Received: 4 June 2021 Revised: 31 August 2021 Accepted: 3 September 2021 DOI: 10.1002/ctm2.575 LETTER TO TH E EDITOR Changes in glucagon-like peptide 1 and 2 levels in people with obesity after a diet-induced weight-loss intervention are related to a specific microbiota signature: A prospective cohort study Dear Editor, main driver of the data variability. Accordingly, we rejected Glucagon-like peptide (GLP)-1 and -2 are enteroendocrine the hypothesis of microbial composition independence hormones released postprandially to regulate glucose among pre- and post-intervention observations from the metabolism.1,2 For reasons that are unclear, GLP-1/2 secre- same patient, which stresses the importance of account- tion is impaired in obesity.3 The enteroendocrine sys- ing for a paired data setting in the downstream analysis tem has been linked to gut microbiota,4 but whether (results obtained at family taxonomic level, Figure 1B). this could be affected by obesity-induced dysbiosis is Association analysis of clinical parameters and gut unknown. Here, we investigated the relationship between microbiota revealed several significant associations GLP dynamics and gut microbiota in a prospective cohort between different bacterial families and body mass index, of patients with obesity enrolled in a 6-month restrictive HOMA-IR, and/or plasma urates (Figure 1C). diet interventional weight-loss trial (trial registration num- Fasting GLP-1/2 levels were inversely related to ber ISRCTN12973246), where effective weight loss (≥10%) Spirochaetaceae, Peptostreptococcaceae, Clostridiaceae, was achieved at the end of the follow-up. We determined Ruminococcaceae, Erysipelotrichaceae, Eubacteriaceae, the association between GLP secretion in response to and Lachnospiraceae, and positively related to Pre- nutrient ingestion and specific metagenomic signatures votellaceae, Acidaminococaeae, Selenomonadaceae, using taxonomic profiling and functional analysis. Coriobacteriaceae, Veillonelaceae, Odoribacteriaceae, The main characteristics of the cohort are described Bacteriodaceae, Rikenellaceae, Porphyromonadaceae, in Table S1. Patients showed improvement in glucose Desulfovibrionaceae, and Sutterellaceae. Notably, we and lipid profiles at 6 months. Notably, fasting GLP-1/2 found a contrasting correlation pattern for GLP-1/2 levels were significantly lower after weight loss, with a response in the MTT, as Spirochaetaceae showed a trend for improved GLP response to a meal tolerance test significant positive correlation with GLP-1 response, (MTT). The intervention had no significant impact on whereas the GLP-2 response was negatively correlated intra-community (alpha) diversity measured by species with Acidaminococcaceae, Desulfovibrionaceae, and richness or evenness (Figure S1), in agreement with Sutterellaceae (Figure 1C). other reports.5–7 Nonetheless, a higher alpha diversity was To address the statistical sparsity issues and the lack of observed in patients with lower plasma urate levels, lower homogeneously distributed variables among individuals systolic blood pressure, and lower weight (p = 0.008, 0.026, (common to microbiome data analysis), we applied the and 0.04, respectively), all related to a “healthy” metabolic FitZig mixture model,8 which allows for discrimina- status. No differences were found in inter-community tion of taxonomic difference abundances between the (beta) diversity, as estimated by the Bray–Curtis distance study groups. Differential abundance analysis pre-/post- (p = 0.216), or when microbial populations at different tax- intervention revealed more than 40 species or genera with onomic levels were compared before and after the inter- a significant variation in terms of fold-change and statis- vention (Figure 1A). Remarkably, the unique patient iden- tical relevance when considered simultaneously (Figure tifier variable correlated with all tested principal compo- S3). Abundance ranking analysis revealed noteworthy nents (Figure S2), pointing to individual variation as the interindividual variations, suggesting that communities This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Clinical and Translational Medicine published by John Wiley & Sons Australia, Ltd on behalf of Shanghai Institute of Clinical Bioinformatics Clin. Transl. Med. 2021;11:e575. wileyonlinelibrary.com/journal/ctm2 1 of 7 https://doi.org/10.1002/ctm2.575 2 of 7 LETTER TO THE EDITOR F I G U R E 1 Taxonomy composition of the study population. (A) Comparison of relative abundances of different taxa at the family level between post-intervention (outerchart) and basal (inner chart) faecal microbiota, showing no significant differences. (B) Individual variability between basal (B) and post-intervention (6 M). The presence of different microorganisms is indicated by colours; the same colour indicates the same family. Microbial taxa are based on the family level. Significant differences were observed for each individual after diet intervention (p-value < 0.05). The chi-square test was used for analysis (n = 18 classified as basal (B) and post-intervention (6 M)). (C) Associations of gut microbiota with clinical variables and intestinal hormones. Heatmap depicting Spearman’s rank correlation coefficient of the relative abundances at the family level of different gut microbiota and clinical variables and incretin hormones in all individuals. For correlational studies, all gathered data (basal and 6 months) are included in the analysis. Adjusted p-values: #padj < 0.25; *padj < 0.05. Peptostreptococcaceae, Clostridiaceae, Ruminococcaceae, Oscillospiraceae, Eubacteriaceae, and Lachnospiraceae showed negative associations with BMI, HOMA-IR index, and/or plasma urates. By contrast, Streptococcaceae, Prevotellaceae, Selenomonadaceae, Coriobacteriaceae, Bacteriodaceae, Rikenellaceae, Porphyromonadaceae, and Sutterellaceae showed positive associations with the afore mentioned parameters. Odoribacteriaceae showed a negative association with body mass index (BMI) and a positive association with plasma urates LETTER TO THE EDITOR 3 of 7 4 of 7 LETTER TO THE EDITOR from the same individual were generally more similar to with GLPs response (Clostridium sp. CAG:75, Clostridium one and other than to those of other individuals, with indi- sp. CAG:127, Clostridium sp. CAG: 230) had similar gene vidual variability more important than diet–responsive association patterns. These patterns contrasted with those variation (Figure 1B). These results are in line with previ- for Mitsuokella_uc, whose abundance decreased at follow- ous studies showing that human gut microbiota response up and inversely correlated with GLP response. Only a to diet has a strong individuality and indicating that the minor fraction of the total studied genes (12%, seven genes) effects of the diet per se as the main source of variability could be projected onto a metabolic map (Figure 3B), indi- are markedly diminished, which is the opposite to that cating that most of the altered genes have other biologi- described in mice.9,10 cal functions. Indeed, no particular pathway was enriched, Most of the significantly increased taxa at follow- suggesting a global modulation of gut microbiota in terms up belonged to the Clostridiaceae family, specifically of metabolism. the genus Clostridium (Figure 2). Analysis of GLP-1/2 Our results lead us to consider that while diet-induced dynamics revealed a positive association between Mit- weight loss impacts gut microbiota composition, and the suokuella_uc and fasting GLP-1/2 levels, but a signifi- global metabolic regulation of microbial community inter- cant negative association between Mitsuokuella_uc and actions maintains a balanced ecosystem. This would agree GLP-2 response in the MTT. Contrastingly, Clostridium sp. with the plasticity of gut microbiota, allowing rapid adap- CAG:75, CAG:230, and CAG:127 and Lachnospiraceae bac- tion to environmental changes. In conclusion, our study terium 5_1_63FAA (which all increased after the weight shows that GLP secretion, which is dependent on body loss intervention) correlated negatively with fasting levels weight and metabolic status, is linked to specific gut of GLP-2 but positively with its response in the MTT. Of microbiota. Further mechanistic studies are warranted to note, Clostridium sp. CAG:127 showed a significant positive understand how gut microbiota regulate GLP secretion