Anorexia nervosa (AN) is a severe and chronic psychiatric disorder, resulting from a voluntary food restriction, vomiting, use of laxatives and excessive exercises, leading in dramatic weight loss and high mortality. AN is a multifactorial disease involving genetic and epigenetic factors supporting that AN is a metabo-psychiatric disorder. The molecular mechanisms involved in the etiology of AN remain unclear. One work reported gene expression by RNA sequencing in peripheral blood before and after weight restoration in 6 AN patients (Kim 2013), and one RNA sequencing in human iiPSC-derived neurons from 4 patients and 4 controls (Negraes 2017). To date, the profile of expression of genes and proteins in AN is undetermined.

In this study, our goal is to identify specific gene expression signatures from circulating blood nuclear cells to decipher the pathophysiology of AN and characterize biomarkers that can be used for diagnostic or prognostic of AN.

All consented participants are recruited at Sainte-Anne Hospital, Paris, France, using DSM5 criteria. They had a blood draw in Paxgene tube for the collection of RNAs. Total RNA was extracted from peripheral blood mononuclear cells of 15 patients suffering of AN and 15 healthy controls. All messenger RNAs are sequenced on a Novaseq plateform. Reads are aligned to the human genome 19 and statistical analyses on the read counts for differentially expressed genes are computed with DESeq2.

The total RNA sequencing allows us to identify 673 dysregulates genes (p adjusted value <0.01, fold change >1,5). Among them, 248 are down-regulated and 425 are up-regulated genes in AN patients compared to controls. From them, 151 transcripts are annotated as pseudogene and 45 are referenced as antisense RNA. Of the 522 remaining transcripts, 424 correspond to a transcript or protein annotated by HGNC and ENSEMBL and 93 are known pseudogenes. A large number of proteins resulting from the expression of deregulated genes interact with each other and form a statistically enriched network impacting biological processes. They are mainly increased and acting in the cellular machinery allowing protein synthesis (biological process: transcription, ribosome, spliceosome and mitochondria). In contrast, down-regulated genes present an enrichment in genes involved in immunity pathways. Finally, several genes are also expressed in the brain. We observed a significant enrichment of genes expressed in the blood and brain tissues.

We identify specific profiles of gene expression in AN. Several genes are both blood and brain tissue expression. Some genes are good candidates for biomarker of the diagnostic in AN that need to be investigated in a longitudinal study to evaluate their useful as prognostic biomarker of AN.

This work is supported by Fondation de France & Fédération Recherche sur le Cerveau.

None Declared

Schizophrenia (SZ) is a severe and frequent mental disorder that has multifactorial origins (genetic but also environmental vulnerability, gene environment interactions). Our team has shown that in France smoking is common among these patients and often begins before the onset of symptoms. This calls into question the hypothesis of self-medication of psychotic symptoms and cognitive disorders by tobacco consumption, and raises the question of specific interactions between nicotine consumption and the genes of the dopaminergic and cholinergic nicotinic systems, in particular with regard to adolescence, period of neurodevelopmental vulnerability.

to study (i) the interactions between the DA system and exogenous nicotine in a pre-clinical mouse model (transgenic for DAT) (ii) the impact of smoking on the psychotic and clinical phenotype in a national cohort of SZ patients (iii) interactions between tobacco consumption and several genetic polymorphisms of the dopaminergic and nicotinic system in SZ population. Hypothesis: Disruption of the balance between DA and nicotine systems by nicotine consumption in adolescence may be a key neurobiological mechanism for the emergence of SZ disorders.

The characterization of the model is behavioral (anxiety, memory, social interactions, locomotion, motor coordination) but also biochemical. For the clinical approach, we exploit the clinical / cognitive / genetic data of a national cohort (Fondamental foundation) and another smaller and local cohort.

