Journal of Clinical Medicine Article Hunger and Satiety Peptides: Is There a Pattern to Classify Patients with Prader-Willi Syndrome? Marta Bueno 1 , Ester Boixadera-Planas 2 , Laura Blanco-Hinojo 3,4 , Susanna Esteba-Castillo 5,6 , Olga Giménez-Palop 7 , David Torrents-Rodas 8 , Jesús Pujol 3,4 , Raquel Corripio 9 , Joan Deus 3,10 and Assumpta Caixàs 7, * 1 Endocrinology and Nutrition Department, University Hospital Arnau de Vilanova, Obesity, Diabetes and Metabolism (ODIM) Research Group, IRBLleida, University of Lleida, 25198 Lleida, Spain; mbuenodiez@gmail.com 2 Servei d’Estadística Aplicada, Universitat Autònoma de Barcelona, 08193 Cerdanyola, Spain; ester.boixadera@uab.cat 3 MRI Research Unit, Department of Radiology, Hospital del Mar, 08003 Barcelona, Spain; laura.blanco02@gmail.com (L.B.-H.); 21404jpn@comb.cat (J.P.); joan.deus@uab.cat (J.D.) 4 Centro Investigación Biomédica en Red de Salud Mental, CIBERSAM G21, 08003 Barcelona, Spain 5 Specialized Service in Mental Health and Intellectual Disability, Institute of Health Assistance (IAS), Parc Hospitalari Martí i Julià, 17190 Girona, Spain; susanna.esteba@ias.cat 6 Neurodevelopment Group [Girona Biomedical Research Institute]-IDIBGI, Institute of Health Assistance (IAS), Parc Hospitalari Martí i Julià, 17190 Girona, Spain  7 Endocrinology and Nutrition Department, Parc Taulí Hospital Universitari, Institut d’Investigació i Innovació  Parc Taulí I3PT, Medicine Department, Universitat Autònoma de Barcelona, 08208 Sabadell, Spain; Citation: Bueno, M.; ogimenez@tauli.cat 8 Biological Psychology Department, Philipps-Universität Marburg, 35037 Marburg, Germany; Boixadera-Planas, E.; Blanco-Hinojo, L.; Esteba-Castillo, S.; Giménez-Palop, torrents@staff.uni-marburg.de 9 Pediatric Endocrine Department, Parc Taulí Hospital Universitari, Institut d’Investigació i Innovació Parc O.; Torrents-Rodas, D.; Pujol, J.; Taulí I3PT, Medicine Department, Universitat Autònoma de Barcelona, 08208 Sabadell, Spain; Corripio, R.; Deus, J.; Caixàs, A. rcorripio@tauli.cat Hunger and Satiety Peptides: Is There 10 Department of Clinical and Health Psychology, Universitat Autònoma de Barcelona, 08193 Cerdanyola, Spain a Pattern to Classify Patients with * Correspondence: acaixas@tauli.cat; Tel.: +34-93-7458-412; Fax: +34-93-7160-646 Prader-Willi Syndrome? J. Clin. Med. 2021, 10, 5170. https://doi.org/ Abstract: Hyperphagia is one of the main problems of patients with Prader-Willi syndrome (PWS) 10.3390/jcm10215170 to cope with everyday life. The underlying mechanisms are not yet well understood. Gut-brain hormones are an interrelated network that may be at least partially involved. We aimed to study the Academic Editors: Kei Nakajima and Assumpta Caixàs hormonal profile of PWS patients in comparison with obese and healthy controls. Thirty adult PWS patients (15 men; age 27.5 ± 8.02 years; BMI 32.4 ± 8.14 kg/m2 ), 30 obese and 30 healthy controls were Received: 3 October 2021 studied before and after eating a hypercaloric liquid diet. Plasma brain-derived neurotrophic factor Accepted: 2 November 2021 (BDNF), leptin, total and active ghrelin, peptide YY (PYY), pancreatic polypeptide (PP), Glucagon-like Published: 4 November 2021 peptide-1 (GLP-1), glucose-dependent insulinotropic polypeptide (GIP) and amylin were determined at times 00 , 300 , 600 and 1200 . Cluster analysis was used. When considering all peptides together, Publisher’s Note: MDPI stays neutral two clusters were established according to fasting hormonal standardized concentrations. Cluster with regard to jurisdictional claims in 1 encompassed most of obese (25/30) and healthy controls (28/30). By contrast, the majority of published maps and institutional affil- patients with PWS were located in Cluster 2 (23/27) and presented a similar fasting profile with iations. hyperghrelinemia, high levels of leptin, PYY, GIP and GLP-1, compared to Cluster 1; that may reflect a dysfunction of these hunger/satiety hormones. When peptide behavior over the time was considered, PP concentrations were not sustained postprandially from 60 min onwards in Cluster 2. BDNF and amylin did not help to differentiate the two clusters. Thus, cluster analysis could be a good tool Copyright: © 2021 by the authors. to distinguish and characterize the differences in hormone responses between PWS and obese or Licensee MDPI, Basel, Switzerland. healthy controls. This article is an open access article distributed under the terms and Keywords: Prader-Willi