correlation with GLP-2 response. Similarly, Clostridium sp. and whether those bacterial strains associated with GLP CAG: 230 correlated negatively with fasting GLP-1 but pos- dynamics might represent novel probiotic approaches to itively with GLP-1 response in the MTT (Figure 2). Overall, restore host mucosal GLP response in obesity. the data indicate that fasting levels of GLPs are inversely related to their response to the MTT, with higher levels AC K N OW L E D G M E N T S signalling poor responsiveness. Notably, this behaviour, The study was supported by grants from the Instituto de which differed for GLP-1 and GLP-2 was, at least in part, Salud Carlos III (PI14/00228, PI17/0153 and PI20/00338 associated with a specific microbiota signature. to J.V.), Ministerio de Ciencia e Innovación (RTI2018- The study at functional gene level identified signifi- 093919-B-I00 to S.F.-V. and PID2019-105969GB-I00 to A.M) cant associations between the expression of non redun- and Generalitat Valenciana (PROMETEO/2018/A/133 to dant genes aligned in the KEGG database. A total of A.M), co-financed by the European Regional Development 116 genes were found to be statistically significant, how- Fund. The Spanish Biomedical Research Center in Dia- ever, only 54 of them could be identified and were betes and Associated Metabolic Disorders (CB07708/0012) distributed into different biologically relevant functions is an initiative of the Instituto de Salud Carlos III. S.F- (Table S2). V. acknowledges the Miguel Servet tenure-track pro- Hierarchical clustering of the correlation data matrix gram (CP10/00438 and CPII16/00008) from Fondo de between genes and taxonomy at the gene level revealed Investigación Sanitaria. We would like to acknowledge the presence of two highly populated clusters (Fig- the patients for participating in the study, and the ure 3A). Remarkably, except for Lachnospiraceae bac- BioBank IISPV (PT17/0015/0029) integrated into the Span- terium 5_1_63FAA, those species that increased in abun- ish National Biobanks Network for its collaboration. IISPV dance after the intervention and were positively associated is a CERCA institution. F I G U R E 2 Diet-induced microbiota changes and associations between species and variables related to metabolic health status. Log fold-change (FC) and Spearman correlation analyses of the relative abundance at the species level of gut microbiota and clinical variables and incretin hormones in all individuals. For FC analysis, a zero-inflated Gaussian mixture model (fitZig) from the metagenomeSeq R package was used, contrasting 6 M/basal. The subject factor in the patient identifier variable (IDPAT) is used as a batch effect, as the inter-individual differences in the microbiota were greater than the changes caused by the diet. As a consequence, the IDPAT variable was introduced as an adjusting covariate in the model to investigate diet-induced changes after the 6-month weight-loss program. Spearman correlation analysis revealed a negative correlation between most of the Clostridium species, Hemophilus_uc, and Eubacterium sp. CAG 192 and plasma urates, whereas Clostridium sp. CAG:75 and Eubacterium sp. CAG 202 were negatively correlated with HOMA-IR index. Proteobacteria sp. CAG:873 and Mitsuokella multocida, showed a significant positive correlation with body mass index (BMI) and HOMA-IR, respectively. All padj < 0.005 in FC and #padj < 0.25; *padj < 0.05 for Spearman correlation analyses LETTER TO THE EDITOR 5 of 7 6 of 7 LETTER TO THE EDITOR CONFLICT OF INTEREST fulfilled by the corresponding authors. The metagenome The authors declare no conflict of interest. data sets from this study are available in the EBI Short Read Archive under the study accession PRJEB43147, with E T H I C S A P P R O VA L A N D C O N S E N T T O accession numbers [ERX5107074 – ERX5107109]. PA R T I C I PA T E All participants gave their informed consent, and the study F U N D I N G I N F O R M AT I O N was reviewed and approved by the ethics and research Instituto de Salud Carlos III, Grant Numbers: PI14/00228, committee of University Hospital Joan XXIII from Tarrag- PI17/0153 and PI20/00338; Ministerio de Ciencia e ona, Spain. Innovación, Grant