The Gut Hormones PYY3-36 and GLP-17-36 amide Reduce Food Intake and Modulate Brain Activity in Appetite Centers in Humans

Summary Obesity is a major public health issue worldwide. Understanding how the brain controls appetite offers promising inroads toward new therapies for obesity. Peptide YY (PYY) and glucagon-like peptide 1 (GLP-1) are coreleased postprandially and reduce appetite and inhibit food intake when administered to humans. However, the effects of GLP-1 and the ways in which PYY and GLP-1 act together to modulate brain activity in humans are unknown. Here, we have used functional MRI to determine these effects in healthy, normal-weight human subjects and compared them to those seen physiologically following a meal. We provide a demonstration that the combined administration of PYY3-36 and GLP-17-36 amide to fasted human subjects leads to similar reductions in subsequent energy intake and brain activity, as observed physiologically following feeding.


Figure S1
Please see Excel file for Table S1.

Peptides
The identity and purity of each peptide was confirmed by matrix-assisted laser desorption

Plasma Hormone Assays
The antibody used in the PYY radioimmunoassay fully cross reacts with only the 3-36 form of human PYY. There is no cross reactivity with the 1-36 form of the hormone. The Millipore PYY 3-36 assay utilizes 125 I labelled PYY and a PYY 3-36 antiserum to determine the plasma level of PYY 3-36 by the double antibody/PEG technique. The detection limit of the assay was 5 pmol/l, with an intra-assay coefficient of variation of 6.4-11.0 %. All samples were measured in one assay to avoid inter-assay variation.
In the GLP-1 ELISA, the monoclonal antibody immobilized in the wells of the microwell plate binds specifically to active forms of GLP-1 only (GLP-1 7-36 amide and GLP-1 7-37 ), with no cross reactivity with any other form of the hormone. Bound GLP-1 is conjugated to an anti-GLP-1 alkaline phosphatise, which produces the fluorescent product umbelliferone when methyl umbelliferyl phosphate is added. The amount of fluorescence generated is proportional to the amount of GLP-1 in the sample. The limit of detection was 2 pmol/l, with an intra-assay variation of 6-9 %. All samples were measured in one assay to avoid inter-assay variation.  (Jack et al., 2008)) were acquired with whole-brain coverage (208 slices) for each participant to facilitate fMRI image co-registration, and ROI definition (TR = 3000 ms, TE = 3.66 ms, flip angle = 9º, voxel size = 1 mm³).

fMRI picture processing task
During the fMRI picture processing task, 75 images were shown, divided into three classes of 25 exemplars: 25 high-calorie foods, 25 low-calorie foods and 25 non-food items (Beaver et al., 2006). There were two sets of 75 images, alternated between each scanning session. Images from each category were presented in counterbalanced order across participants and sessions. The images appeared for 5 s, interspersed with periods of rest (fixation on a blue octagon), with each block lasting 25 seconds. Subjects were instructed to press a button in response to the images shown, and asked to alter the length of time they pressed the button depending on how pleasant they found each image.

fMRI analysis
fMRI data were preprocessed with FSL software (www.fmrib.ox.ac.uk/fsl/) in order to correct for motion, to register the echo planar functional images to the high-resolution anatomical (T1-weighted) scan of each individual, and to overlay images on a standardized atlas (MNI) in order to allow for comparisons across individuals. Data were high-pass filtered and spatially smoothed (5mm Full Width Half Maximum Gaussian kernel) to allow for gyral variability across subjects and to improve signal-to-noise ratio at the intra-subject level.
Individual sessions with greater than 3.5mm absolute head motion or severe stimuluscorrelated motion (as exhibited by marked rim artifacts in intra-subject statistic maps) were excluded from further analysis. This led to a total of 15 usable subjects with complete data from all infusion groups.
We estimated the difference in regional mean BOLD signal intensity between periods of subject exposure to palatable food images in relation to balanced exposure to periods of non-food images, at each of 6 pre-specified ROIs. All ROIs were defined as the conjunction between the full group main effect of task (food vs. non-food, using parametric testing at the level of spatially contiguous supra-thresholded clusters, while controlling the family-wise probability of type 1 error at p<0.05, corrected) and the relevant atlas region. We confirmed the accuracy of registration of ROIs on the BOLD functional data by visual inspection of their overlay in relation to the functional images i.e. in subject space. These estimates were then averaged over right and left homologous regions of amygdala, caudate, insula, nucleus accumbens, OFC, and putamen. For each infusion, we tested the null hypothesis that the within-subject difference between fasted saline, fed saline, PYY 3-36 infusion, GLP-1 7-36 amide infusion or combined PYY 3-36 + GLP-1 7-36 amide infusion in regional BOLD response to food images vs. non-food images was zero, i.e. [the gut hormone infusion or fed saline BOLD response to palatable food exposures] -[fasted saline BOLD response to palatable food exposures] = ΔBOLD = 0. The statistical model for the primary study objectives was a repeated measures analysis of difference of means, between control (fasted saline) and each infusion session's BOLD data. A paired t-test model was also used to test intra-regional difference effects of the three gut hormone infusion states and the fed saline state against fasted saline.