Fecal transplantation from humans 
with obesity to mice drives a 
selective microbial signature 
without impacting behavioral and 
metabolic health
Audrey M. Neyrinck1,9, Hany Ahmed2,9, Quentin Leyrolle1,3,9, Sophie Leclercq4, 
Camille Amadieu1,3, Topi Meuronen2, Sophie Layé3, Patrice D. Cani1,5,6, Olli Kärkkäinen7,8, 
Laure B. Bindels1,5, Kati Hanhineva2,10 & Nathalie M. Delzenne1,10
Obesity is associated with alterations in the gut microbiome that may contribute to metabolic and 
mental health disturbances. Fecal microbiota transplantation (FMT) from humans to mice is a model 
proposed to study human microbiota-associated disorders. In this study, we investigated whether 
gut microbiota from human donors with obesity could affect behavior and metabolomic profiles of 
mice. Stools from donors with obesity and from lean donors were inoculated to antibiotic-pretreated 
mice fed a standard low-fat diet throughout the experiment. Obese-recipient mice exhibited a lower 
bacterial alpha-diversity and limited changes in specific taxa (e.g., an increase in Eubacterium) but were 
similar to lean-recipient mice in terms of dietary intake, body weight, fat mass, anxiety/depression-like 
behavior and glucose homeostasis. Non-targeted LC–MS metabolomic analysis revealed no change 
in the portal and cava serum samples. However, 1-methylnicotinamide, indole-3-acetic acid (I3A) and 
methyllysine were increased in the cecal content of obese-recipient compared to lean-recipient mice. 
Microbial metabolites derived from amino acids were positively correlated with Eubacterium. These 
results indicate that FMT from donors with obesity to mice fed chow diet (low in lipids) leads to minor 
but persistent change in intestinal microbial-derived metabolites, without recapitulating the metabolic 
and behavioral alterations of obesity.
The etiological triad of diet, genome, and gut microbiota plays a crucial role in the pathophysiology of obesity, 
considered a pandemic of the 21st century. The exploration of the role of the microbiome in obesity and 
metabolic diseases has been supported by experiments of fecal microbiome transfer (FMT). Studies conducted 
in diabetic patients have shown that heterologous FMT from non-diabetic to diabetic subjects can improve 
insulin sensitivity. However, the effectiveness of this treatment is temporary and varies between individuals1,2. 
FMT from human donors to recipient mice has also been performed to investigate the causal role of the gut 
microbiota on physiology and its impact on responses to nutritional disorders. Notably, humanized mice fed the 
Western diet exhibit increased adiposity, a trait that can be transmitted through microbiota transplantation3. We 
have previously demonstrated that the improvement of metabolic alterations by a dietary fiber (inulin) in mice 
transplanted with the gut microbiome from patients with obesity was variable. This variability depended on both 
the microbiome characteristics of the donor and changes in specific bacterial taxa following inulin feeding4. 
1Metabolism and Nutrition Research Group, Louvain Drug Research Institute (LDRI), UCLouvain, Université 
catholique de Louvain, Brussels, Belgium. 2Food Sciences Unit, Department of Life Technologies, University of Turku, 
Turku, Finland. 3INRAE, Bordeaux INP, NutriNeurO, UMR 1286, Université de Bordeaux, 33000 Bordeaux, France. 
4Laboratory of Nutritional Psychiatry, Institute of Neuroscience, UCLouvain, Université catholique de Louvain, 
Brussels, Belgium. 5WELBIO Department, WEL Research Institute, Wavre, Belgium. 6Institute of Experimental and 
Clinical Research (IREC), UCLouvain, Université catholique de Louvain, Brussels, Belgium. 7School of Pharmacy, 
University of Eastern Finland, Kuopio, Finland. 8 Institute of Public Health and Clinical Nutrition, University of 
Eastern Finland, Kuopio, Finland. 9Audrey M. Neyrinck, Hany Ahmed and Quentin Leyrolle contributed equally 
to this work. 10These authors jointly supervised the work: Kati Hanhineva and Nathalie M. Delzenne. email:  
Nathalie.Delzenne@uclouvain.be
OPEN
Scientific Reports |        (2025) 15:15455 1| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports
Obesity is often accompanied by alterations in psychological traits, which can lead to depression, altered social 
behavior or addictions5. The gut-brain axis plays as crucial in human mental health and behavior, with the gut 
microbiome being a key component of this axis. This is primarily due to its production of neurotransmitters or 
other bioactive metabolites involved in neuroendocrine processes6–8.
Metabolomics, i.e. the systematic identification and quantitation of all compounds involved in metabolic 
pathways (metabolites) within a given organism or biological sample9, is a valuable method for elucidating 
the crosstalk between host metabolism and gut microbiota8,10. Numerous studies have shown that the serum 
metabolome contains a plethora of biomarkers and causative agents of various diseases. Some of these 
metabolites have been taken up from the environment while others are produced endogenously. The origins of 
specific metabolites are known, including those that are highly heritable, as well as those influenced by lifestyle 
choices (e.g. smoking), diet or the gut microbiome. Analysis of the fecal gut microbiota and plasma non-
targeted metabolomics in subjects with increased negative mood revealed elevated levels of Coprococcus and 
decreased levels of  Sutterella  and  Lactobacillus11. Serum metabolomic analysis in these subjects showed 
altered levels of several amino acid-derived metabolites, including an increase in L-histidine and a decrease 
in phenylacetylglutamine. These changes are associated with altered gut microbiota composition and function 
rather than differences in dietary amino acid intake11. However, the proof of concept that the gut microbiome 
characteristics drive the occurrence of psychological trait has not yet been fully demonstrated.
In a previous study, we demonstrated that transplanting the gut microbiota from patients with alcohol-
use disorder into mice replicated some behavioral changes linked to alcohol dependence, including reduced 
sociability, increased depression-like behavior, and higher stress level. We also showed that FMT lead to changes 
in the production of bacterial-derived metabolites (namely ethanol) which may explain the brain-dysfunction 
associated with the altered microbiome12. Here, we hypothesized that the gut microbiome associated with 
obesity in humans might drive behavioral changes independent of effects on adiposity or metabolic disorders. 
The objective of this study was to determine whether FMT from human donors with obesity—who exhibit 
varying levels of psychological mood scores—can trigger changes in the production of metabolites related to 
metabolic and behavioral traits (sociability, depression) in recipient mice. To investigate this, we conducted 
untargeted metabolomics analyses of the serum and intestinal contents of the recipient mice to explore the 
potential contribution of gut microbiome, nutrition- or host-derived metabolites.
Results
Experimental design
Ten donors with obesity (7 men and 3 women) and 5 lean control donors (3 men and 2 women) were selected 
based on their body mass index (BMI) in order to inoculate fecal microbiota into recipient antibiotic pretreated-
mice (Table S1). Although the psychological mood score, assessed using the Positive and Negative Affect Schedule 
(PANAS), were slightly lower in individuals with obesity, the difference was not significant (p = 0.09). Forty mice 
receiving “obese” microbiota (obese-recipient mice) and 20 mice that receive a “lean” microbiota (lean-recipient 
mice) were fed a normal chow diet for the duration of the experiment. They also underwent behavioral tests and 
oral glucose tolerance test according to the experimental design outlined in Fig. 1 (4 recipient mice/donor; 2 
independent experiments). Based on the Aitchison distance-based hierarchical clustering of the microbiome of 
donors and recipient mice (Fig. S1), we decided to exclude 1 donor and 6 mice from the statistical analyse for all 
the parameters discussed in the next sections.
