Dietary supplementation with probiotics promotes weight loss by reshaping the gut microbiome and energy metabolism in obese dogs

C. elegans lifespan analysis was utilized to evaluate the probiotic capability of E. faecium IDCC 2102 and B. lactis IDCC 4301. C. elegans is a well-established genetic model with the advantage of relatively simple genetic editing and a shorter life expectancy than other animal models (28). Both E. faecium IDCC 2102 and B. lactis IDCC 4301 extended the longevity of C. elegans in comparison to Escherichia coli OP50 (OP50), and IDCC 2102 elevated the longevity even more than Lacticaseibacillus rhamnosus GG (LGG), one of the most clinically renowned probiotics worldwide (29) (Fig. 1a). Next, the lipogenic inhibitory capacity of E. faecium IDCC 2102 and B. lactis IDCC 4301 was evaluated by lipid staining with the hyperlipidemic C. elegans VS29 strain. The lipid content was determined by oil red O and Nile red staining; E. faecium IDCC 2102 and B. lactis IDCC 4301 treatment significantly decreased the lipid accumulation of VS29 stains (Fig. 1b and c).

Furthermore, to assess the efficacy of the probiotic intervention in reducing lipid accumulation in the hyperlipidemic C. elegans model, a metabolite analysis was conducted. Significantly, the hyperlipidemic control group exhibited elevated levels of L-serine, L-alanine, and hydroxyproline, key components of glycine. Normally, glycine plays a crucial role in the regulation of food consumption and glucose homeostasis by promoting insulin and GSH (glutathione) production. However, excess amino acids are known to be converted into fat and stored in fat depots; alternatively, they can be utilized for glucose synthesis through gluconeogenesis. In the context of C. elegans, an excess of glycine proves detrimental to its longevity, as the organism heavily relies on a tightly regulated insulin/IGF-1 signaling pathway (30–32). Consequently, it appears that E. faecium IDCC 2102 and B. lactis IDCC 4301 may enhance the longevity of C. elegans by modulating glycine synthesis (Fig. 1a; Fig. S1b and c).

The hyperlipidemic C. elegans strain VS29 (cont) displayed elevated levels of factors associated with fatty acid metabolism, such as L-gluconolactone and oxalic acid, alongside reduced levels of L-tryptophan. Additionally, the accumulation of glycine in tissues, particularly the brain, can disrupt normal brain functions. Glycine serves as an inhibitory neurotransmitter capable of suppressing dopamine (31, 33, 34). This pattern was reversed following treatment with E. faecium IDCC 2102 and B. lactis IDCC 4301, as they increased the levels of beneficial bacterial short-chain acids such as 5-aminovaleric acid and acetic acid, as well as the serotonin precursor L-tryptophan. Simultaneously, they decreased the levels of glycine and its precursor molecules. This outcome suggests that probiotic treatment indeed proves effective in the hyperlipidemic C. elegans model, exerting a regulatory effect on lipid accumulation and neurotoxicity (Fig. S1).

Canines were fed a high-fat diet to induce obesity for 9 weeks, during which E. faecium IDCC 2102 and B. lactis IDCC 4301 were administered concurrently (10¹⁰ CFU/day) to determine the anti-obesity impact of these two probiotics. As anticipated, the body weight of the high-fat diet (HFD) group without probiotics increased and, beginning in week 6, was considerably higher than that of the control group (cont) fed normal chow. At the end of the experiment, the HFD group weighed 127.8% ± 10.2% more than the control group. However, probiotic treatments altered the features. Even with a high-fat diet, probiotic treatments significantly reduced the body weight gain ratio. Compared to the HFD group, E. faecium IDCC 2102 and B. lactis IDCC 4301 significantly decreased the final body weight by 17.3% ± 6.0% and 15.4% ± 4.3%, respectively (Fig. 2a and b). In addition, the body condition score based on body shape exhibited significant lipid accumulation in the HFD group and improvement in the probiotic-treated groups (Fig. 2d; Fig. S2). Given that the amount of food consumed did not differ significantly regardless of probiotic therapy (Fig. 2c), probiotic treatment could prevent lipid accumulation.