We demonstrate, for the clinical study, that some clinical and cognitive characteristics are associated with tobacco use in schizophrenia patients, with more cognitive distubrances in smokers, against the self medication hypothesis. Genes envionment interactions also demonstrate associations with genes involved in dopaminergic system. Regarding the preclinical study, we show a gene environment interaction, as heterozygote mice for the DAT gene exposed to Nicotine durig a critical neurodevelopmental window ( adolescence) show a loss of familiarity in the social memory test and a loss of cognitive flexibility at the set shifting test, similar to what is found in schizophrenia.

TBetter characterization of patients with schizophrenia is necessary to better understand the physiopathology of this disease and explore new personalized preventive and therapeutical targets. We show that tobacco is associated with cognitive disturbances in schizophrenia patients, against selfmedication hypothesis; and in our animal model, that nicotine exposition during the adolescence combined to a moderate cortical and limbic hyperdopaminergia, is associated with persistent cognitive and social deficits, in favor of a gene environment interaction.

None Declared

Anorexia nervosa (AN) is a complex mental disorder mainly characterized by a voluntary food restriction and excessive physical activity resulting in dramatic weight loss. Changes in the brain-derived neurotropic factor (BDNF) have been reported in AN patients compared to controls. According to meta-analysis, functional variant rs6265 Val66Met of the BDNF gene has been found genetically associated to AN. We also reported an association of this functional variant and electrodermal response to images of thinness suggesting an association between rs6265 and a reward effect of weight loss in AN. In animal models, BDNF modulates negatively the central control of food intake and its injection in rodents induces weight loss and anorexia. Thus, besides its function on neuronal survival, synaptic plasticity and mood, BDNF was also reported to have a metabolic effect via both central nervous system and peripheral organs, which makes BDNF a good candidate for AN diagnosis biomarker.

Our study investigates the levels of expression of Bdnf, gene and protein, taking advantage of the mouse AN-like model by measuring Bdnf levels in specific brain areas and blood in food-restricted and refeed animals.

We used a mouse AN-like model combining a phase of chronic food restriction (50%) during 15 days followed by an ad libitum refeeding period of one week. Female mice have or not access to a running with wheel to create a similar metabolic environment that those patients suffering from AN during restriction and recovery once hospitalised. The Bdnf mRNA and protein levels were measured in samples of blood and brain regions (prefrontal cortex, hippocampus, hypothalamus, dorsal striatum, nucleus Accumbens, ventral tegmental area and amygdala) using quantitative PCR and ELISA methods in the different groups of mice (ad libitum, ad libitum with wheel, food restriction and food restriction with wheel). Statistical analysis will compare the measures for different samples by one-way or two-way ANOVAs depending the group of animals or brain regions and blood.

To date, no difference of the level of transcription for Bdnf was observed between the different groups of mice (ad libitum, ad libitum with wheel, food restriction and food restriction with wheel) in the prefrontal cortex, hippocampus and hypothalamus. We expect significant differences of Bdnf expression in the other brain regions of interest for the food restricted animals with or without the wheel compared to ad libitum animals. We expect also differences in the level of expression of Bdnf in fasted animals compared to the refeed animals.

The BDNF could represent a potential biomarker of AN for the diagnostic and the prognosis in the evolution to the remission when weight recover and thus will allow a better understanding of the aetiology of AN. This study is supported by Fédération pour la Recherche sur le Cerveau.

None Declared

Anorexia nervosa (AN) is a complex psychiatric disorder characterized by a persistant decrease in food intake leading to dramatic weight loss and energy deficit. The ghrelin system is a key regulator of appetite and food intake across species. LEAP-2, a recently discovered ghrelin antagonist, appears to be up-regulated in obesity and opposes to the orexigenic drive of ghrelin. The evolution of LEAP-2 levels could be an interesting insight to reflect the regulation of appetite in eating disorders such as anorexia nervosa (AN).

We tested this hypothesis and here provide the first study exploring the ghrelin and LEAP-2 regulation in long-term food restriction followed by refeeding in both mice and patients suffering from AN.