Fig. 1. Experimental design of mice study. Fecal microbiota transfer (FMT) from donors with obesity (7 men, 
3 women) or lean donors (3 men, 2 women) into antibiotic-pretreated male mice. Mice were fed a normal 
chow diet for 6–7 to weeks following the FMT. Behavioral tests (1. elevated plus maze (EPM); 2. open field 
(OF); 3. object location (OL); 4. light and dark box (LDB); 5. forced swim (FS)) were performed 3–4 weeks 
after FMT. Oral glucose tolerance test (OGTT) were assessed 5–6 weeks after FMT. Created in BioRender. 
Neyrinck, A. (2025) https://BioRender.com/j25w456.
 
Scientific Reports |        (2025) 15:15455 2| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
Characterization of the caecal microbiota of the recipient mice
The number of total bacteria per gram of caecal matter of recipient mice was similar to the number of total 
bacteria per gram of fecal matter of human donors Fig. S1. Finally, 48% and 53% of genera present in the control 
and obese inocula (the percentages were determined from the number of genera found in the cecal content of 
recipient mice relative to the number of genera detected in the inoculum) were successfully transferred to the 
respective recipient mice.
Compared with the microbiota from lean-recipient mice, the microbiota of obese-recipient mice exhibited 
a reduced bacterial α-diversity (Shannon index, Fig. S2) and a modest change in β-diversity (only at the genus 
level, evidenced from the PERMANOVA analyses) (Fig. 2a). Eubacterium was the main genus explaining the 
shift alongside the first component of the PCA of the obese-recipient mice compared to the lean-recipient mice 
(Fig. 2a). Eubacteriaceae is the sole family significantly increased in obese-recipient versus lean-recipient mice 
(Supplementary Dataset 1). Consistently, at the genus level, Eubacterium was more abundant in the caecum of 
obese-recipient than in lean-recipient mice (Fig. 2b, Supplementary Dataset 1). Another genus, Eisenbergiella, 
also increased following the transfer of an obese microbiota. Conversely, two genera, Flavonifractor and 
Barnesiella, significantly decreased in the caecal content of obese-recipient compared to lean-recipient mice 
(Fig. 2b, Supplementary Dataset 1).
Mice inoculated with the obese microbiota exhibit no change regarding behavior and 
metabolic and inflammatory markers
Body weight evolution and glucose homeostasis, assessed by measuring glycemia and insulinemia at fasting state 
and after a load of glucose (OGTT), were not affected by the transfer of humanized microbiota from individuals 
with obesity versus lean-recipient mice (Fig. S3). In addition, there were no significant differences between 
the groups in circulating levels of corticosterone, leptin, brain-derived neurotrophic factor (BDNF) or several 
inflammatory parameters (Table S2). To assess the impact of gut microbiota on anxiety/depression-like behavior 
and memory, recipient mice underwent 6 different tests 3 or 4 weeks after FMT. The light/dark box test, open 
field test, object location test, elevated-plus maze test and the forced swim test showed no differences in memory 
and anxiety/depression-like behavior between obese-recipient and lean-recipient mice (Fig. 3).
Fig. 2. Impact of the transfer of obese and lean human fecal microbiota on mouse caecal microbiota. 
Antibiotic-pretreated recipient mice were inoculated with microbiota from donors with obesity or from lean 
donors. PCA plot of all genera based on the Aitchinson distance, with PERMANOVA analysis and correlation 
circle plot (a). Caecal genera significantly (Data are means ± SD; *q < 0.05, Mann Whitney) altered by obese 
microbiota transfer (b).
 
Scientific Reports |        (2025) 15:15455 3| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
Inoculation with the obese microbiota modulates the metabolome of recipient mice, mainly 
in the caecum
Metabolomic profiles were analyzed in the caecal content and in the blood from portal vein or vena cava of 
recipient mice. The complete set of statistical analysis of data is presented in Supplementary Dataset 2. As 
shown in Fig. 4a, we observed that metabolomic fingerprints were more separated in the caecal content than 
in blood. Among annotated metabolites that discriminated (i.e. VIP > 2) obese from control recipient mice 
in the caecal content (Fig. 4b), 3-indoleacetic acid (I3A), 1-methylnicotinamide and N(6)-methyllysine are 
those that significantly (at the q value < 0.05) increased in obese-recipient versus lean-recipient mice (Fig. 5). 
Although no metabolite was significantly different in the blood after adjusting for false discovery rate (portal 
and systemic) between both groups, I3A is one of the metabolites that also increased in the portal blood (+ 
57%, p = 0.07, VIP score = 1.74) and the vena cava blood (+ 36%, p = 0.07, VIP score = 1.27), in contrast 
to 1-methylnicotinamide and N(6)-methyllysine that were not significantly modified in the portal or systemic 
serum (Fig. 4b, Supplementary Dataset 2).
Increase of 3-indole acetic acid occurring due to inoculation with the obese microbiota is 
associated to specific bacterial changes
We conducted a correlation analyse between metabolites and bacteria significantly impacted by the inoculation 
with obese microbiota (Fig. 6a). Interestingly, the Eubacteriaceae family, and coherently Eubacterium, was 
identified as a key contributor to the elevated levels of I3A -and to a lesser extent N(6)-methyllysine—in 
the caecal content of obese-recipient mice. However, no significant correlation was found between the three 
highlighted caecal metabolites and the behavioral parameters (Fig. 6b). Of note, I3A in the plasma from vena 
cava was negatively and significantly correlated with anxiety parameter obtained from EPM test (Spearman rank 
correlation r = − 0.295, p = 0.035).
Fig. 3. Impact of obese microbiota transfer to mice on anxiety- and depression-like behavior. Antibiotic-
pretreated recipient mice were inoculated with microbiota from donors with obesity or from lean donors. 
Behavioral tests were performed 3–4 weeks after FMT. Data are mean ± SD. p > 0.05 (student t test or Mann 
Whitney). OFT open field test, EPM elevated plus maze test, FST forced swim test, OLT object location test, 
LDBT light and dark box test. Parts of the image come from Fig. 1.
 
Scientific Reports |        (2025) 15:15455 4| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
Discussion
Several studies suggest that the microbiome in obese individuals has an increased capacity to harvest energy 
from the diet and that this trait is transmissible identifying the gut microbiota as a key contributing factor to the 
pathophysiology of obesity13. For instance, microbiota transplantation from mice previously fed an obesogenic 
diet to lean germ-free recipients results in a greater increase in fat mass compared to transplants from lean 
donors. Furthermore, FMT from leptin-deficient ob/ob mice to germ free mice leads to a significantly greater 
increase in total body fat than FMT from lean littermates ob/+ or +/+ mice14. However, our data indicated that 
when mice are fed a chow diet (low in fat) after FMT from a human obese donor compared to a lean donor, there 
is no difference in fat mass, glycemia, or markers of inflammation. This finding contradicts previous observations 
of an obese phenotype transfer from obese human donor to germ-free mice: germ-free mice colonized with fecal 
microbiota from each member of four twin pairs discordant for obesity conveyed significantly greater increases 
in body mass and adiposity than those of lean communities4. The lack of effect on body weight, adiposity and 
glucose homeostasis due to FMT from obese human donors has already been described by Rodriguez et al.15. 