The serum was then subjected to DRI-CHEM analysis to determine the various lipid levels standardized by individual body weight, as individual body weight varied. TCHO was increased by 4.7% ± 12.3% in the HFD group compared to the control, and E. faecium IDCC 2102 and B. lactis IDCC 4301 decreased the TCHO level by 26.1% ± 10.5% and 24.9% ± 13.6% compared to the HFD group. TG, a store lipid, was increased by 50.2% ± 17.9% in the HFD group compared to the control group, whereas E. faecium IDCC 2102 and B. lactis IDCC 4301 treatment reduced TG levels by 31.5% ± 26.9% and 10.4% ± 25.7%, respectively. HDL-C, which inhibits cholesterol biosynthesis, was decreased by 35.4% ± 17.0% in the HFD group and increased by 71.7% ± 47.6% and 77.4% ± 356.2% in the E. faecium IDCC 2102 and B. lactis IDCC 4301 groups, respectively. In the same context, the ratio of LDL-C, which transfers cholesterol to the periphery, increased by 218.9% ± 82.6% in the HFD group compared to the control group, while it decreased by 60.5% ± 21.1% and 72.3% ± 22.6% in the E. faecium IDCC 2102 and B. lactis IDCC 4301 groups.

Lipid accumulation could induce sensitization of the immune system, which could lead to chronic inflammation in the long term. Therefore, inflammatory markers were analyzed in serum and whole blood samples. To investigate the systemic inflammatory marker, aspartate aminotransferase (AST) was normalized by alanine aminotransferase (ALT); the AST/ALT level was significantly increased in the HFD group by 40.9% ± 10.3% compared to the normal chow control group, and E. faecium IDCC 2102 and B. lactis IDCC 4301 treatment reduced the AST/ALT level by 40.8% ± 6.0% and 27.8% ± 11.6%, respectively (Fig. 3a).

The term CBC stands for complete blood counting, and it is a means of evaluating general health for various illness conditions, such as infection, by validating the amount of active cells in the blood. WBC count and PLT count elevations represent immune disorders and an acute inflammatory state. During the experiment, CBC tests were taken with blood for 9 weeks. The results of these measurements showed that all of the test subjects were within the normal range. Nevertheless, when compared to the control group, the WBC level increased by 39.9% ± 37.3% in the HFD group and reduced by 22.9% ± 10.8% in the IDCC 2102 and 23.3% ± 10.1% in the IDCC 4301 treatment group. Specifically, the IDCC 4301 treatment group confirmed that the PLT count inflammatory factor was reduced by 28.3% ± 36.3% in comparison to the HFD group (Fig. 3b).

To ensure that the probiotics employed in this trial had an anti-inflammatory effect, serum inflammatory markers were identified using an ELISA. Treatment with E. faecium IDCC 2102 and B. lactis IDCC 4301 changed the immune response by lowering the levels of the pro-inflammatory adipokines TNF-α and IL-1β by a large amount. They also lowered the level of IL-6 by a large amount, which helped restore the hormonal balance needed for adipocyte differentiation and lipid metabolism (Fig. 3c). Furthermore, the administration of probiotics significantly improved insulin sensitivity as well as the levels of leptin and adiponectin, which are involved in the suppression of fat storage by balancing the intake/consumption of energy in the body as well as hunger regulation (35) (Fig. 3d).

A metagenomic analysis of canine feces from each group was performed to examine the microbial shift after probiotic administration. NGS (next generation sequencing) analysis was used to sequence the V4 region of 16S rRNA. A total of 9,604 OTUs were examined, and 301 of them were identified as bacteria with >97% identification reliability using the NCBI database. To assess the diversity among groups, alpha diversity was analyzed; the Chao index and Shannon index measure the number of species and effective number of species, respectively. Although there was no significant difference between groups in the Chao index, B. lactis IDCC 4301 demonstrated high variety when compared to the HFD group in the Shannon index (Fig. 4a). In addition, weighted and unweighted beta-diversity analyses revealed that each probiotic treatment resulted in unique microbiota composition clustering (Fig. 4b).

As observed earlier, it is evident that diet can influence microbial composition. In contrast to the regular chow control group, the HFD-induced obese beagles experienced a reduction of 3.9% in their genus diversity, namely, Enterococcus, Bacteroides, and Lactobacillus. On the other hand, E. faecium IDCC 2102 treatment restored Enterococcus, Fusobacterium, and Bacteroides genus levels, and B. lactis IDCC 4301 treatment increased Fusobacterium, Bacteroides, and Lactobacillus at the genus level (Fig. 4c). The microbial changes at the genus level were reflected at the family level. E. faecium IDCC 2102 treatment particularly increased the composition of Lachnospiraceae, Enterobacteriaceae, and Lactobacillaceae, and B. lactis IDCC 4301 treatment specifically increased Ruminococcaceae, S24-7, and Lactobacillaceae at the family level compared to the HFD-induced obese canines (Fig. 4d). Moreover, the Firmicutes/Bacteroides ratio was exceptionally high in the HFD group, which was not surprising given that prior research has suggested this ratio as a potential biomarker of obesity (Fig. 4e).