Using a translational strategy, we compared the regulation of ghrelin and LEAP-2 concentrations in blood during food restriction and after refeeding i/ in female mice exposed to a 14 days protocol combining quantitative food restriction and running wheel activity followed by 10 days of progressive refeeding; ii/ in an ongoing longitudinal study of patients with AN evaluated before and after refeeding (n=30) as well as 6 months after hospital discharge to evaluate if the weight gain was stable (n=7) or unstable (n=10). Plasma concentrations of ghrelin and LEAP-2 were measured with selective immunoassays.

Long-term food restriction in mice was associated with increased ghrelin (p<0.001) and decreased LEAP-2 concentrations (p=0.006) compared to ad libitum fed controls. Refeeding led to a decrease in ghrelin (p<0.01) and increase in LEAP-2 concentrations (p<0.01). Patients with AN displayed increased ghrelin levels (p<0.01) but also higher LEAP-2 concentrations on admission than after refeeding (p=0.04). LEAP-2 decreased with refeeding. On 17 patients re-evaluated 6 months after discharge, patients with unstable weight gain exhibited a greater decrease of LEAP-2 concentrations during refeeding compared to patient with stable weight gain (p=0.02). Decreasing LEAP-2 concentrations was able to predict a negative outcome (i.e. unstable weight gain) in 80% of the cases.

We provide evidence that the ghrelin/LEAP-2 system is not regulated according to the nutritional status in AN as it is in the case of a physiological adaptation to food restriction. Results from an ongoing longitudinal study exploring remission in AN suggest that the evolution of LEAP-2 concentrations during refeeding is opposed to data from preclinical model and could give new insights on the outcome of weight gain in AN.

None Declared

Gambling Disorder (GD) is considered a multifactorial behavioral addictive disorder, leading to severe psychological, social and economic consequences. Previous studies have investigated genetic mechanisms underlying GD. Growing literature showed a possible link between addiction-related disorders and neurotrophic factors (NTF), involved in synaptic plasticity and neuronal survival. Thus, the study of NTF genes emerged as promising targets in the field of GD.

To evaluate genetic implications of the NTF family in the pathophysiology GD. We hypothesized the involvement of some NTF genes polymorphisms in the onset and progression of GD as potential biological risk factors.

The sample was composed by 166 individuals with GD and 191 healthy controls. 36 Single nucleotide polymorphisms (SNPs) from NTF (NGF, NGFR, NTRK1, BDNF, NTRK2, NTF3, NTRK3, NTF4, CNTF and CNTFR) were selected and genotyped. Linkage disequilibrium and haplotype constructions were assessed, related to the presence of GD. Moreover, regulatory elements overlapping the identified SNPs variants associated with GD was also analyzed.

6 SNPs were potentially associated to GD after the comparisons of allele frequencies between groups. Single and multiple-marker analyses showed a strong association between both NTF3 and NTRK2 genes, and GD.

This study suggests the implication of NTF genes in the development of GD, being the altered cross-regulation of some NTF factors signalling pathways, a potential biological vulnerability factor in GD. Fundings and Acknowledgements: Ministerio de Ciencia, Innovación y Universidades (RTI2018-101837-B-100) Delegación del Gobierno para el Plan Nacional sobre Drogas (2017I067, 2019I47), Instituto Salud Carlos III (ISCIII) (PI17/01167, PI20/00132) and CIBERObn, an initiative of ISCIII.

No significant relationships.

Anorexia nervosa (AN) is a chronic psychiatric disorder resulting from abnormal eating habits with a high prevalence (0.5%). AN involves genetic and epigenetic factors supporting that AN is a metabo-psychiatric disorder. One candidate gene for AN, validated by meta-analyzes, is BDNF which encodes the brain-derived neurotrophic factor. BDNF negatively modulates the central control of food intake and its injection in rodents induces weight loss and anorexia. In humans, we observed an association of its functional variant Val66Met/rs6265 and electrodermal response to images of thinness suggesting an association between rs6265 and a reward effect of weight loss in AN.