They showed that mice inoculated with fecal matter from obese human donors (after microbiome depletion 
with antibiotics) had a significantly different microbiome compared to mice that received a FMT from lean 
donors. Interestingly, Rodriguez et al. highlighted that diet influenced the microbiome composition irrespective 
of donor body type after 22 weeks of treatment, suggesting that diet is a key variable in the shaping of the gut 
microbiome after FMT. It is important to note that in the present study, we deliberately chose to provide a chow 
diet low in lipid to the mice to avoid any interference from nutrients and fat components on inflammation.
The data presented in this paper refute the initial hypothesis that transferring the gut microbiome from 
patients with obesity modifies anxiety/depressive behavior in mice. Even if there is no change in metabolism and 
behavior in mice transferred with the gut microbiome from donors with obese versus lean phenotype, it is quite 
interesting to note that some specific modifications of selected bacteria, and of bacterial-related metabolites 
are present up to 7 weeks after the transplantation process. The question of the relevance of those microbial 
Fig. 4. Obese microbiota transfer to mice shapes the caecal and blood metabolome. Antibiotic-pretreated 
recipient mice were inoculated with microbiota from donors with obesity or from lean donors. Scores plot 
of the sparse partial least square discriminant analysis (sPLS-DA) of the metabolomic features between 
obese recipient mice and lean recipient mice (a). Variable importance in projection (VIP) plots with the 
top discriminating metabolites (VIP score > 1) identified through PLS-DA analyses in descending order of 
importance (b).
 
Scientific Reports |        (2025) 15:15455 5| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
components merits attention. Our data revealed that 3 amino acid-derived metabolites increased in the caecal 
content of obese-recipient mice compared to lean-recipient mice. The fold changes in abundance were 1.6, 3.0 
and 7.1 for N(6)-methyllysine, 1-methylnicotinamide and I3A, respectively. Methyllysine was first described 
on a bacterial flagellar protein and thereafter was identified on eukaryotic histone proteins16. In fact, the lysine 
methylation of proteins has been established to regulate many cellular processes, including protein interactions 
and cellular signaling transduction. More importantly, I3A was also increased in the portal and systemic plasma 
even if the level of significance was not reached. The plasma and stool levels of I3A are directly regulated by the 
gut microbiota17. I3A is a tryptophan metabolite derived from gut microbiota. Tryptophan is one of the nine 
essential amino acids commonly found in protein-rich foods, such as milk, cheese, eggs, meat, fish, bananas, 
oats, nuts, and beans but also in normal chow diet for mice. Tryptophan is largely absorbed in the small intestine, 
and the fraction that reaches the colon is catabolized by numerous bacterial species, including Escherichia 
coli, Clostridium, Bacteroides, Peptostreptococcus, Eubacterium, and Lactobacillus species, and Ruminococcus 
gnavus18. Here, level of caecal I3A was particularly correlated to Eubacteriacea and Eubacterium; the abundance 
of this family and genus being significantly increased in the caecal content of obese-recipient mice. Of note, the 
human origin of this microbe, particularly in the large intestine, is well documented19. Tryptophan-metabolizing 
pathways into I3A have already been identified in specific members of the human gut microbiota such as 
Eubacterium hallii20. This microbiota-derived indole metabolite has been identified as a postbiotic and a potent 
modulator of host physiology, maintaining in particular the homeostasis of the gut and systematic immunity21. 
It has been shown to ameliorate nonalcoholic fatty liver disease after sleeve gastrectomy and dextran-sulfate 
sodium-induced colitis in mice and has been proposed as potential treatment or prevention for dysbiosis-driven 
diseases17,22. There are several studies linking I3A to behavior23–25, but also that I3A supplementation can exert 
anti-depressive effects on an animal model of chronic stress26. However, while the present study highlights 
a negative correlation between circulating I3A and one of the behavioral parameters (anxiety via EPM test), 
no relationship between this metabolite and depression/anxiety mood in individuals with obesity has been 
documented in the literature. Inoculation of mice with microbiota from obese human donors stimulated dietary 
Fig. 5. Impact of obese microbiota transfer to mice on caecal metabolome. Volcano plot depicting the effect 
size (Cohen’s d) and -Log10 transformed q values derived from Welch’s t-test analyse of the caecal metabolomic 
features different between obese recipient mice and lean recipient mice (a). The grey dots are features not 
significantly different, the red dots are features significantly up-regulated while the blue dots are features 
significantly down-regulated. Significantly altered (*q < 0.05, Welch’s t-test) caecal identified metabolites (b). 
Data are means ± SD.
 
Scientific Reports |        (2025) 15:15455 6| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
tryptophan-dependent production of methylnicotinamide in the ceacal content. Although no relationship was 
found with the abundance of a specific caecal bacteria, this vitamin B3 degradative product, was also shown 
to protect gut barrier function during experimental colitis27,28. In addition, this metabolite regulated insulin 
sensitivity in type 2 diabetic mice, leading to a reduction of hepatic glucose output and improvement of insulin 
resistance29. Early evidence of the association between methylnicotinamide and obesity was observed in 
metabolomic works since methylnicotinamide levels in human urine are positively correlated with the body 
mass index (BMI). Furthermore, methylnicotinamide levels in serum were found increased in the obese people 
and the methylnicotinamide levels in urine were found increased both in obese db/db mice and in obese Zucker 
rats30.
Altogether, our study suggests that inoculation of fecal microbiota from obese human donor to antibiotic-
pretreated mice fed with chow diet leads to minor but persistent change in gut bacteria and intestinal microbial-
derived metabolites without consequence on inflammation, metabolism and behavior.
Methods
Ethics statement
Mice experiments were performed in strict accordance with relevant guidelines and regulations for the care 
and use of animals and in accordance with the EU directive. All mouse experiments were approved by and 
performed in accordance with the guidelines of the local ethics committee for animal care of the Health Sector 
of the Université catholique de Louvain under the specific agreement numbers 2019/UCL/MD/009. Housing 
conditions were as specified by the Belgian Law of 29 May 2013, on the protection of laboratory animals. The 
study is reported in accordance with ARRIVE guidelines (https://arriveguidelines.org).
Human donors
Ten donors with obesity were selected from the FOOD4GUT study that was a 3-month-long, multicentric, single-
blind, placebo-controlled trial. The trial protocol was published on protocols.io  ( h t t p s : / / d o i . o r g / 1 0 . 1 7 5 0 4 / p r o t 
o c o l s . i o . b a i d i c a 6 ) and the trial was registered at ClinicalTrials.gov under identification number NCT03852069. 
Five control donors were obtained from the GUT2BRAIN cohort and were recruited in 2019. Recruitment, 
enrollment, randomization, sample size determination, inclusion and exclusion criteria, and outcomes have 
been previously described31,32 for FOOD4GUT and GUT2BRAIN studies, respectively. Characteristics of the 
Fig. 6. Correlation analyses with caecal metabolites significantly altered with obese microbiota transfer. 
Heatmap of Spearman rank correlation between caecal metabolites and gut bacteria significantly altered by the 
obese microbiota transfer (a). Heatmap of Spearman rank correlation between caecal metabolites significantly 
altered by the obese microbiota transfer and the behavioral parameters (b). *p < 0.05.
 
Scientific Reports |        (2025) 15:15455 7| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
donors are given in the Table S1. Both studies were approved by the “Comité d’éthique Hospitalo-facultaire 
Saint-Luc UCLouvain” (2014/26SEP/48 with amendment obtained on 18/03/2019 and 2014/14AOU/438 with 
amendment obtained on 22/10/2015, respectively) and in accordance to the written informed consent provided 
by participants (leaving the possibility to use biological material for future research outside the context of the 
FOOD4GUT study and GUT2BRAIN study). Both studies have been carried out in accordance with followed 
the ethical guidelines set out in the Declaration of Helsinki. All participants provided written informed consent 
in compliance with the European law 2001/20/CE guidelines, before inclusion.