syndrome; obesity; hunger; satiety; BDNF; PYY; ghrelin; clusters conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). J. Clin. Med. 2021, 10, 5170. https://doi.org/10.3390/jcm10215170 https://www.mdpi.com/journal/jcm J. Clin. Med. 2021, 10, 5170 2 of 14 1. Introduction Prader-Willi syndrome (PWS) is the first genetic syndromic cause of obesity and is caused by the lack of expression of the genes contained in the 15q11-q13 region of chromo- some 15 of paternal origin. The syndrome includes features such as neonatal hypotonia, intellectual disability, growth hormone deficiency or hypogonadism. Undoubtedly, its main characteristic is hyperphagia, which leads to the development of morbid obesity, the main cause of morbidity and mortality in these patients [1]. Knowing the underlying causes behind hyperphagia would enable the development of new therapeutic targets. However, although many studies have tried to delve into them, they are not yet well known. In this sense, there are several works that have tried to find alterations in different hormones involved in the regulation of appetite. Thus, hyperghrelinemia is a constant finding in these patients [2–5]. They have high total ghrelin concentrations with high levels of active or acylated ghrelin and normal levels of unacylated ghrelin, reflecting a relative unacylated ghrelin deficit [6]. Most studies also agree finding low concentrations of pancreatic polypeptide (PP) [7–9]. The results are controversial with regard to other hormones such as peptide YY (PYY) [10–15] or Glucagon-like peptide-1 (GLP-1) [16–19], and there are few studies exploring others such as glucose-dependent insulinotropic polypeptide (GIP) [8,15], amylin [20] or brain-derived neurotrophic factor (BDNF) [21,22]. Given that ghrelin is a potent orexigen [23], the constant finding of high concentrations of such peptide in subjects with PWS suggests that this may be the cause of its hyperphagia. However, some studies show that hyperghrelinemia is present in very young children before the onset of hyperphagia [5,24] and the reduction of ghrelin concentrations with an infusion of somatostatin or with analogues thereof does not induce a reduction in appetite or body weight or a change in eating behavior in these patients [25,26]. Livoletide, an unacylated ghrelin analogue, showed initially a significant reduction in hyperphagia, waist circumference and fat mass in PWS subjects, without changes in ghrelin concentrations [27], but these effects were not sustained after three months of treatment [28]. PP is a peptide that induces satiety [29]. The finding of low concentrations in patients with PWS suggests that it may be involved in their lack of satiety. However, the effects of its infusion are controversial. In children at high doses, it does not decrease food intake [9]. With more physiological and repeated doses, a reduction in intake is achieved in adult women with PWS [30]. Treatment with GLP-1 receptor agonists is used in subjects with PWS. There is a study in rats [31] in which exendin-4 induces satiety and decreases ghrelin by 74%, and a report of a female PWS patient [32] in which liraglutide increases insulin and reduces GLP-1 and ghrelin concentrations. However, in other studies in rats [33] and humans [34,35], treatment with GLP-1 receptor agonists did not induce changes in ghrelin concentrations. No infusion or blocking experiments have been performed on BDNF, amylin or GIP in PWS or PWS animal models. Treating children with PWS using brain peptides such as oxytocin, for a short period of time, resulted encouraging. A double-blind, placebo-controlled, crossover study with intranasal oxytocin (low doses) for five days or intranasal placebo spray, was safe and showed a reduction in appetite drive, and improvements in socialization, anxiety, and repetitive behaviors [36]. However, other studies for a longer period of time found no effect of oxytocin at 8 weeks of treatment [37] or a different degree of response at 3 months of treatment depending on sex and genetic subtype (boys and deletion were responders) [38]. Thus, appetite regulation is an extraordinarily complex process involving multiple peptides that interact with each other in a synergistic or antagonistic way. Therefore, it is plausible that hyperphagia in PWS is due to a failure of some overlapped peptide signals rather than to a single peptide alteration. In the present study, we aimed to evaluate different peptides involved in the regulation of appetite in the