Background. Anorexia nervosa (AN) still carries the highest fatality rate of any psychiatric disease, and less than half of the patients recover, completely refractory to any treatment. The etiology remains unknown and evidence for treatment is lacking. The intestinal microbiota and its microbiome provide humans with additional gene products which may be regarded as an organ, which contributes to multiple host metabolic pathways. Recent advances in microbial DNA sequencing technologies have resulted in metagenomic DNA analysis of whole ecosystems such as the human gut. Altered intestinal microbiota has been related to obesity and insulin resistance. Hypothetically, the intestinal microbiota could play a role in the generation and/or maintenance of the emaciation in AN.

Aim. The aim of the present study is to investigate whether gut microbiota is altered in patients with AN.

Subjects and methods. A cross sectional study of the gut microbiome profiles in 75 clinical, psychometric and biochemical well characterized treatment seeking females with AN. The microbiome profiles are compared with 75 age- matched healthy Danish control subjects.

Perspectives. Clarifying whether the intestinal flora is implicated in the susceptibility to or maintenance of AN may provide the basis for development of new highly required treatments.

BACKGROUND Anorexia nervosa (AN) is a complex disorder characterized by disturbed body image, ego-syntonic neglect, ambivalence, self-starvation, loss of body weight, obsessive thoughts of food, ritualistic patterns of food intake, elevated physical activity, depression, anxiety and emotional rigidity. The evidence base for the treatment is very limited, considering the extent to which this disorder erodes quality of life and still carries the highest fatality rate of any psychiatric diseases. In fact, there is no evidence that the prognosis has improved throughout the 20th century.

AN remains a syndrome, i.e. collections of symptoms, as it is not defined by etiology. Twin studies have consistently demonstrated that AN is strongly familial predominantly due to genetic factors. However, neither candidate gene nor genome-wide studies have identified truly validated genes for AN. AN is associated with multiple, profound endocrine disturbances. However, so far the extensive research has not led to any breakthrough discovery. There is a need for new thinking.

Colonization of the human gut begins at birth with bacteria from the mother's vagina and labor-released maternal feces. In adults the human digestive tract contains complex assemblies of microorganisms which actually outnumber the number of the host cells 10:1. This intestinal microbiota and its microbiome provide humans with additional gene products, which may be regarded as an organ, which contributes to multiple host metabolic pathways. Thus, a whole range of substances are produced which are not yet identified. Potentially some of these substances may affect functions in the brain, such as appetite and emotion regulations. Recent advances in microbial DNA sequencing technologies have resulted in metagenomics DNA analysis of whole ecosystems such as the human gut. An important step forward in investigating the bacterial composition of the human microbiome was taken with the Metagenome of the Human Intestinal Tract consortium reporting the first complete microbial gene atlas of the human distal gut, using deep next generation metagenome sequencing.

AIM To investigate whether gut microbiota is altered in patients with AN. If confirmed, this potentially may play a role in the pathogenesis, promoting susceptibility to development and/or maintenance of AN.

HYPOTHESIS Patients with AN harbor an altered gut microbiota composition and functional potential as determined by quantitative metagenomic analyses of microbial DNA isolated from stool samples and sequenced applying a combination of deep and untargeted shotgun sequencing.

STUDY DESIGN A cross sectional multicenter study where the gut microbiome profiles in women with AN are compared with age matched healthy, normal weight women.

SUBJECTS The patients are characterized according to the national guidelines. The following data will be extracted from the medical files and included in the descriptive and multivariate analyzes.

Clinical variables: Age, disease duration, BMI, %weight change during the last 4 weeks, in/out patient status, medication, somatic and psychiatric comorbidity, smoking status.

Psychometric variables: Eating Disorder Inventory (EDI)-3.

Control subjects: The control subjects will be matched as described, and allocated from siblings of patients, local health workers, or if necessary, from announcement at an official Danish website for participation in clinical trials. The age matched women fill in a 17 items health related questionnaire. Any chronic disease, medication or a history of psychiatric illness results in their exclusion from participation. They are, however, allowed to use oral contraceptives.

FECAL SAMPLING Fecal samples are collected in order to characterize the composition of the gut microbiome. A sample of about 5 g is collected with the help of a collection tray. Participants are instructed in immediately storing the sample in their freezer at -18°C. Transportation is done at cold temperatures in an isolated box containing frozen elements, and handed over to study personnel maximal 24 hours after delivery where it is stored at -80 C until DNA extraction.

MORNING FASTING URINE 5 ml morning fasting urine is collected.