Experimental design
Specific pathogen-free (SPF) C57BL/6J male mice (Janvier Labs, Le Genest St Isle, France) were housed in a 
controlled environment (12 h daylight cycle) in individually ventilated cages with free access to food and water. 
Mice (aged 4–5 weeks) were inoculated with the fecal material from 10 patients with obesity (Obese-recipient 
mice) versus 5 control individuals (Control-recipient mice) (Fig. 1) according to previous procedures4,12. We have 
already demonstrated that, with such a protocol, the microbial composition of the recipient mice was similar to 
their human inoculum and was stable up to 6 weeks12. Mice were first administered broad-spectrum antibiotics 
cocktail (ampicillin, neomycin, metronidazole: 100 mg/kg; vancomycin: 50 mg/kg) by oral gavage once a day 
using flexible plastic feeding tubes (Instech Laboratories, the Netherlands) for 8 days. Amphothericin-B (1 mg/
kg) was added to the antibiotic cocktail for the first 3 days to prevent fungi overgrowth. At the end of the 
antibiotic treatment, mice were subjected to intestinal purge with PEG 4000 (polytethylene glycol, macrogolum; 
59 g/l supplemented with NaCl: 1.46 g/l; KCl: 0.75 g/l; Na2SO4: 5.68 g/l; NaHCO3: 1.68 g/l) in order to wash out 
the antibiotics remaining in the intestine. 1.5 ml of this intestinal bowel cleansing solution was administered by 
oral gavage to each mouse (2 × 500 μl the day before FMT with 1 h interval, and 1 × 500 μl 4h before the FMT). 
Mice were fasted 4 h before PEG administration. Handling of human fecal samples (previously collected in 
sterile containers and stored at − 80 °C) was performed under anaerobic conditions (vinyl anaerobic chamber, 
Coy Lab, MI, USA). Each fecal sample (0.5 g) was suspended in 5 ml of reduced sterile PBS (0.05% L-cystein-
HCL and N2), then vortexed for 5 min followed by sedimentation for 5 min. The slurry was then passed through 
a 100 μM cell strainer and 200 μl of the suspension was immediately administered to the mice by oral gavage 
with flexible disposable plastic tubes. Aliquots of this suspension were frozen at − 80 °C with 10% glycerol for 
the next gavages. Stool samples were inoculated 3 times in total (one time per day every 2 days) into recipient-
mice (4 mice per donors). Before and after the inoculation with stool samples, all mice were fed a standard diet 
(AIN-93M en pellets, mature rodent diet, ResearchDiets) for up to 7 weeks.
Behavioral testing
Behavioral testing started 20 days after the first FMT. All behavioral tests were recorded by a video camera 
connected to a computer and data analyse was performed by using EthoVision XT software (Noldus, the 
Netherlands), except for the forced swim test which was assessed manually. Three sessions of animal handling 
were performed 1 week before the first behavioral test to reduce the level of stress induced by the experimenter. 
Mice were transported to the behavioral testing room for a 1 h habituation period. All tests have been performed 
during the light phase. Each apparatus was thoroughly cleaned with water and dried between each animal. 
Control-recipient and Obese-recipient mice were tested in a random order and allowed to rest for several days 
between each test. The experimenter was not blinded to the group allocation when performing the testing.
Elevated plus maze (EPM) test
Mice were tested in the EPM apparatus which is elevated at 76 cm off the ground and consisting of 4 arms—2 
open arms and 2 closed arms made with black Plexiglas walls33. The mouse was placed in the intersection of the 4 
arms, facing an open arm, and allowed to explore for 5 min. Time spent and number of entries in the open arms 
were used to assess anxiety-like behavior.
Open field (OF) test
Animal was transferred to the open-field apparatus (40 cm × 40 cm) and was allowed to freely explore the new 
environment for 10 min. EthoVision XT software allowed the analyse of mobility pattern of each mouse, total 
distance traveled (cm) and time spent in the central area of the arena (percentage).
Object location (OL) test
Visual cues were placed in the room and kept constant during all tests. These tests consist of an acquisition phase 
(5 min) and a test phase (5 min), with an inter-test interval (ITI) of 24 h. The arena is open, made of plastic and 
measuring 40 cm × 40 cm with walls 16 cm high and a light intensity of 15 Lux. During the 5 min training, mice 
were placed in the arena and were allowed to freely explore two new and similar objects spaced 15 cm apart. We 
measured the time spent exploring each object. During the test phase, one of the objects from the training phase 
was moved. Both objects and arena were cleaned with 10% of ethanol between tested animals. We measured the 
time spent exploring each object for 5 min.
Light-dark box (LDB) test
Mice were placed for 10 min in a clear Plexiglas enclosure containing a black insert on one side. Time spent and 
number of entries in the light zone were used to assess anxiety-like behavior34.
Forced swim (FS) test
Mice were placed in an inescapable transparent Plexiglas tank (30 cm 320 cm) fill with autoclaved water (15 cm 
from the bottom) at room temperature35. A video camera recorded each mouse for 6 min and manual scoring 
Scientific Reports |        (2025) 15:15455 8| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
was performed by a trained experimenter who was blinded to the group allocation. Duration of immobility and 
latency to immobility were used to assess depression-like behavior. Water was renewed between each mouse. 
Before returning to their home cage, mice were gently dried with paper towels.
Oral glucose tolerance test (OGTT)
One week before the end of the experiment, mice were fasted for 6 hours before being given an oral gavage 
glucose load (2 g glucose per kg body weight). Blood glucose was measured 30 minutes before (timepoint −30), 
just prior to the oral glucose load (timepoint 0) and then after 15, 30, 60, 90 and 120 min. Blood glucose was 
determined with a glucometer (Accu Chek Performa, Roche, Basel, Switzerland) on blood samples collected 
from the tip of the tail vein. Plasma insulin concentration was determined by ELISA (Mercodia, Uppsala, 
Sweden) according to the manufacturer’s instructions.
Blood and tissue sampling
Mice were anesthetized with isoflurane (Forene®, Abbott, England) and blood from portal vein and vena cava 
was centrifuged and serum/plasma samples were stored at − 80 °C until analyses. Mice were then killed by 
decapitation. Brain, liver, subcutaneous adipose tissue and gut (ileum, colon, caecum) were weighted and 
immediately snap-frozen with liquid nitrogen then stored at − 80 °C.
Plasma parameters
Plasma samples obtained from vena cava were used to measure corticosterone (Enzo Life Sciences, Belgium) 
following manufacturer’s instructions. Other markers -Tumor necrosis factor alpha, interleukin 1 beta, 
monocyte chemoattractant protein-1, C-X-C motif chemokine ligand 1, fibroblast growth factor 21, brain-
derived neurotrophic factor, leptin- were determined using the Meso Scale Discovery (MSD) U-PLEX assay 
(Rockville, MD, USA) following the manufacturer’s instructions.