PWS individuals and assess the possible existence of a hormonal pattern that may help to explain the exaggerated hyperphagia in this syndrome. We compared the hormonal pattern to age and sex matched obese and lean controls. J. Clin. Med. 2021, 10, 5170 3 of 14 2. Materials and Methods 2.1. Participants We studied 30 adult patients with genetically confirmed PWS, 30 obese controls matched by age, sex and body mass index (BMI) and 30 lean controls matched by age and sex. Genetic subtype was type 1 deletion (n = 7), type 2 deletion (n = 13), uniparental disomy (n = 7) and imprinting defects (n = 3). Patients with PWS and obese controls were recruited from the Endocrinology and Nutrition Department and lean controls were hospital staff or acquaintances volunteers. The study was approved by the local ethics committee (Comitè d’Ètica d’Investigació amb medicaments del Parc Taulí Ref 2011/511) in January 2011. The informed consent of the participants or their parents or guardians was obtained, and the investigations were carried out in accordance with the Helsinki Declaration. All subjects had a stable weight for at least three months before the inclusion. No patient with PWS was receiving GH treatment at the time of the study, although 14 had pre- viously received it. Twenty-seven had hypogonadism, of which 10 were in treatment. Seven PWS patients and one obese control had type 2 diabetes, being treated with metformin alone or in combination with GLP1 receptor agonists, DPP4 inhibitors or insulin. 2.2. Experimental Methods Subjects were admitted at 8 a.m. after an overnight fast of at least 10 hours. An- thropometric measures were taken, and body composition analysis was performed using bioelectrical impedance (TANITA, body composition analyzer BC-418 MA Biologica Tec- nologia Medica SL-BCN, Tokyo, Japan). A blood sample was taken for routine analysis and fasting hormonal analysis. For the postprandial study, subjects ingested a standard liquid diet (Resource 2.0, Nestle Lab, 1200 kcal, 43% carbohydrate, 39% fat, 18% protein) and blood samples were drawn at 30, 60 and 120 min after the intake. All samples were collected on ice and spun at 4 ◦ C. Plasma was centrifuged and stored at −80 ◦ C until processed. Assays were performed using commercially available methods. BDNF was quantified by an enzyme-linked immunosorbent assay (ELISA) kit (Cat. No. CYT306, Millipore; Billerica, MA, USA) following the manufacturers’ instructions. This kit’s intra-assay variation is ±3.7% and its inter-assay variation is ±8.5% (125 pg/mL). Ghrelin (total and active), amylin, GIP, GLP-1, insulin, leptin, PP and PYY concentrations were determined using Luminex 200 and the human metabolic hormone magnetic bead panel (HMHMAG- 34K, Merck Millipore; Billerica, MA, USA) following the manufacturer’s instructions. The sensitivity of this kit is as follows: total and active ghrelin 11 pg/mL, amylin 4 pg/mL, GIP 0.7 pg/mL, GLP-1 20 pg/mL, insulin 48 pg/mL, leptin 184 pg/mL, PP 5 pg/mL, PYY 26 pg/mL. The intra and inter assay coefficients of variation are, respectively as follows: total and active ghrelin 3% and 21%, amylin 3% and 7%, GIP 4% and 10%, GLP-1 2% and 13%, insulin 3% and 10%, leptin 7% and 10%, PP 6% and 9%, PYY 7% and 8%. Subjects (three patients with PWS) receiving GLP1 receptor agonists or dipeptidil peptidasa-4 (DPP4) inhibitors for diabetes treatment were excluded for GLP-1 and GIP determinations and for cluster analysis. Before the meal and 600 and 1200 after the meal, subjects quantified their hunger on a visual analogue scale ranging from 0 to 100. 2.3. Statistical Methods We used a Kruskal-Wallis test and a Mann-Whitney test with Bonferroni correction for pairwise comparisons between groups. We used Wilcoxon signed rank test to detect statistically significant differences of the concentration evolution (from baseline). We elaborated a generalized linear regression model with repeated measures for an indicator of values higher than the median score for each peptide including cluster_t0 and time and the interaction between them. J. Clin. Med. 2021, 10, 5170 4 of 14 A hierarchical cluster analysis [39] was conducted to reveal groups of patients with similar hormonal concentrations based on the factors derived from the principal compo- nents analysis. The clusters obtained are represented in a tree dendrogram. The distribution of the sample within each cluster was analyzed using Mann-Whitney-Wilcoxon test for continuous variables and chi-square test for categorical variables. A classification tree analysis [40] was conducted to discriminate patient groups (PWS, obese, lean controls) through hormonal concentrations levels. Entropy and the cost- complexity criterion were used to build the tree. The number of terminal nodes was determined using 10-fold cross validation. We used SAS system v9.4 (SAS Institute Inc., Cary, NC, USA) for all analyses. 