activity of MC2R receptor led to enhanced gene and protein expressions of key players in corticosterone synthetic pathway, as SR-B1, StAR, Cyp11a1 and 3-HSD; after recovery, the entire pathway was reduced, resulting in lower production and release of corticosterone. In the amygdala and the hippocampus, at the acute phase, the ABA induction reduced GR levels in cytosolic and nuclear compartments and disrupted its transcriptional activity, as shown by expressions of GR-responsive genes, as Gr, Gr, Fkbp5, Sgk1 and Gilz. After weight recovery, GR levels were still reduced in the nuclear fraction of the amygdala while restored in the hippocampus, suggesting a region-specific recovery of glucocorticoid functions. In the hippocampus, glucocorticoid-dependent genomic alterations were coupled with enduring downregulation of their non-genomic mechanisms, as shown by reduced protein levels of membrane-bound GR, of structural markers such as CaD1, PSD95, nCadherin and Neuroligin-1. These alterations contributed to morphological rearrangement of dendritic spines and to spatial-memory impairments, which persisted even through weight recovery. Interestingly, the treatment of ABA rats with a GR antagonist partially ameliorated the ABA phenotype, reducing hyperactive behaviors on the last day of food restriction and restoring physiological GR subcellular localization specifically in the hippocampus. In addition, the deleterious impact of ABA induction during adolescence persisted even after three weeks of weight recovery: ABA animals were unable to respond to an acute challenge in an emotional-demanding task in adulthood, exhibiting a blunted emotional response to the acute restraint stress. This effect was sustained by an improper recruitment of BDNF- dependent neuroplastic mechanisms in the prelimbic and infralimbic cortices. Finally, SERT ablation induced an anxiety-like phenotype prior to ABA exposure and exacerbated features of the ABA phenotype in a sex-specific manner, with male animals showing heightened vulnerability to hyperactivity. In conclusion, our findings support the hypothesis that AN is characterized by a maladaptive response of the HPA axis that persists from the acute phase through weight recovery, ultimately impairing cognitive and emotional functioning. Pharmacological targeting of the HPA axis may therefore represent a promising strategy to ameliorate some of the behavioral and molecular alterations associated with the disorder. In addition, our data revealed that a history of AN during adolescence can give rise to psychiatric comorbidities even later in life, suggesting that these conditions may represent risk factors for relapse or unsuccessful recovery. Our data also provide novel insights into the characterization of the ABA phenotype in male individuals, which have historically been underrepresented in both clinical and preclinical research, underscoring the importance of investigating the behavioral and molecular bases of psychiatric disorders in a sex-specific manner. Although these studies represent only an initial step toward disentangling the causal relationship between AN and psychiatric comorbidities, our results emphasize the need to consider individuals’ emotional state across the entire course of the disease, both as prodromal and consequential factors, since they may inform the development of personalized therapies tailored to individual characteristics.