Gut microbiota analyses
The composition of the gut microbiota was analyzed by Illumina sequencing of the 16S rRNA gene, as previously 
described36. We followed the STORMS guidelines for human microbiome research37. DNA was extracted from 
faecal samples following the protocol Q described by Costea et al.38 and uses the QIAamp DNA Stool Mini Kit 
(Qiagen, Germany) and includes a bead-beating step. Treatment with RNAse A was performed (10 mg/ml, 
Thermo Fisher Scientific, USA). DNA concentration was determined, and purity (A260/A280) was checked 
using a NanoDrop 2000 (Thermo Fisher Scientific, USA). Absolute quantification of the total bacterial load 
was performed by quantitative polymerase chain reaction (qPCR) using the primers Bacteria Universal P338F 
( A C T C C T A C G G G A G G C A G C A G) and P518r ( A T T A C C G C G G C T G C T G G). Real-time PCR was performed 
with a QuantStudio3 (Applied Biosystems, The Netherlands) using SYBR Green (GoTaq® qPCR mix, Promega, 
USA) for detection. All samples (0.1 ng/µl) were run in duplicate in a single 96-well reaction plate. Final 
concentrations were as follow: cDNA 2 µl/25 µl, primers 300 nM, and SyberGreen mix 1X (MeteorTaq DNA 
polymerase, dNTP, RT buffer, MgCl2 4 mM, SYBR® Green I, ROX passive reference and stabilizers, as provided 
by the manufacturer). Thermocycling conditions were as follow: initiation step at 95 °C 2 min; cycling stage 
at 95 °C 30 s, 60 °C 30 s, 72 °C 30 s, 40 cycles; melt curve stage at 95 °C 1 s, 65 °C 20 s, increment of 0.1 °C 
every 1 s until reaching 95 °C. Threshold was manually adjusted to reach the linear range of the log-fluorescent 
curves and CT values were determined using the QuantStudio Software (Version 1.4.3, Applied Biosystems, 
The Netherlands). Absolute quantification was achieved through the inclusion of a standard curve (performed 
in duplicate) on each plate generated by diluting DNA from pure culture of L. acidophilus NCFM (fivefold 
serial dilution). Cell counts were determined by plating and expressed as “colony-forming unit” (CFU) before 
DNA isolation. Amplicon sequencing of the microbiome was done at the University of Minnesota Genomics 
Centre. Briefly, the V5–V6 region of the 16S rRNA gene was PCR-enriched using the primer pair V5F_
Nextera (TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGRGGATTAGATACCC) and V6R_Nextera 
(GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGCGACRRCCATGCANCACCT) in a 25 μl PCR 
reaction containing 5 μl of template DNA, 5 μl of 2× HotStar PCR master mix, 500 nM of final concentration of 
primers and 0.025 U/μl of HostStar Taq+ polymerase (QIAGEN). PCR-enrichment reactions were conducted 
as follows: an initial denaturation step at 95 °C for 5 min followed by 25 cycles of denaturation (20 s at 98 °C), 
annealing (15 s at 55 °C), and elongation (1 min at 72 °C), and a final elongation step (5 min at 72 °C). Next, 
the PCR-enriched samples were diluted 1:100 in water for input into library tailing PCR. The PCR reaction was 
analogous to the one conducted for enrichment except with a KAPA HiFi Hot Start Polymerase concentration of 
0.25 U/μl, while the cycling conditions used were as follows: initial denaturation at 95 °C for 5 min followed by 
10 cycles of denaturation (20 s at 98 °C), annealing (15 s at 55 °C), and elongation (1 min at 72 °C), and a final 
elongation step (5 min at 72 °C). The primers used for tailing are the following: F-indexing primer  A A T G A T A 
C G G C G A C C A C C G A G A T C T A C A C[i5] T C G T C G G C A G C G T C and R-indexing primer  C A A G C A G A A G A C G 
G C A T A C G A G A T[i7] G T C T C G T G G GCTCGG, where [i5] and [i7] refer to the index sequence codes used by 
Illumina. The resulting 10 μl indexing PCR reactions were normalized using a SequalPrep normalization plate 
according to the manufacturer’s instructions (Life Technologies). 20 μl of each normalized sample was pooled 
into a trough, and a SpeedVac was used to concentrate the sample pool down to 100 μl. The pool was then 
cleaned using 1X AMPureXP beads and eluted in 25 μl of nuclease-free water. The final pool was quantitated by 
QUBIT (Life Technologies) and checked on a Bioanalyzer High-Sensitivity DNA Chip (Agilent Technologies) 
to ensure correct amplicon size. The final pool was then normalized to 2 nM, denatured with NaOH, diluted 
to 8 pM in Illumina’s HT1 buffer, spiked with 20% PhiX, and heat denatured at 96 °C for 2 min immediately 
prior to loading. A MiSeq 600 cycle v3 kit was used to sequence the pool. Subsequent bioinformatics analyses 
were performed in-house as previously described36. Initial quality filtering of the reads was performed with 
Scientific Reports |        (2025) 15:15455 9| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
the Illumina Software, yielding an average of 91468 pass-filter reads per sample. Quality scores were visualized 
with the FastQC software (http://www.bioinformatics.babraham.ac.uk/ publications.html), and reads were 
trimmed to 220 bp (R1) and 200 bp (R2) with the FASTX-Toolkit  ( h t t p : / / h a n n o n l a b . c s h l . e d u / f a s t x _ t o o l k i t / ) . 
Next, reads were merged with the merge-illumina-pairs application v1.4.2 (with p = 0.03, enforced Q30 check, 
perfect matching to primers which are removed by the software, and otherwise default settings including no 
ambiguous nucleotides allowed). The UPARSE pipeline implemented in USEARCH v11 was used to further 
process the sequences. Amplicon sequencing variants (ASVs) were identified using UNOISE3. Such method 
infers the biological sequences in the sample prior to the introduction of amplification and sequencing errors 
and distinguishes sequence variants differing by as little as one nucleotide. The analysis allowed the identification 
of 4047 ASVs. ASVs were identified using the RDP database v16. Taxonomic prediction was performed using the 
nbc_tax function, an implementation of the RDP Naive Bayesian Classifier algorithm. Taxonomic identification 
of ASV of interest was refined using EZBioCloud (database version 2023.08.23) and cross-validated using blastn 
and the Prokaryota experimental tanoxonomic nt database. Alpha diversity indexes were calculated using 
QIIME on the rarefied ASV table. Rarefaction was performed using Mothur 1.32.1 by randomly selecting 43323 
sequences for all samples. The dendrogram was generated through a hierachical clustering (Ward D2 method) 
of the Aitchinson distance between samples (namely the euclidian distance between CLR-transformed relative 
abundance levels).
Nontargeted metabolomics analyses
Sample preparation
Three sample types were used for the non-targeted metabolomics analysis in this study—vena cava serum, 
portal serum and caecal content. Serum samples were prepared and analyzed separetely from caecal content. 
All samples were randomized prior to pre-treatment and prepared as described in Klåvus et al.39. In brief, serum 
samples were allowed to thaw on wet ice and once thawn, 100 µl of sample was mixed with 400 µl of ice-cold 
LC-MS grade acetonitrile (Riedel de Haën, Honeywell, Seelze, Germany) in a separate tube. Precipitation was 
facilitated by 2-minute vortexing in room temperature followed by centrifugation for 5 min at 7000×g at 4 °C. 
Supernatant was filtered using a PTFE 0.2 µm syringe filter into a new tube and aliquoted into a brown LC-MS 
vial holding an insert. A separate QC vial was prepared by combining 10 µl of filtered supernatant from all 
samples into a single vial. Ready samples were kept at − 20 °C until analysis. Solvent blanks consisted of solvent 
without additional treatments while extraction blanks devoid of sample followed the same treatment steps as 
samples.