3. Results Table 1 shows subjects’ baseline characteristics. Lean controls had a lower BMI, percentage of body fat and waist circumference than PWS subjects and obese controls. Five subjects on insulin treatment were excluded for insulin determinations. Fasting glucose levels were lower in lean subjects than in obese controls. Insulin and insulin resistance determined by homeostatic model assessment (HOMA-IR-index) were higher in obese controls than in PWS subjects and lean controls. Table 1. Baseline characteristics of all groups. PWS (n = 30) Obese Controls (n = 30) Lean Controls (n = 30) Sex (M/F) 15/15 15/15 15/15 Age (year) 27.5 ± 8.02 28.4 ± 7.13 27.9 ± 7.77 BMI (kg/m2 ) 32.4 ± 8.14 33.7 ± 6.88 22.1 ± 2.05 *$ Body fat (%) 37.0 ± 8.39 35.7 ± 9.70 20.3 ± 7.23 *$ Waist (cm) 105.0 ± 18.3 105.3 ± 15.1 78.3 ± 7.43 *$ Glucose (mg/dL) 94.1 ± 22.0 98.2 ± 32.1 86.7 ± 6.64 $ Insulin (µU/mL) 333.7 ± 218.5 493.2 ± 213.7 * 304.8 ± 134.7 $ HOMA-IR 1.69 ± 0.84 2.94 ± 1.57 * 1.69 ± 0.77 $ p < 0.05 vs. PWS * and obese $ patients. All quantitative values are expressed as mean ± SD. PWS: Prader-Willi Syndrome; BMI: body mass index; HOMA-IR: homeostatic model assessment insulin resistance. 3.1. Fasting Study Table 2 shows the results of the fasting study. Table 2. Fasting hormonal concentrations. PWS Obese Controls Lean Controls BDNF (pg/mL) 113.7 (75.6–237.7) 187.6 (119.3–303.0) 158.2 (118.0–275.8) Leptin (ng/mL) 35.3 (20.8–42.9) 18.5 (12.8–25.4) * 4.4 (2.7–9.2) *$ Total ghrelin (pg/mL) 572.4 (379.8–812.5) 279.0 (199.4–396.2) * 477.2 (350.7–633.6) $ Active ghrelin (pg/mL) 29.0 (1.0–35.8) 9.5 (5.5–15.0) 12.3 (8.6–16.0) PYY (pg/mL) 149.2 (134.2–176.6) 101.3 (72.0–134.3) * 34.6 (1.0–67.7) *$ PP (pg/mL) 29.9 (9.5–55.3) 23.5 (12.5–38.3) 31.8 (23.7–61.2) GLP-1 (pg/mL) 68.5 (55.8–77.0) 38.1 (33.7–53.3) * 32.5 (19.3–49.6) * GIP (pg/mL) 30.2 (21.1–47.4) 27.5 (18.5–35.6) 24.8 (19.6–32.7) Amylin (pg/mL) 19.9 (7.8–28.5) 22.4 (15.6–27.1) 20.9 (16.2–24.2) p < 0.05 vs. PWS * and obese $ patients. All quantitative values are expressed as median (interquartile range). Plasma brain-derived neurotrophic factor (BDNF), peptide YY (PYY), pancreatic polypeptide (PP), Glucagon-like peptide-1 (GLP-1), glucose-dependent insulinotropic polypeptide (GIP). Total ghrelin concentrations were higher in PWS patients than in obese controls (stat = 604.00, p < 0.001). Active ghrelin tended to higher values in PWS subjects than in controls. Plasma BDNF levels tended to be lower (χ2 = 5.78, p = 0.05) in PWS patients than in obese and lean controls. Leptin concentrations were higher in PWS subjects than in controls and lower in lean than in obese controls (χ2 = 55.7, p < 0.001). Fasting PYY was higher in PWS than in controls and higher in obese than in lean subjects (χ2 = 49.05, p < 0.001). GLP-1 concentrations were higher in PWS (χ2 = 30.91, p < 0.001). No differences Total ghrelin concentrations were higher in PWS patients than in obese controls (stat = 604.00, p < 0.001). Active ghrelin tended to higher values in PWS subjects than in controls. Plasma BDNF levels tended to be lower (χ2 = 5.78, p = 0.05) in PWS patients than in obese and lean controls. Leptin concentrations were higher in PWS subjects than in controls and J. Clin. Med. 2021, 10, 5170 5 of 14 lower in lean than in obese controls (χ2 = 55.7, p < 0.001). Fasting PYY was higher in PWS than in controls and higher in obese than in lean subjects (χ2 = 49.05, p < 0.001). GLP‐1 concentrations were higher in PWS (χ2 = 30.91, p < 0.001). No differences were found in fasting concentrations were found of GIP, amylinoforGIP, in fasting concentrations PP. amylin Statistically or PP.significant pairwise Statistically comparisons significant pairwise are indicatedare comparisons in Table 2. in Table 2. indicated 3.2. 3.2. Postprandial Study The 60′0 sample The 60 sample could could not not be be collected collected inin one onesubject subjectwith withPWS. PWS. Patients Patients with withPWSPWSpresented presented higher concentrations higher of leptin concentrations than both of leptin thancontrol both groups control (p < 0.001), groups (p