Glucose, Brain and Microbiota

Overview

The accumulation of iron is known to affect the functions of the liver, adipose tissue and muscle. The brain is a well-known place of iron deposition, which is associated with cognitive parameters of subjects with obesity. The hypothesis is that certain parameters related to glucose metabolism (glycemic variability, the circulating concentration of AGE receptor agonists, pentosidine and HbA1c) are associated with cognitive function, brain iron content and gut microbiota composition in subjects with obesity. The study includes both a cross-sectional (comparison of subjects with and without obesity) and a longitudinal design (evaluation one year after weight loss induced by bariatric surgery or by diet in patient with obesity) to evaluate the associations between continuous glucose monitoring, brain iron content (by magnetic resonance), cognitive function (by means of cognitive tests), physical activity (measured by activity and sleep tracker device) and the composition of the microbiota, evaluated by metagenomics.

Subjects and methods: A. Cross-sectional study: Patients with obesity previously scheduled at the Service of Endocrinology, Diabetes and Nutrition (UDEN) of the Hospital "Dr. Josep Trueta" of Girona (Spain) will be recruited and studied. Subjects without obesity will also be recruited through a public announcement. A blood glucose sensor will be implanted for ten days, as well as an activity and sleep tracker device to record physical activity during this period of time. Interstitial subcutaneous glucose concentrations will be monitored on an outpatient basis for a period of time of 10 consecutive days using a glucose sensor validated by the FDA (Dexcom G6 ®). The sensor will be implanted in on day 0 and will retire on day 10 midmorning. Glucose records will preferably be evaluated on days 2 to 9 to avoid the bias caused by the insertion and removal of the sensor, which prevents a sufficient stabilization of the monitoring system. The characteristic glycemic pattern of each patient will be calculated on average from the profiles obtained on days 2 to 9. At the end of the week an magnetic resonance imaging will be done to evaluate the iron content in the brain and parameters of "Diffusion Tensor Imaging" in different brain territories. Cognitive tests will be carried out and feces will be collected for the study of the microbiota. The project will be carried out in subjects with obesity (20 men, 20 premenopausal women and 20 women postmenopausal, BMI \> = 30kg/m2) and subjects without obesity, similar in age, sex and menopausal status (20 men, 20 premenopausal women and 20 postmenopausal women, BMI \<30kg/m2). B. Longitudinal study: After one year of follow-up, in which, subjects with obesity will undergo conventional treatment (hypocaloric diet and physical activity advise) or bariatric surgery for weight loss, a second visit will be carried out. For comparison, the same protocol of the cross-sectional study will be done again. See information above. Data collection of subjects of cross-sectional and longitudinal studies: * Subsidiary data: Age, sex and birth date. * Clinical variables: Weight, height, body mass index, waist and hip perimeters, waist-to-hip ratio, blood pressure (systolic and diastolic), fat mass and fat free-mass (bioelectric impedance and DEXA), smoking status, alcohol intake, registry of usual medicines and registry of antecedent relatives with obesity, diabetes and comorbidities. * Laboratory variables: 15cc of blood will be extracted from fasted subjects to determine the following variables using the usual routine techniques of the clinical laboratory (hemogram, glucose, bilirubin, aspartate aminotransferase (AST/GOT), alanine aminotransferase (ALT/GPT), gamma-glutamyl transpeptidase (GGT), urea, creatinine, uric acid, total proteins, albumin, total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, glycated haemoglobin (HbA1c), ferritin, soluble transferrin receptor, ultrasensitive C reactive protein, erythrocyte sedimentation rate, lipopolysaccharide binding protein, free thyroxine (free T4), thyroid stimulating hormone (TSH) and baseline cortisol). An additional 15cc of blood (plasma-EDTA) will be extracted for further analyses. * Stool samples collection: A stool sample will be provided from each patient. The sample should be collected at home or in the hospital, sent to the laboratory within 4 hours from the collection, fragmented and stored at -80ºC. * Magnetic Resonance Imaging: All MRI examinations will be performed on a 1.5-T scanner (Ingenia ®; Philips Medical Systems). First, fluid-attenuated inversion recovery (FLAIR) sequence will be used to exclude subjects with preexisting brain lesions. Brain iron load will be assessed by means of R2\* values. T2\* relaxation data will be acquired with a multi-echo gradient-echo sequence with 10 equally spaced echoes (first echo=4.6ms; inter echo spacing=4.6ms; repetition time=1300ms). T2\* will be calculated by fitting the single exponential terms to the signal decay curves of the respective multi-echo data.R2\* values will be calculated as R2\*=1/T2\* and expressed as Hz. In addition, R2\* values will be converted to μmol Fe/g units as previously validated on phantom tests. Brain iron images from control subjects will be normalized to a standard space using a template image for this purpose (EPI MNI template). Subsequently, all normalized images will be averaged for the determination of normal iron content. Normal values (mean and SD) will be also calculated for anatomical regions of interest using different atlas masks, addressing possible differences between gender and age. The brain iron comparison between control and obese subjects will be performed using voxel-based analysis. Obese-subjects images will be normalized to a standard space. The normalized image will be compared to normal population using t-test analysis with age and sex as co-variables. As result, a parametric map will show individual differences in iron deposition. Based on previous observational studies showing increased brain iron load at some specific regions and the evidence suggesting hippocampal and hypothalamic changes in association with obesity and insulin resistance, the statistical and image analyses will be focused on iron differences at the caudate, lenticular, thalamus, hypothalamus, hippocampus, and amygdala. * Neuropsychological examination: General cognitive functioning will be measured using the Vocabulary and Similarities subtests of the Wechsler Adult Intelligence Scale-III (WAIS-III); attention and working memory by the Forward and Backward Digit Span subtest of the WAIS-II; memory using the California Verbal Learning Test II; executive functions by the Trail Making Test, the Color-Word Stroop Test and the Verbal Fluency; mood using the Patient Health Questionnaire-9 and impulsive behaviors using the Iowa Gambling Task. * Microbiota composition: the microbiota composition will be analyzed according to a previous described protocol. 16s rRNA qPCR and LPS-binding protein in blood samples will be used for detection of bacterial translocation. The information will remain registered in a notebook and will be computerized in the database of the study. Statistical methods: Sample size: There are no previous data showing expected differences for sample size estimation regarding glucose variability, physical activity, composition of gut microbiota and cognitive function. In a previous study, differences in brain iron content were observed in 20 obese vs. 20 nonobese subjects. Thus, the proposed sample size is at least 20 individuals per group, with balanced age and gender (pre- and postmenopausal women) representation. Statistical analyses: Firstly, normal distribution and homogeneity of variances will be tested. To determine differences between study groups, it will be used χ2 for categorical variables, unpaired Student's t-test in normal quantitative and Mann-Whitney U test for non-normal quantitative variables. Nonparametric Spearman analysis will be used to determine the correlation between quantitative variables. The same tests will also be used to study differences before and after follow-up. The significant associations, whether positive or negative, will be explored more-in-depth (simple and multivariate linear regression analyses). The microbiota composition will be analyzed and compared using HeatMaps, Principal Component Analysis (PCA) and PLSDA. For multivariate statistics (PLSDA and hierarchical clustering), variables comprising morphological tissue characteristics, gut microbiota and functional test will be log transformed, filtered using interquartile range estimate and scaled using auto-scaling calculation (mean-centered and divided by the standard deviation of each variable) by using the Metaboanalyst ® platform, the R ® package ropls and MATLAB ® scripts. Alpha and beta biodiversity will be compared according to obesity, insulin resistance and iron status. It will also be used SPSS ® statistical software and Minitab ®.