Pre-weighted caecal samples were allowed to thaw on wet ice and mixed with 80% LC-MS grade MeOH 
(Riedel de Haën, Honeywell, Seelze, Germany) in a ratio of 500 µl solvent per 100 mg of sample. Precipitation 
was facilitated by 5-minute vortexing in room temperature followed by centrifugation for 20 min at 17,000×g 
at 4 °C. Filtration, QC and blanks preparations were the same as with the serum samples. Ready caecal extracts 
were kept at − 20 °C until analysis.
LC-MS measurement
Bruker Impact II QTOF-MS instrument couple to an Elute UHPLC 1300 LC system (Bruker Daltonic, Bremen, 
Germany) was utilized for all non-targeted metabolomics data acquisition. Prior data acquisition the MS was 
calibrated using a sodium formate solution. Sample tray was at 4 °C and injection volume at 2 µl for all samples. 
Serum samples were run in a separately from caucum samples. For both sample types, the sequence consisted 
of blank, extraction blank and QC injections prior to sample analysis with QC injections repeated every 13th 
injection. MS/MS data was acquired from QC sample and a randomly selected sample for each sample type using 
separately collision energies of 10, 20 and 40 eV. Data was acquired with both ESI+ and ESI− ionization. For the 
chromatographic separation a reversed-phase and HILIC chromatographies were employed resulting in a total 
of four datasets. The column applied in reversed-phase chromatography was Zorbax Eclipse XDB-C18, 1.8 µm, 
2.1 × 100 mm (Agilent Technologies, Santa Clara, CA, USA) and the gradient at 0.4 ml/min flow rate consisted 
of 0–10 min: 2 → 100% B, 10–14.5 min 100% B, 14.5–14.51 min 100 → 2% B, 14.51–16.5 min 2% B where A: H2O 
+ 0.1 % formic acid and B: MeOH + 0.1% formic acid (VWR HiPerSolv CHROMANORM, Radnor, PA, USA). 
The column applied in HILIC chromatography was Acquity UPLC® BEH Amide 1.7 µm, 2.1 × 100 mm (Waters 
Corporation, Milford, MA, USA) and gradient at 0.6 ml/min flow rate consisted of 0–2.5 min: 100% B, 2.5–10 
min: 100% B → 0% B, 10–10.01 min: 0% B → 100% B, 10.01–12.5 min: 100% B where A: ACN:H2O (1:1) and B 
(ACN:H2O (9:1)) both containing 20 mM of ammonium formate (Arcos Organics, Geel, Belgium).
MS instrument source parameters were selected as indicated; a capillary voltage of 3500C, end plate offset 
voltage of 500V, temperature of 325 °C, dry gas flow of 10 L/min and nebulizer pressure of 45 psi. Parameters 
for the full scan mode were mass range of 50–1600 m/z, abundance threshold of 150 and scan time of 1.67 Hz. 
Parameters for the MS/MS scans were mass range of 50–1600 m/z, scan time of 1.67 Hz, number of precursors 
set to 4, active exclusion enabled with exclusion after 2 spectra and release after 0.25 min. Data was stored in 
centroid mode and instrument operated with the Bruker Compass HyStar SR 5.0 software
Peak picking, alignment and data pre-processing
Before proceeding to peak picking, raw instrumental file (.d) were transformed into .mzML-format using the 
Bruker DataAnalysis 5.1 software. mzML-files were imported into MS-DIAL v 4.80 software used for peak 
picking and alignment40. Parameters for the peak picking included MS1 tolerance 0.005 Da, MS2 tolerance 
0.015 Da, m/z range 0–2000 Da, minimum peak height 8000 (serum) or 10,000 (caecum), minimum peak width 
8 scans, retention time tolerance 0.1 minutes, N% detected in at least one group 10 (serum) or 25 (caecum) and 
gap filling by compulsion activated. Peak picking and alignment was separately for each of the four datasets, 
results exported as .xlsx table and finally merged into a single .xlsx table.
Scientific Reports |        (2025) 15:15455 10| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
The resulting .xlsx table was processed with notame v. 0.2.1 R package using R v. 4.0.3 (CRAN hosted by the 
Institute for Statistics and Mathematics of Wirtschafts Universität Wien, Austria). Each mode was pre-processed 
individually including drift correction, feature quality assessment and relevant visualizations. Shortly, first zero 
values were replaced with NA and features with QC detection below 30 % were flagged. Next, log-transformation 
was applied following drift correction by regularized cubic spline regression based on the QC samples. To prevent 
overfitting, smoothing parameter ranging from 0.5 to 1.5 was applied by leave-one-out cross-validation. After 
drift correction, log-transformed features were reverted to original scale to assess quality by flagging additional 
poor features if for example presence was less than 20 % in any group or having an RSD value of over 20 %. 
Random forest imputation was applied to impute missing values first on the good-quality (unflagged) features 
and then on all features.
Metabolite identification
Features were annotated using MS-DIAL v.4.80 and Bruker DataAnalysis 5.1 softwares to compare m/z and 
retention time values and MS/MS fragmentation spectra against in-house and publicly available databases such 
as LIPIDMAPS (https://lipidmaps.org/) and METLIN41. Features were selected for annotation based on their 
group-wise significance in multivariate and univariate analysis in addition to presence of MS/MS spectrum. 
Level of identification was assessed according to the recommendations listed by Chemical Analysis Working 
Group (CAWG) Metabolomics Standards Initiative (MSI)42: 1 = identified based on a reference standard, 2 
= putatively annotated based on physicochemical properties or similarity with public spectral libraries, 3 = 
putatively annotated to a chemical class and 4 = unknown (2). Additionally, the remaining unknowns with MS/
MS spectrum were subjected for analysis by MS-FINDER v 3.5043 to characterize their molecular formula.
Statistical analysis
Data are expressed as means ± SD or medians with interquartile ranges. Unpaired t tests or Mann-Whitney 
tests were used to compare Obese-recipient versus Control-recipient mice considering the normality of the 
distribution. For metabolomic analyse, multivariate analyses, namely PCA for dimension reduction and sPLS-
DA for group discrimination, were conducted by ‘mixOmics’ R package v. 6.14.1 (58). For the sPLS-DA model, 
we used a cross-validation (CV) procedure of 10-fold CV repeated 50 times. Univariate analyses were conducted 
by ‘notame’ R package v. 0.2.139. Significant features were shortlisted using Welch’s t-test and VIP value > 1.5 
retrieved from the sPLS-DA model. All p-values were corrected using the Benjamini–Hochberg false discovery 
rate (FDR) to calculate the q-value. Correlation between the variation of bacteria (at the family level or the genus 
level) and metabolites significantly affected (Obese- versus Control-recipient mice) was assessed by Spearman’s 
correlation tests. Statistical p values < 0.05 were considered significant except for metabolomic and microbial 
abundances analyses for which only q values < 0.05 were considered statistically significant. All analyses and 
graphs were conducted with Graphpad Prism software v. 10.1.2 (San Diego, CA, USA; www.graphpad.com) 
except for the gut microbiota analyse and metabolomic analyse where we used R version 4.2.1 (CRAN hosted by 
the Institute for Statistics and Mathematics of Wirtschafts Universität Wien, Austria).
Data availability
Raw 16S rRNA gene sequences can be found in the SRA database (project ID: PRJNA1217873) and will be made 
public upon acceptance (person to contact Audrey Neyrinck: audrey.neyrinck@uclouvain.be). All data will be 
available from the corresponding author on reasonable request.