As the composition of the microbiome varied based on the diet of each group, it was discovered that the bacterial metabolites were also quite distinct. Obese canines produced higher levels of histamine, malonic acid, picolinic acid, methylsuccinic acid, and 5-hydroxy-L-tryptophan levels than the normal chow control group, although tyramine synthesis was decreased. Interestingly, treatment with E. faecium IDCC 2102 and B. lactis IDCC 4301 resulted in the majority of the elevated metabolites in the HFD group being downregulated. Tyramine, alpha-tocopherol, tryptamine, D-proline, and stearic acid were all upregulated in the E. faecium IDCC 2012 group, while methylsuccinic acid, picolinic acid, 5-hydroxy-L-tryptophan, acetic acid, butyric acid, L-valine, and histamine were all downregulated. B. lactis IDCC 4301 treatment specifically decreased malonic acid and phenethylamine while elevating isocaproic acid, adipic acid, succinic acid, and beta-alanine (Fig. 5a).

Since bacterial metabolites are thought to be strongly related to the microbiome structure, we examined the interaction between microbiota and metabolites generated by each probiotic treatment by normalizing the metabolite levels by microbiome composition. Treatment with E. faecium IDCC 2102 stimulated the colonization of lactic acid bacteria, such as Ruminococcus, Lactobacillus, and S24-7, which led to a general increase in amino acids and short-chain fatty acids, such as oxalic acid, phosphoric acid, and isocaproic acid. In particular, the Enterococcus genus colonized and significantly elevated dopamine, valeric acid, acetic acid, and glutaric acid levels (Fig. 5b and c, Fig. 6a and b). Treatment with E. faecium IDCC 2102 increased the proportion of Firmicutes, Bacteroides, Fusobacteria, and Proteobacteria, resulting in the proportional synthesis of fatty acids, phenols, organic acids, and indoles (Fig. 6c and d).

B. lactis IDCC 4301 exhibited somewhat different metabolomics results. B. lactis IDCC 4301 treatment also increased the colonization of general lactic acid bacteria and increased the general amino acids, but the colonization of Bifidobacterium increased the short- and medium-chain fatty acids such as isocaproic acid and amino acids such as adipic acid, succinic acid, and beta-alanine more dramatically (Fig. 5c and d, Fig. 7a and b). Treatment with B. lactis IDCC4301 increased the proportion of Firmicutes, Proteobacteria, Bacteroidetes, Actinobacteria, and Fusobacteria, resulting in the proportional synthesis of fatty acids, organic acids, indoles, and phenols (Fig. 7c and d).

Obesity induction and probiotic treatments modified the metabolic pathways. The HFD obese canine group significantly increased the functional and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways of aminoacyl-tRNA, fatty acid biosynthesis (valine, leucine, isoleucine, pantothenate, and CoA), and amino acid metabolism (arginine, proline, and tryptophan) (Fig. 8a and c). Probiotic treatments with E. faecium IDCC 2102 and B. lactis IDCC 4301 enhanced pyruvate metabolism, glucose metabolism (glycolysis and gluconeogenesis), energy production (glyoxylate and dicarboxylate metabolism), and amino acid metabolism (arginine and proline) (Fig. 8b and c).

Ildong Bioscience Co., Ltd. (Gyeonggi-do, Republic of Korea) provided the probiotics for this experiment, Enterococcus faecium IDCC 2102 (IDCC 2102) and Bifidobacterium lactis IDCC 4301 (IDCC 4301), and they were cultured in de Man, Rogosa, and Sharpe (MRS; BD Difco, New Jersey, USA) at 37°C for 48 h under aerobic conditions. Lacticaseibacillus rhamnosus GG (LGG) was cultured in MRS at 37°C for 24 h under aerobic conditions. The bacterial cultures were freeze-dried and kept at −80°C before use.