Numbers: RTI2018-093919-B-I00 and PID2019-105969GB-I00; Generalitat Valenciana, Grant AU T H O R CO N T R I B U T I O N S Number: PROMETEO/2018/A/133. M-Mar Rodríguez-Peña, Gemma Llauradó, and Brenno Astiarraga participated in the sample processing, func- M-Mar Rodríguez-Peña1,2 tional studies, laboratory determinations, and statisti- Brenno Astiarraga1,3 cal analysis. M-Mar Rodríguez-Peña, Jesus Seco, Brenno Jesus Seco1 Astiarraga, and Victoria Ceperuelo-Mallafré wrote the Victoria Ceperuelo-Mallafré1,3 manuscript. Jesus Seco, Adrià Caballé, Camille Stephan- Adrià Caballé4 Otto Attolini, and Vicente Pérez-Brocal performed statisti- Vicente Pérez-Brocal5,6 cal analysis of taxonomical and clinical features. Andrés Camille Stephan-Otto Attolini4 Moya and Ana Megía provided scientific discussion and Andrés Moya5,6,7 revised the manuscript. Silvia Pellitero and Núria Vilar- Gemma Llauradó2,8 rasa participated in the human sample recruitment and Ana Megía1,2 conducted the clinical trial. Joan Vendrell and Sonia Silvia Pellitero9 Fernández-Veledo conceived, designed, and supervised the Núria Vilarrasa2,10 research project and wrote the manuscript. Joan Vendrell Sonia Fernández-Veledo1,2 and Sonia Fernández-Veledo are also the guarantors of this Joan Vendrell1,2,3 work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data 1 Endocrinologyand Nutrition Service, Hospital and the accuracy of the data analysis. Universitari de Tarragona Joan XXIII, Institut d’Investigació Sanitària Pere Virgili, Tarragona, Spain D A T A AVA I L A B I L I T Y S T A T E M E N T 2 CIBER de Diabetes y Enfermedades Metabólicas All data generated or analyzed during this study are Asociadas (CIBERDEM)-Instituto de Salud Carlos III, included in this published article and its Supplemen- Madrid, Spain tary Information. Further information and requests for 3 Departament de Medicina i Cirurgia, Universitat Rovira i resources and reagents should be directed to and will be Virgili, Tarragona, Spain F I G U R E 3 Spearman correlation analysis of microbiota species and metabolic bacterial function for a set of 45 genes. (A) Hierarchical clustering of the correlation data matrix was conducted at the gene level, which revealed the presence of two highly populated clusters (labelled as 1 and 2). Annotation of metabolic pathways according to the KEGG database, sorted in alphabetical order: [C]: energy production and conversion; [E]: amino acid transport and metabolism; [G]: carbohydrate transport and metabolism; [H]: coenzyme transport and metabolism; [J]: translation; ribosomal structure and biogenesis; [K]: transcription; [L]: replication, recombination, and repair; [M]: cell wall/membrane/envelope biogenesis; [O]: post-translational modification; protein turnover, chaperones; [P]: inorganic ion transport and metabolism; [T]: signal transduction mechanism; [U|W]: intracellular trafficking and vesicular transport; [V]: defence mechanism. Correlations where an associated p-adjusted value was greater than 0.25 were neglected for additional analysis. Genes associated with replication, recombination, and repair (encoded as [L]), such as MobA/MobB, site-specific recombinases, integrases, and transposases, were found to positively correlate with an increase in the presence of Clostridium species. A similar pattern was found (with an equivalent number of enzymes, 4) with post-translational modification enzymes ([O]) and enzymes involved in signal transduction ([T]). An increase in the abundance of the metabolic enzymes glycosyltransferase_36, dicarboxylate transporter, and carbohydrate-solute binding proteins were positively correlated with Clostridiaceae family members. Genes associated with amino-acid transport and metabolism ([E]) and inorganic ion transport ([P]), such as branched-amino acid transporter and ABC-related transporter, were also modified. (B) Gene projection of statistically relevant enzyme subset on a KEGG-modified metabolic enzyme pathway. This analysis revealed only a minor fraction of the total studied genes (12%, seven genes) that could be projected into a metabolic map constellation:Aln-; Gln-t-RNA synthase ([J]), PALM hydrolase ([L]), nicotinate-phosphoribosyltransferase ([H]), N-acetylmuranoyl-l-alanine amidase ([M]), glycerol-phosphoryl-diester-phospho-diesterase ([C]), glycosyltransferase 36 ([G]), and branched-amino acid transporter ([E]). All correspond to different biological pathways LETTER TO THE EDITOR 7 of 7 4 Institute for Researchin Biomedicine (IRB Barcelona), REFERENCES The Barcelona Institute of Science and Technology, 1. Drucker DJ. Mechanisms of action and therapeutic application Barcelona, Spain of glucagon-like peptide-1. Cell Metab. 