Received: 30 September 2024; Accepted: 16 April 2025
References
 1. Kootte, R. S. et al. Improvement of insulin sensitivity after lean donor feces in metabolic syndrome is driven by baseline intestinal 
microbiota composition. Cell Metab. 26, 611–619. https://doi.org/10.1016/j.cmet.2017.09.008 (2017).
 2. de Groot, P. F., Frissen, M. N., de Clercq, N. C. & Nieuwdorp, M. Fecal microbiota transplantation in metabolic syndrome: History, 
present and future. Gut Microbes 8, 253–267. https://doi.org/10.1080/19490976.2017.1293224 (2017).
 3. Turnbaugh, P. J. et al. The effect of diet on the human gut microbiome: A metagenomic analysis in humanized gnotobiotic mice. 
Sci. Transl. Med. 1, 6–14. https://doi.org/10.1126/scitranslmed.3000322 (2009).
 4. Rodriguez, J. et al. Discovery of the gut microbial signature driving the efficacy of prebiotic intervention in obese patients. Gut 69, 
1975–1987. https://doi.org/10.1136/gutjnl-2019-319726 (2020).
 5. Clark, T. D., Reichelt, A. C., Ghosh-Swaby, O., Simpson, S. J. & Crean, A. J. Nutrition, anxiety and hormones. Why sex differences 
matter in the link between obesity and behavior. Physiol. Behav. 247, 113713. https://doi.org/10.1016/j.physbeh.2022.113713 
(2022).
 6. Donoso, F., Cryan, J. F., Olavarría-Ramírez, L., Nolan, Y. M. & Clarke, G. Inflammation, lifestyle factors, and the microbiome-gut-
brain axis: Relevance to depression and antidepressant action. Clin. Pharmacol. Ther. 113, 246–259.  h t t p s : / / d o i . o r g / 1 0 . 1 0 0 2 / c p t . 2 5 
8 1 (2023).
 7. Cryan, J. F. et al. The microbiota-gut-brain axis. Physiol. Rev. 99, 1877–2013. https://doi.org/10.1152/physrev.00018.2018 (2019).
 8. Ahmed, H. et al. Microbiota-derived metabolites as drivers of gut-brain communication. Gut Microbes 14, 2102878.  h t t p s : / / d o i . o r 
g / 1 0 . 1 0 8 0 / 1 9 4 9 0 9 7 6 . 2 0 2 2 . 2 1 0 2 8 7 8 (2022).
 9. Idle, J. R. & Gonzalez, F. J. Metabolomics. Cell Metab. 6, 348–351. https://doi.org/10.1016/j.cmet.2007.10.005 (2007).
 10. Puljiz, Z. et al. Obesity, gut microbiota, and metabolome: From pathophysiology to nutritional interventions. Nutrients 15, 236. 
https://doi.org/10.3390/nu15102236 (2023).
 11. Leyrolle, Q. et al. Microbiota and metabolite profiling as markers of mood disorders: A cross-sectional study in obese patients. 
Nutrients 14, 147. https://doi.org/10.3390/nu14010147 (2021).
 12. Leclercq, S. et al. Gut microbiota-induced changes in β-hydroxybutyrate metabolism are linked to altered sociability and depression 
in alcohol use disorder. Cell Rep. 33, 108238. https://doi.org/10.1016/j.celrep.2020.108238 (2020).
Scientific Reports |        (2025) 15:15455 11| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
 13. Turnbaugh, P. J., Bäckhed, F., Fulton, L. & Gordon, J. I. Diet-induced obesity is linked to marked but reversible alterations in the 
mouse distal gut microbiome. Cell Host Microbe 3, 213–223. https://doi.org/10.1016/j.chom.2008.02.015 (2008).
 14. Turnbaugh, P. J. et al. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 444, 1027–1031. 
https://doi.org/10.1038/nature05414 (2006).
 15. Rodriguez, D. M., Benninghoff, A. D., Aardema, N. D. J., Phatak, S. & Hintze, K. J. Basal diet determined long-term composition 
of the gut microbiome and mouse phenotype to a greater extent than fecal microbiome transfer from lean or obese human donors. 
Nutrients 11, 630. https://doi.org/10.3390/nu11071630 (2019).
 16. Bhat, K. P., Ümit Kaniskan, H., Jin, J. & Gozani, O. Epigenetics and beyond: Targeting writers of protein lysine methylation to treat 
disease. Nat. Rev. Drug Discov. 20, 265–286. https://doi.org/10.1038/s41573-020-00108-x (2021).
 17. Wang, Y. et al. Role of indole-3-acetic acid in NAFLD amelioration after sleeve gastrectomy. Obes. Surg. 31, 3040–3052.  h t t p s : / / d o 
i . o r g / 1 0 . 1 0 0 7 / s 1 1 6 9 5 - 0 2 1 - 0 5 3 2 1 - 0 (2021).
 18. Liu, M., Nieuwdorp, M., de Vos, W. M. & Rampanelli, E. Microbial tryptophan metabolism tunes host immunity, metabolism, and 
extraintestinal disorders. Metabolites 12, 834. https://doi.org/10.3390/metabo12090834 (2022).
 19. Wishart, D. S. et al. MiMeDB: The human microbial metabolome database. Nucleic Acids Res. 51, D611–D620.  h t t p s : / / d o i . o r g / 1 0 . 
1 0 9 3 / n a r / g k a c 8 6 8 (2023).
 20. Russell, W. R. et al. Major phenylpropanoid-derived metabolites in the human gut can arise from microbial fermentation of 
protein. Mol. Nutr. Food Res. 57, 523–535. https://doi.org/10.1002/mnfr.201200594 (2013).
 21. Su, X., Gao, Y. & Yang, R. Gut microbiota-derived tryptophan metabolites maintain gut and systemic homeostasis. Cells 11, 96. 
https://doi.org/10.3390/cells11152296 (2022).
 22. Qu, X. et al. Indole-3-acetic acid ameliorates dextran sulfate sodium-induced colitis via the ERK signaling pathway. Arch. Pharm. 
Res. 47, 288–299. https://doi.org/10.1007/s12272-024-01488-z (2024).
 23. Karu, N. et al. Tryptophan metabolism, its relation to inflammation and stress markers and association with psychological and 
cognitive functioning: Tasmanian Chronic Kidney Disease pilot study. BMC Nephrol. 17, 171.  h t t p s : / / d o i . o r g / 1 0 . 1 1 8 6 / s 1 2 8 8 2 - 0 1 
6 - 0 3 8 7 - 3 (2016).
 24. Anderson, G. M., Gerner, R. H., Cohen, D. J. & Fairbanks, L. Central tryptamine turnover in depression, schizophrenia, and 
anorexia: Measurement of indoleacetic acid in cerebrospinal fluid. Biol. Psychiatry 19, 1427–1435 (1984).
 25. Zheng, S. et al. Urinary metabonomic study on biochemical changes in chronic unpredictable mild stress model of depression. 
Clin. Chim. Acta 411, 204–209. https://doi.org/10.1016/j.cca.2009.11.003 (2010).
 26. Chen, Y. et al. Indole acetic acid exerts anti-depressive effects on an animal model of chronic mild stress. Nutrients 14, 19.  h t t p s : / / 
d o i . o r g / 1 0 . 3 3 9 0 / n u 1 4 2 3 5 0 1 9 (2022).