C. elegans fer15(b26)II;fem-1(hc17)IV and C. elegans strain VS29 (vha-6p::GFP::dgat-2) were purchased from the Caenorhabditis Genetic Center (Minnesota, USA) and maintained on nematode growth medium (NGM) plates at 15°C. E. coli strain OP50 (OP50), the standard feed for C. elegans, was cultured in the Luria–Bertani medium (LB Broth, Miller; BD Difco) at 37°C for 24 h with 225 rpm shaking. The live bacterial lawn was prepared for C. elegans feeding by centrifuging the bacterial pellet at 13,000 rpm for 1 minute, washing it twice with sterile M9 buffer (3 g KH₂PO₄, 6 g Na₂HPO₄, and 5 g NaCl mixed in 1 L distilled water, autoclaved), and then adding 1 mL of 1 M MgSO₄ (Sigma‒Aldrich, St. Louis, MO, USA). The bacterial pellet was then concentrated to a final concentration of 2.5 mg/L (wet weight) in M9 buffer and suspended on NGM plates [3.5 g Bacto Peptone (BD Difco), 3 g NaCl (Sigma‒Aldrich), and 20 g agar (BD Difco)] (11).

For the lifespan analysis, egg-bearing worms were bleached using a sodium hypochlorite‒sodium hydroxide solution (Sigma‒Aldrich), and the newborn worms were synchronized to L1 stage worms on NGM plates at 25°C (57). After 3 days, young adult L4 stage worms were plated on 35-mm-diameter NGM plates seeded with OP50, IDCC 2102, IDCC 4301, and LGG. The experiment was carried out in triplicate, and all worms were transplanted to new bacterial lawns on a daily basis until all worms died.

To quantify the lipid accumulation of C. elegans, featured staining methods were applied to oil red O and Nile red O (58, 59). First, C. elegans strain VS29 was synchronized to the young adult L4 stage as described above and exposed to E. coli OP50, IDCC 2102, and IDCC 4301 for 48 h. The staining method followed bioprotocols using oil red O and Nile red solutions (Sigma‒Aldrich). The result was observed using an Olympus IX53 (Olympus Life Science, Tokyo, Japan), and the intensity was analyzed using i-Solution (IMT i-Solution Inc, New York, USA).

Beagles aged 2–4 years and weighing 9.0 ± 2.14 kg were purchased from Orient Bio (Gyeonggi-do, Republic of Korea) for the experiment (8 males and 12 females). They were given 2 weeks to acclimate to the breeding room atmosphere before the experiment began ad libitum. After acclimation, all subjects were randomly separated into four groups of five individuals based on body weight (9.24 ± 2.17 kg; n = 3 per group). The control group (cont) received a standard diet consisting of medium adult dry dog food (Royal Canin, Gard, France; ME (metabolizable energy) = 357.0 kcal/100 g), while the other groups were induced to develop obesity with a 22% high-fat diet (HFD; Jeilfeed, Seoul, Republic of Korea; ME (metabolizable energy) = 483.0 kcal/100 g). The probiotic groups were administered with 10¹⁰ CFU/day of IDCC 2102 and 10¹⁰ CFU/day of IDCC 4301 for 9 weeks, mixed with the high-fat diet. The nutritional compositions of normal chow and HFD are stated in Table S1. Each experimental animal was bred in a separate cage with an identity card during the experiment, including the acclimation period. The breeding room was maintained at a constant temperature (21°C ± 2°C), humidity (50% ± 20%), and photoperiod of 12 h (08:00–20:00). This animal experiment was conducted with the approval of the Institutional Animal Care and Use Committee at Chungnam National University (202109A-CNU-149).

The body weight was obtained once a week following the commencement of the test at a specific time to observe the change in body weight during the experiment. To assess feed intake, the provided quantity of feed was measured daily at a specific time, and the residual amount was gauged 24 h later consistently throughout the experiment.

To monitor changes in the body shape of the beagle animal model as a result of a high-fat diet and probiotic feeding, the width of the abdomen and hips was measured once a week to score the body condition score, which is an index for determining the body shape of companion animals, and the score during the test period. Table S2 provides the body condition score index. The higher the degree of obesity, the higher the score, and the requirements for each score component are stated in Table S2 (60). The average value between the groups was calculated.

At the end of the 9-week experiment, all animals were fasted for 14 h, blood was collected and centrifuged (4°C, 3,000 rpm, 20 minutes), and serum was separated to analyze various physiological changes. The separated serum was kept at −70°C, and the experiment was carried out using ELISA. TNF-α, IL-1β, IL-6 (R&D Systems, Minnesota, USA), and C-reactive protein (Cusabio Technology LLC, Texas, USA) concentrations in the blood were analyzed to assess changes in inflammation-related cytokines. For the examination of energy metabolism regulators, the concentrations of leptin (Millipore, USA), adiponectin (ABclonal Technology, Massachusetts, USA), and malondialdehyde (Biocompare, California, USA) in the blood were measured according to the manufacturer’s instructions. Serum ALT, AST, HDL-C, TG, and TCHO were analyzed using a DRI-CHEM Analyzer (Fujifilm, Tokyo, Japan).