This work study the impact of the functional polymorphism at risk rs6265, epigenetic variations in DNA methylation of BDNF gene and consequences on the concentrations of BDNF in AN patients.

DNA was isolated from 24 AN patients and 48 controls. DNA methylation was measured for sites spanning the BDNF gene using Infinium HumanMethylation450 BeadChip technology. The genotyping of rs6265 was performed by Taqman-SNP assay. The BDNF was dosaged by ELISA from plasmas.

We observe that several sites are significantly hypermethylated in AN patients compared to controls. AN patients show significantly higher BDNF levels than controls. Finally, this BDNF concentration is significantly higher in AN carrying the risk Met66 allele.

This work demonstrates the effects of genetic and epigenetic variations of BDNF, which could constitute relevant diagnostic biomarkers of AN, and their likely consequences in the pathophysiology of AN. This work was supported by the Nestlé Foundation.

No significant relationships.

Patients with anorexia nervosa (AN) show impaired decision-making ability, but it is still unclear if this is a trait marker, i.e. a stable endophenotype of AN, or a state parameter, i.e. being explained by present symptoms and associated comorbidity.

We aimed to determine whether decision-making impairment is an endophenotype of AN. We hypothesized that decision-making alteration would not respect the criteria of an endophenotype, and that these alterations would have a relationship with illness severity.

Ninety-one patients with acute AN (A-AN), 90 unaffected relatives (UR), 23 patients remitted from AN (R-AN) and 204 healthy controls (HC) underwent the Iowa Gambling Task (IGT) and psychometric assessments. Prospective Valence Learning model (PVL) was employed to distinguish the cognitive dimensions underlying the decision-making process. Performance at the IGT was compared between the four groups and then analysed according to clinical and psychometric variables.

Patients with A-AN scored worse than UR and HC at the IGT (p<.01). PVL-feedback sensitivity parameter was lower in patients with R-AN and A-AN than in the two other groups (p<.01) and PVL-loss aversion parameter was lower in A-AN than in UR and R-AN (p<.01). Decision-making style, in particular learning and loss aversion parameters, accounted for a significant part of variance of psychopathology in patients with AN (p<.01).

Impaired decision-making represents a state-associated, cognitive hallmark of AN. The aggravation of reward modulation along with illness progression may explain the persistence of symptoms despite their consequences on health. Reversal of decision-making impairment should not be limited by inherited vulnerability.

No significant relationships.

Depression is the most frequent psychiatric disorder. Antidepressant drug treatment response is modulated by genetic factors. Tianeptine molecule does not inhibit serotonin reuptake. Possible targets to this treatment are the serotonin transporter (5HTT/SLC6A4), the serotonin receptor 2A (5HT2A) and, the neuronal-specific enzyme that controls brain serotonin synthesis, the tryptophan hydroxylase-2 (TPH2).

We screened variants of candidate genes 5HTT/SLC6A4, 5HTR2A and THP2 in 3500 outpatients treated with tianeptine for a major depressive episode (MDE) to search for an association to a positive treatment response.

The goal of this work was to study the pharmacogenetic response to tianeptine treatment in MDE.

A total of 3500 outpatients were treated with tianeptine for a MDE. The criteria for a MDE were examined by the clinicians according to the DSM-IV diagnosis and the duration of each symptom was recorded during the inclusion and at 4 to 8 weeks of treatment. The Hospital Anxiety and Depression Scale (HAD) was evaluated at the two visits. DNA was extract from saliva sample and genotyping of single nucleotide polymorphisms (SNPs) was performed by Taqman assay.

All clinical and genotype data were collected for 1855 tianeptine-treated patients. The SNP rs6354 in SLC6A4 gene was significantly associated with response to tianeptine (p = 0.009; Odds ratio = 1.26; 95% confident interval = 1.06–1.50). Two SNPs in 5HT2A and one in TPH2 were also associated to treatment response (rs7322347 p = 0.03, rs7997012 p = 0.04 and rs7955501 p = 0.04).