2018;27(4):740-756. https: 5 Department of Genomics and Health, Foundation for the //doi.org/10.1016/j.cmet.2018.03.001. 2. Guan X. The CNS glucagon-like peptide-2 receptor in the Promotion of Health and Biomedical Research of Valencia control of energy balance and glucose homeostasis. Am. J Region (FISABIO-Public Health), Valencia, Spain Physiol. 2014;307(6):R585-R596. https://doi.org/10.1152/ajpregu. 6 Biomedical Research Networking Centre for Epidemiology 00096.2014. and Public Health (CIBERESP), Madrid, Spain 3. Carr RD, Larsen MO, Jelic K, et al. Secretion and dipeptidyl 7 Spanish National Research Council (CSIC), Institute for peptidase-4-mediated metabolism of incretin hormones after a Integrative Systems Biology (I2SysBio), University of mixed meal or glucose ingestion in obese compared to lean, Valencia, Valencia, Spain nondiabetic men. J Clin Endocrinol Metab. 2010;95(2):872-878. 8 Department of Endocrinology and Nutrition, Hospital del https://doi.org/10.1210/jc.2009-2054. 4. Greiner TU, Bäckhed F. Microbial regulation of GLP-1 and L-cell Mar, Institut Hospital del Mar d’Investigacions Mèdiques biology. Mol Metab. 2016;5(9):753-758. https://doi.org/10.1016/j. (IMIM), Universitat Autònoma de Barcelona, Barcelona, molmet.2016.05.012. Spain 5. Seganfredo FB, Blume CA, Moehlecke M, et al. Weight-loss 9 Endocrinology and Nutrition Department, Institute interventions and gut microbiota changes in overweight and Research and Hospital Germans Trias i Pujol, Universitat obese patients: a systematic review. Obes Rev. 2017;18(8):832-851. Autònoma de Barcelona, Barcelona, Spain https://doi.org/10.1111/obr.12541. 10 Department of Endocrinology and Nutrition, 6. Pataky Z, Genton L, Spahr L, et al. Impact of hypocaloric hyper- proteic diet on gut microbiota in overweight or obese patients L’Hospitalet de Llobregat, Bellvitge University-IDIBELL, with nonalcoholic fatty liver disease: a pilot study. Dig Dis Sci. Barcelona, Spain 2016;61(9):2721-2731. https://doi.org/10.1007/s10620-016-4179-1. 7. Louis S, Tappu RM, Damms-Machado A, Huson DH, Bischoff Correspondence SC. Characterization of the gut microbial community of obese Joan Vendrell and Sonia Fernández-Veledo, Research patients following a weight-loss intervention using whole Unit, University Hospital of Tarragona Joan XXIII, c/ metagenome shotgun sequencing. PLoS One. 2016;11(2). https: Dr.Mallafré Guasch, 4, 43007 Tarragona. Spain. //doi.org/10.1371/journal.pone.0149564. Email: jvo@comt.es; sonia.fernandez@iispv.cat; 8. Paulson JN, Colin Stine O, Bravo HC, Pop M. Differential abun- dance analysis for microbial marker-gene surveys. Nat Methods. sonia.fernandezveledo@gmail.com 2013;10(12):1200-1202. https://doi.org/10.1038/nmeth.2658. 9. Zhang C, Zhang M, Wang S, et al. Interactions between gut ORCID microbiota, host genetics and diet relevant to development of M-Mar Rodríguez-Peña https://orcid.org/0000-0003- metabolic syndromes in mice. ISME J. 2010;4(2):232-241. https: 2553-9769 //doi.org/10.1038/ismej.2009.112. Brenno Astiarraga https://orcid.org/0000-0003-2216- 10. Healey GR, Murphy R, Brough L, Butts CA, Coad J. Interindi- 8974 vidual variability in gut microbiota and host response to dietary Jesus Seco https://orcid.org/0000-0002-4049-8873 interventions. Nutr Rev. 2017;75(12):1059-1080. https://doi.org/ 10.1093/nutrit/nux062. Victoria Ceperuelo-Mallafré https://orcid.org/0000- 0002-4460-9761 Ana Megía https://orcid.org/0000-0002-5101-9452 Sonia Fernández-Veledo https://orcid.org/0000-0003- S U P P O RT I N G I N F O R M AT I O N 2906-3788 Additional supporting information may be found in the Joan Vendrell https://orcid.org/0000-0002-6994-6115 online version of the article at the publisher’s website.