ANALYZING FOR DIFFERENCES IN GUT MICROBIOTA COMPOSITION The feces specimens are processed and analyzed as follows: Bacterial DNA extraction, shotgun illumine DNA sequencing, microbial gene analyses, taxonomy analyses including entero types, known species and unknown meta-species (=quantitative metagenomics) as well as functional gene annotation. The DNA purification and the shotgun sequencing is done at Unité MetaGenoPolis - Centre de recherche de Jouy-en-Josas, MGP-INRA, JOUY en JOSAS, FRANCE. Quantitative metagenomics as outlined above is performed by the unit of Bioinformatics and Computational Metagenomics at CBMR applying established state-of-the art methodologies.

BIOCHEMICAL VARIABLES Basic parameters: The following analyses are performed as an integral part of diagnosis and treatment in specialist units, and will be extracted from the medical files: Na, K, creatinine, carbamide, bicarbonate, Ca-ion, phosphate, Hemoglobin, C-reactive protein, leukocytes, mean cell volume, mean cell hemoglobin concentrate, alanine aminotransferase, alkaline phosphatase, bilirubin, pancreatic amylase, Mg, Zn, thyroxine, triiodothyronine, thyrotropin, cobalamine, folate, follicle stimulating hormone, luteinizing hormone, prolactin, estradiol, and vitamin D. Values will be adjusted and expressed as ratio relative to the normal value on the basis of average standard values for each centers.

SAMPLES

All samples are stored at -80 C:

• Faeces 4 x 5 g.
• Serum 20 x 500 microliter serum.
• Plasma 20 x 500 microliter plasm.
• Buffy coat x 10.
• Morning urine 2 ml.

Biochemical analyses: From the sampled serum/plasma, following analyses are performed: Insulin, Insulin-like growth factor-I (IGF-I), IGF-1 binding protein-3, leptin, and adiponectin.

BIOBANK The fecal samples and the above mentioned materials will be stored in a biobank in up to 30 years. If a microbiome genotype for AN is disclosed, further genomic analyses, mass spectroscopy for metabolites and maybe inoculation studies will be performed (followed by new applications to the Ethical Committee).

STATISTICS Power calculation. There is no data available regarding microbiota in AN. The number of participants is based on ongoing studies of gut microbiome in pregnant women and patients with multiple sclerosis where a population size of 50 was estimated to be appropriate for providing a power greater than 80% and a significance of 5% for observing differences in the present quantitative metagenomics approach.

Multivariate analyze. Statistical methods will build upon previously developed methods and multivariate statistical approaches applied within the Metagenomics of the Human Intestinal Tract (MetaHIT) consortium.

ETHICS The study is approved and registered at The Regional Scientific Ethical Committee for Southern Denmark (file no 42053 S-20140040).

The participants will be asked whether they are willing to donate the material (feces, blood, urine) to a biobank for later analysis, and will only be used in related future studies, for which approval will first be obtained from the Scientific Ethical Committee. The biobank samples are stored up to 30 years. Clinical, psychometric and biochemical variables mentioned above are extracted from medical files. Patients from the three participating centres will receive an invitation to an information meeting with one of the health workers, who are one of the local investigators. For subjects below age of 18 years (15 - 18 y) one or both parents must participate and must provide written informed consent. Subjects above age 18 years are allowed to bring a friend or family member as an observer. It will be emphasized that it is voluntary to participate. During the information meeting an official document published by The National Committee on Health Research Ethics 2012 is handed.

The invitation of siblings is motivated by the fact that future studies of biobank material can prove to be of particular interest, if there is a high level of genetic relatedness between the test patients and the control subjects. However, the patient inclusion in the present project will not depend on whether there are siblings who wish to participate.

Recruitment of patients down to the age of 15 are motivated by the fact that the highest prevalence of AN is in this age group. The older the patient, the greater is the likelihood for psychiatric co-morbidity. It therefore will be important for the validity of the results that the age spectrum also covers 15 - 18 years of age.

HANDLING AND ARCHIVING DATA The current protocol are approved by The Danish Data Protection Agency (case files 1429534, 2008-58-035) and all the documents and materials related to the clinical study will be stored according to the Act on Processing of Personal Data. All sensitive information of personal, physical or biological origin related to the individual participants will be treated confidentially according to the Danish legislation.

FINANCE This is a non-profit study, completely independent from commercial interests. There will be applied for public and private grants to finance this study.

PERSPECTIVES If the present study identifies an altered gut flora in AN, it provides the basis for several important new studies of gen products in faces and blood in AN, and also inoculation experiments with gnotobiotic mice. Clarifying whether the intestinal flora is implicated in the susceptibility to or maintenance of AN may provide the basis for development of new highly required treatments of AN and potentially also obesity.

Inclusion:

• Women.
• Referred to specialized eating disorder center.
• Fulfill criteria in Diagnostic and Statistical Manual of Mental Disorders - 5.

Exclusion:

• Age <15 years.
• Known ongoing abuse of laxatives.
• Antibiotic treatment within the last three months.
• Inflammatory bowel disease, colorectal cancer or other bowel disease.

Patients referred to specialist unit.