We detected a pharmacogenetic association between serotonin genes and the response to tianeptine in major depressive episode.

Selon le modèle multifactoriel, la schizophrénie résulte de l’interaction de plusieurs gènes avec différents facteurs environnementaux, dont le cannabis qui peut jouer un rôle chez des sujets vulnérables [1]. Certains résultats issus de l’hypothèse neurodéveloppementale suggèrent une perturbation précoce du neuro-développement moins importante chez les consommateurs de cannabis et une sensibilité aux effets pharmacologiques du cannabis [3]. Les signes neurologiques mineurs (SNM) sont le reflet d’anomalies neuro-développementales précoces, ils sont fréquents dans les pathologies développementales à début précoce et remplissent les critères d’un marqueur endophénotypique [2]. Nous avons fait l’hypothèse que les patients consommateurs réguliers de cannabis avant le début des troubles constituent une entité phénotypique homogène avec une atteinte neuro-développementale précoce moins marquée.

Nous avons évalué prospectivement la symptomatologie négative, les déficits cognitifs et les SNM chez 61 patients schizophrènes suivis à l’hôpital Louis-Mourier (Colombes). L’évaluation était clinique avec la Positive And Negative Symptoms Scale, cognitive (fonctions exécutives, attentionnelles et mnésiques) et neurologique à l’aide de la Neurological Evaluation Scale. Les patients ayant consommé du cannabis avant le début des troubles étaient comparés aux autres patients (test t de Student pour les données quantitatives, test du Δ-2 ou test de Fischer pour les données qualitatives).

Les résultats principaux sont présentés dans le Tableau 1. On constate une perturbation plus importante des fonctions exécutives, attentionnelles et mnésiques dans le groupe des non-consommateurs de cannabis. Les SNM sont plus fréquents dans le sous-groupe non-consommateur de cannabis.

Tableau 1

Comparaison des caractéristiques cliniques, cognitives et neurologiques entre les patients ayant fumé régulièrement du cannabis avant le début des troubles et les autres patients.

Les patients consommateurs réguliers de cannabis avant le début des troubles représenteraient une entité phénotypique homogène. Le cannabis est un facteur environnemental pouvant intervenir tardivement sur le neurodéveloppement et pouvant jouer un rôle causal chez certains patients.

Autism (MIM#209850) is a complex neurodevelopmental affection that is largely genetic psychiatric disorder. Several genes have been found associated with autism but their expression levels and neuropathological effects remain unknown in autistic brain.

We compare the level of expression of autism candidate genes in post-mortem brain region samples between controls and patients. We studied Brodmann area (BA) 46 and the granule cells of the cerebellum lobule 6, for which neuropathological findings and functional abnormalities have been reported in autism.

Different levels of transcription for SLC25A12/AGC1, EN2 and Nr-CAM genes are observed in the cortex and granule cells. Difference of expression are observed between patients and controls. We focused on SLC25A12 for which polymorphisms have been associated to autism in various studies. SLC25A12 encodes the mitochondrial aspartate/glutamate carrier and its function is requested to produce energy in neurons. By hybridation in situ, we analysed the expression pattern of SLC25A12 in human development and we studied the effects of SLC25A12 over-expression on mouse embryonic cortical neurons.

Convergent evidence suggest that level of expression of candidate genes may be involved in autism pathophysiology by modifying neuronal networks and molecular plasticity in specific brain subregions at both pre- and postnatal stages.

Autism Tissue Program and Fédération pour la Recherche sur le Cerveau

The risk of suicide attempt and death by suicide during an antidepressant drug treatment for a major depression episode (MDE) remains unclear. Furthermore, this risk can be influenced by genetic factors, especially those involved in the serotonine pathway.