 27. Suntornsaratoon, P. et al. Lactobacillus rhamnosus GG stimulates dietary tryptophan-dependent production of barrier-protecting 
methylnicotinamide. Cell Mol. Gastroenterol. Hepatol. 18, 101346. https://doi.org/10.1016/j.jcmgh.2024.04.003 (2024).
 28. Nejabati, H. R. et al. N1-methylnicotinamide: Is it time to consider it as a dietary supplement for athletes? Curr. Pharm. Des. 28, 
800–805. https://doi.org/10.2174/1381612828666220211151204 (2022).
 29. Zhang, J., Chen, Y., Liu, C., Li, L. & Li, P. N(1)-methylnicotinamide improves hepatic insulin sensitivity via activation of SIRT1 and 
inhibition of FOXO1 acetylation. J. Diabetes Res. 2020, 1080152. https://doi.org/10.1155/2020/1080152 (2020).
 30. Liu, J. R. et al. Roles of nicotinamide N-methyltransferase in obesity and type 2 diabetes. Biomed. Res. Int. 2021, 9924314.  h t t p s : / / 
d o i . o r g / 1 0 . 1 1 5 5 / 2 0 2 1 / 9 9 2 4 3 1 4 (2021).
 31. Hiel, S. et al. Link between gut microbiota and health outcomes in inulin-treated obese patients: Lessons from the Food4Gut 
multicenter randomized placebo-controlled trial. Clin. Nutr. https://doi.org/10.1016/j.clnu.2020.04.005 (2020).
 32. Leclercq, S. et al. Blood metabolomic profiling reveals new targets in the management of psychological symptoms associated with 
severe alcohol use disorder. Elife 13, 937. https://doi.org/10.7554/eLife.96937 (2024).
 33. Sidor, M. M., Rilett, K. & Foster, J. A. Validation of an automated system for measuring anxiety-related behaviours in the elevated 
plus maze. J. Neurosci. Methods 188, 7–13. https://doi.org/10.1016/j.jneumeth.2010.01.021 (2010).
 34. Bourin, M. & Hascoët, M. The mouse light/dark box test. Eur. J. Pharmacol. 463, 55–65.  h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / S 0 0 1 4 - 2 9 9 9 ( 0 3 ) 0 
1 2 7 4 - 3 (2003).
 35. Can, A. et al. The mouse forced swim test. J. Vis. Exp. 1, e3638. https://doi.org/10.3791/3638 (2012).
 36. Thibaut, M. M. et al. The microbiota-derived bile acid taurodeoxycholic acid improves hepatic cholesterol levels in mice with 
cancer cachexia. Gut Microbes 17, 2449586. https://doi.org/10.1080/19490976.2025.2449586 (2025).
 37. Mirzayi, C. et al. Reporting guidelines for human microbiome research: the STORMS checklist. Nat. Med. 27, 1885–1892.  h t t p s : / / 
d o i . o r g / 1 0 . 1 0 3 8 / s 4 1 5 9 1 - 0 2 1 - 0 1 5 5 2 - x (2021).
 38. Costea, P. I. et al. Towards standards for human fecal sample processing in metagenomic studies. Nat. Biotechnol. 35, 1069–1076. 
https://doi.org/10.1038/nbt.3960 (2017).
 39. Klåvus, A. et al. “notame”: Workflow for non-targeted LC-MS metabolic profiling. Metabolites 10, 135.  h t t p s : / / d o i . o r g / 1 0 . 3 3 9 0 / m e 
t a b o 1 0 0 4 0 1 3 5 (2020).
 40. Tsugawa, H. et al. MS-DIAL: Data-independent MS/MS deconvolution for comprehensive metabolome analysis. Nat. Methods 12, 
523–526. https://doi.org/10.1038/nmeth.3393 (2015).
 41. Smith, C. A. et al. METLIN: A metabolite mass spectral database. Ther. Drug Monit. 27, 747–751.  h t t p s : / / d o i . o r g / 1 0 . 1 0 9 7 / 0 1 . ft  d . 0 
0 0 0 1 7 9 8 4 5 . 5 3 2 1 3 . 3 9 (2005).
 42. Sumner, L. W. et al. Proposed minimum reporting standards for chemical analysis Chemical Analysis Working Group (CAWG) 
metabolomics standards initiative (MSI). Metabolomics 3, 211–221. https://doi.org/10.1007/s11306-007-0082-2 (2007).
 43. Tsugawa, H. et al. Hydrogen rearrangement rules: Computational MS/MS fragmentation and structure elucidation using MS-
FINDER software. Anal. Chem. 88, 7946–7958. https://doi.org/10.1021/acs.analchem.6b00770 (2016).
Acknowledgements
We thank prof. Thissen and prof. de Timary to give access to stool sample collection from FOOD4GUT study 
and GUT2BRAIN study, respectively. We thank A. Flandin and B. Es Saadi for their excellent technical assis-
tance. NMD is a recipient of grants from the Fonds de la Recherche Scientifique (FRS)‐FNRS (Grant Numbers: 
PINT‐MULTI R.8013.19 (NEURON, call 2019) and PDR T.0068.19). LBB is a Collen-Francqui Research Pro-
fessor and grateful for the support of the Francqui Fondation. PDC is honorary research director at FRS-FNRS 
(Fonds de la Recherche Scientifique) and the recipient of grants from FNRS (Projet de Recherche PDR-conven-
tion: FNRS T.0030.21, FRFS-WELBIO: WELBIO-CR-2022A-02P, EOS: program no. 40007505) KH is a recipient 
of grants from the Research Council of Finland (grant numbers 321716 and 334814), and Jane and Aatos Erkko 
Foundation.
Author contributions
N.M.D., A.M.N. and Q.L. conceived the experiment(s). Q.L. conducted the experiment(s). H.A., T.M., O.K. and 
K.H. performed the analyse of the metabolomics data.  A.M.N., H.A., Q.L., C.A. and L.B.B performed data ana-
Scientific Reports |        (2025) 15:15455 12| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/
lyse. A.M.N. and N.M.D. wrote the paper. S.Le, S.La., P.D.C., O.K., L.B.B. and K.H. provided intellectual input on 
the paper and reviewed the paper. N.M.D. planned and supervised all experiments and manuscript preparation. 
All authors reviewed the manuscript.
Declarations
Competing interests
OK and KH are founders of Afekta Technologies Ltd. PDC is inventor on patent applications dealing with 
the use of gut bacteria and their components in the treatment of diseases. PDC was co-founder of The 
Akkermansia Company and Enterosys. The other authors report no financial interests or potential conflicts of 
interest.
Additional information
Supplementary Information The online version contains supplementary material available at  h t t p s : / / d o i . o r g / 1 
0 . 1 0 3 8 / s 4 1 5 9 8 - 0 2 5 - 9 9 0 4 7 - z . 
Correspondence and requests for materials should be addressed to N.M.D.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and 
institutional affiliations.
Open Access  This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 
4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in 
any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide 
a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have 
permission under this licence to share adapted material derived from this article or parts of it. The images or 
other third party material in this article are included in the article’s Creative Commons licence, unless indicated 
otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence 
and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to 
obtain permission directly from the copyright holder. To view a copy of this licence, visit  h t t p : / / c r e a t i v e c o m m o 
n s . o r g / l i c e n s e s / b y - n c - n d / 4 . 0 / . 
© The Author(s) 2025 
Scientific Reports |        (2025) 15:15455 13| https://doi.org/10.1038/s41598-025-99047-z
www.nature.com/scientificreports/