After 9 weeks of experimentation, fecal samples were obtained from each subject, frozen at −70°C, and then used for the experiment to examine the intestinal microbial community’s response to feed. The first step was to extract the gDNA from each sample in accordance with the DNeasy PowerSoil Pro Kit’s manufacturer’s instructions (Qiagen, Hilden, Germany). Next, the V4 segment of the 16S ribosomal RNA gene was amplified using the extracted DNA sample, and Illumina iSeq 100 (Illumina, Inc., California, USA) was used for next-generation sequencing (61). The following is the primer sequence for the synthesis of the V4 region: 515F, TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTGCCAGCMGCCGCGGTAA; 806R, GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGGACTACHVGGGTWTCTAAT.

The pooled fecal samples of canines and C. elegans VS29 strains exposed to probiotics were weighed and diluted in methanol to a final concentration of 20 mg/mL on ice. To ensure significance, the metabolite analysis was repeated three times. The supernatant was filtered using a 0.2-µm pore size polyvinylidene fluoride syringe filter after centrifugation at 15,000 × g for 5 minutes at 4°C. Two hundred-microliter aliquots of the filtered supernatant were concentrated in a vacuum concentrator and kept at −81°C prior to derivatization and gas chromatography mass spectrometry (GC-MS) analysis. The extract was derivatized at 30°C for 90 minutes with 30 µL of 20 mg/mL methoxyamine hydrochloride in pyridine (Sigma‒Aldrich), followed by 50 µL of N,O-bis(trimethylsilyl)trifluoroacetamide (Sigma‒Aldrich). As an internal standard, fluoranthene was added to the extract. A Thermo Trace 1310 GC (Waltham, MA, USA) linked to a Thermo ISQ LT single quadrupole mass spectrometer was used for the GC-MS analysis (Waltham, MA, USA). A 60-µm long DB-5MS column with an inner diameter of 0.2 mm was used. The derivatives were separated using a 0.25 m film thickness (Agilent, Santa Clara, CA, USA). The sample was injected at 300°C with a split ratio of 1:60, and the helium split flow was 90 mL/min for analysis. The metabolites were separated using 1.5 mL of continuous flow helium and an oven ramp of 50°C (2 minute hold) to 180°C (8 minute hold) at 5°C/min, 210°C at 2.5°C/min, and 325°C (10 minute hold) at 5°C/min. The mass spectra peaked at a rate of 5 spectra/s throughout the scan in the range of 35–650 m/z. The ionization mode was exposed to electron impact with the ion source temperature adjusted to 270°C. Thermo Xcalibur software was used to analyze the spectral peaks, and the metabolites were identified by matching the mass spectra and retention indices of the NIST Mass spectral search tool (version 2.0, Gaithersburg, MD, USA) (62, 63).

The Shannon and Chao indices of bacterial alpha diversity were analyzed using a non-parametric one-way analysis of variance (Kruskal–Wallis test) and Tukey’s post hoc analysis if a significant difference (P < 0.05) was found. Welch’s t-test was used to compare the differences in the relative abundance of bacterial composition (64, 65). The correlation between the microbiota and metabolites was conducted using the M²IA (66). Microbiome–metadata correlation analyses and analyses of the different levels of microbial communities by genus were performed. A network was used to examine the relationship between the functional metabolites and the bacteria. The pink nodes are the functional metabolites, and the other nodes of bacteria were grouped by color (functional false discovery rate <0.05). To discover functional pathways in the fecal microbiome, PICRUSt and KEGG (level 2) were utilized to predict the existence of functional genes in the sample. PICRUSt was completed via an online application (67, 68). Using STAMP v0.2.1.3 and an extended error bar plot, the difference in the relative abundance of bacteria and significantly different KEGG pathways within groups was calculated.

A volcano plot and principal component analysis (PCA) of metabolomics data were processed using GraphPad Prism 7.0.4 (GraphPad Software, California, USA), and heatmap analysis was processed using MetaboAnalyst 5.0 (69).

All data points used for this study were analyzed in triplicate, and the data are expressed as the mean ± standard deviation. Significant differences were determined according to Student’s t-test and one-way ANOVA using GraphPad Prism 7.0.4 (GraphPad Software) and SigmaPlot 13 (Systat Software, California, USA) followed by Tukey’s post hoc test. C. elegans survival analysis was performed using the Kaplan–Meier method, and the significance of differences between survival curves was determined using the log-rank test (STATA6; STATA, College Station, TX, USA).