We genotyped the variants of 6 candidate genes (including 5HTT/SLC6A4, TPH2, 5HTR1A, 1B & 2A) in 4000 outpatients treated with tianeptine for a MDE to search for an association with a suicide risk and the treatment response.

The goal of this study was to evaluate the pharmacogenetic role of the serotonin pathway in suicide risk during an antidepressant treatment.

A total of 4000 outpatients for a MDE were treated with tianeptine. The criteria for a MDE were examined by the clinicians according to the DSM-IV diagnosis and the duration of each symptom was recorded during the inclusion and at 4 to 8 weeks of treatment. The Hospital Anxiety and Depression Scale (HAD) was evaluated at the two visits. Suicide attempts, number and suicide idea of MADRS, before and during treatment, were recorded. DNA was extract from saliva sample and genotyping of 22 single nucleotide polymorphisms (SNPs) was performed by Taqman assay.

All clinical and genotyping data were collected for 3771 tianeptine-treated outpatients. A total of 450 subjects have reported suicide attemps before treatment and 24 during treatment.

The results of pharmacogenetic association for suicide risk during an antidepressant treatment will be presented during the workshop.

The most frequent psychiatric disorder is depression. Half of patients treated by tianeptine present a positive response. In Fact, tianeptine response, as well as, other antidepressants is modulated by genetic factors. This molecule was shown acting on glutamate and serotonin pathways.

We analyzed polymorphisms of 12 candidate genes involved in glutamate and serotonin pathways in 3536 outpatients treated with tianeptine for a major depressive episode (MDE) to find an association to a positive treatment response.

The aim of this study was to identify the pharmacogenetic response to tianeptine treatment in MDE.

A total of 3536 outpatients with a MDE were treated with tianeptine. The criteria for a MDE were characterized by clinicians according to the DSM-IV diagnosis and, the duration and intensity of each symptom was recorded during the inclusion at 4 to 8 weeks of treatment. The Hospital Anxiety and Depression Scale (HAD) was evaluated at the two visits. DNA was extract from saliva samples and genotyping of 49 single nucleotide polymorphisms (SNPs) in 12 genes was performed by Taqman assay.

All clinical and genotype data were collected for 3536 tianeptine-treated patients. A total of 7 SNPs, located in 4 different genes, were found associated to a positive treatment response, corresponding to a reduction of 50% of the HAD. This was confirmed in the Caucasians subgroup of 3040 patients.

We detected pharmacogenetic associations between glutamate and serotonin genes and the response to tianeptine in major depressive episode.

Identifying schizophrenia vulnerability factors requires testing specific hypotheses concerning etiopathogenic models, candidate genes, and environmental factors involved. We hypothesized that patients developing schizophrenia after using regularly cannabis represent an homogenous phenotypic entity which could result from a specific interaction between genetic polymorphisms (SNPs) of CNR1, CNR2 and COMT genes and cannabis use (an environmental factor intervening lately in the neurodevelopment).

The main objective was to confirm that patients using cannabis before the onset of schizophrenia had fewer markers of early-neurodevelopmental impairment. The secondary objective was to identify a gene environment interaction in this subset of patients.

Patients were prospectively included and underwent a standardized interview and an evaluation of cognitive functions and neurological soft signs. Blood and saliva samples were collected. For the case-case genetic study, those patients were pooled with patients included in a cannabisdependent schizophrenic-patients cohort. Eight CNR1, CNR2 and COMT SNPs were analyzed.

We included 61 patients in the clinical part. Regular cannabis users before the onset of the disorder had fewer neurological soft signs, less cognitive impairment and a different clinical symptomatology. For the genetic part 26 patients were added to the 57 patients from the clinical part (4 non Caucasian patients excluded). Results were in favor of a gene environment interaction for rs1049353 of CNR1.

We found that regular cannabis users before the onset of schizophrenia had less neurodevelopmental impairment. In this subset of patients, cannabis could play a causal role in a gene environment interaction leading to schizophrenia.

Deregulations of peripheral control of food intake in anorexia nervosa (AN), hunger signals (Ghrelin & obestatin) and satiety signals (leptin & insulin), have been reported. However, these differences could be a consequence or a risk factor of the disease. The genetic approach seems to be a good strategy to analyse this issue. Indeed, AN has a high heritability of ~50-80%.

The aims of this study are to confirm deregulation of hunger and satiety signals in AN and to analyze these results according to genetic polymorphisms. Furthermore, we search for an endophenotype by the screening of the AN mothers.

This work recruited 3 populations: 100 anorexic patients recruited at CMME (Sainte-Anne Hospital, Paris), their mothers, and 200 control women, matched with the patients or relatives for age. All subjects were assessed during a morning day at CMME. They arrived at 8:30am, fasting since the day before, to take a blood sample, in order to carry out genetic and physiologic analyses. Then, they were analysed for Body Mass Index and eating behaviors, including a psychiatric interview (DSM-IV-TR criteria) and self-questionnaires. Dosage of ghrelin, obestatin, and leptin were done by EIA or RIA. Single nucleotide polymorphisms were genotyped by Taqman assay.

We have confirmed that peripheral control mechanism of food intake is deregulated in AN patients and we observe difference also between mothers and controls.

Leptin dosage might be an endophenotype for anorexia nervosa.

Anorexia nervosa (AN) is the most severe in terms of morbidity psychiatric illness with the highest mortality rate increased by 23 fold. Treatments are limited effectiveness. AN has a strong genetic component with heritability at 70% but despite ∼ 200 studies no major gene was identified. Epigenetics, such as DNA methylation, is another component of heritability that could explain the high heritability. Methylation is poorly studied in AN from small samples, and is focused on few candidate genes among publications. Under publication, a first genome-wide methylation study investigated 10 restrictive type AN patients, 19 binging/purging type of AN patients and 15 normal eaters using DNAs from whole blood (Booij, 2015). Of the 480K CpG sites that can be methylated of Infinium Human Methylation450 BeadChip Kit, authors focused on 24,000 sites located close to genes and they identified candidate genes with a different profile of methylation between AN and controls.

Our work is to replicate the results of Booji and also to investigate the AN remitters.

Our goal is to identify epigenetic signatures of the AN disorder and the prognostic of remission.

Twenty-four AN patients, 24 AN remitters will be compared to 48 healthy control women for methylation using the Infinium Human Methylation450.

As Booji et al., we will compare methylation for 24,000 sites located close to genes for 24 AN, 24 remitters and 48 controls.

We expected to replicate the published results of Booji and to identify genes with a methylation signature specific of the AN remission.

The authors have not supplied their declaration of competing interest.

Anorexia nervosa (AN) is a severe psychiatric disorder. The epigenetic regulations are strongly suggested in AN. We and other groups have performed a whole-genome methylation study (methylome) in AN. We found that the differentially methylated CpG sites are located around genes involved in biological processes in link with embryonic morphogenesis, brain development and its plasticity, in particular adhesion and axon guidance. Here, we study an independent group of 40 AN patients. Furthermore, we have done a follow-up during more than one year, to compare the methylation profiles in subjects that evolve to the remission.

Our work is to replicate the methylome study in an independent AN cohort and to characterize profiles of methylation at two times for the same subjects to compare the AN patients that convert to remitters.

Our goal is to identify diagnostic and prognostic epigenetic signatures for AN.

Of the 40 AN patients, 18 evolved to remission. Furthermore, the blood samples of the subjects from the 2 times will be investigated, like this, each subject is its own control. Methylation of DNA is measured by using the Infinium HumanMethylation450 BeadChip technology.

Comparisons of AN to controls showed similar profiles of methylation involving the same biological processes as previously identified. We are comparing now the difference of methylation between the 18 remitters and the 18 actual AN, taking into account of the two times of samples.

We expect to characterize specific methylation signature of the prognostic of the AN remission.

The authors have not supplied their declaration of competing interest.