Baseline data

Between November 2023 and May 2024, 72 participants were randomly assigned to the probiotic or the placebo group. After excluding four participants who took concomitantly contraindicated drugs, one with poor compliance, and one withdrawal the consent, the per-protocol set included 33 participants in the BL21 group and 33 participants in the placebo group. The baseline characteristics of participants are presented in Table 1. There were no statistically significant differences in age, gender, or BMI between the two groups (P > 0.05).

Table 1 Baseline characteristics of participantsThe effect of probiotic intervention on body weight and lipid profiles of the participants

After the 8-week treatment period, we compared the effects on participants’ body weight between the BL21 group and the placebo group. As shown in Table 2, both the BL21 group (P = 0.02) and placebo group (P = 0.01) demonstrated significant body weight reductions, with no statistically significant intergroup difference observed (P = 0.81). We also compared the lipid profiles between the BL21 group and the placebo group (Table 2). Notably, it was found that the triglyceride (TG) significantly decreased in BL21 group after the treatment (P = 0.04). Meanwhile, neither the probiotic BL21 nor the placebo showed a significant effect on the total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) at week 8.

Table 2 The effect of BL21 on the weight and lipid profilesThe diversity of the gut microbiota changed during probiotic intervention

A total of 130 samples were successfully sequenced except two sample failed. A total of 7,517,102 raw reads were obtained from high-throughput sequencing and the number of raw reads per sample ranges from 40,553 to 74,391. To minimize the effects of sequencing depth on microbiota analysis, the sequences from each sample were rarefied to 27,697, with an average Good’s coverage above 99.7%. The α-diversity of the microbiota is shown in Table S1. There was no change in index of α- diversity during the intervention process in both groups, including Shannon, Chao, Ace and Simpson index. The β-diversity of both groups were similar before intervention process (R = 0.02, P = 0.43) (Fig. 2A). After 8 weeks probiotic intervention, the β-diversity of BL21 group was significantly different from placebo group (R = 0.04, P = 0.04) (Fig. 2B). However, there was no significant change in β-diversity of the BL21 group and the placebo group before and after intervention (Fig. 2C and D).

Fig. 2

The β-diversity of the gut microbiota changed during probiotic intervention. PCoA, principal coordinate analysis; Pro, probiotic group; Con, placebo group; ProB/ConB, before intervention; ProA/ConA, after intervention

Changes in microbial community composition during probiotic intervention

Then, we analyzed the microbial community composition during probiotic intervention and the results are shown in Fig. 3. The main phyla that account for over 1% in both groups included Bacteroidota, Firmicutes, Proteobacteria and Actinobacteriota. While the top 10 genera in different group included Bacteroides, Prevotella, Faecalibacterium, Roseburia, Romboutsia, Parasutterella, unclassified_f__Lachnospiraceae, Ruminococcus, Parabacteroides, and Blautia. (Figure 3A and B).

Fig. 3

Changes in microbial community composition during probiotic intervention. Pro, probiotic group; Con, placebo group; ProB/ConB, before intervention; ProA/ConA, after intervention

We further analyzed the differences in microbial communities among different groups at the same time point by LEfSe, and the results are shown in Fig. 3C-E. Before intervention, the genera of Blautia, Escherichia-Shigella, Coprococcus, Klebsiella, Akkermansiaceae, Turicibacter enriched in placebo group. After 8 weeks intervention, genera of Klebsiella, Turicibacter still enriched in placebo group. However, genera of Parasutterella, Parabacteroides, Blautia, Dorea, Butyricicoccus enriched in BL21 group. Besides, the control group showed no genus of significant changes before and after intervention. For probiotic group, genera of Acinetobacter, Bradyrhizobium, Stenotrophomonas, and Pseudomonas enriched before intervention. While the genera of Romboutsia, Blautia, Dorea, and Butyricicoccus enriched after intervention.

Metabolomics profiling of plasma during probiotic intervention

To further investigate the impact of probiotic intervention on the serum microbiome, we conducted LC/MS analysis across a total of 132 samples. Between-group differences and within-group pre-post changes were analyzed (Fig. 4). Through, total of 835 differential metabolites were obtained based on inter group and intra group comparisons (Fig. 4A and B). After intervention with probiotics, the enriched metabolic groups in serum significantly increased, with 353 metabolic groups increasing and 118 metabolic groups decreasing compared to the control group. Compared with the baseline period, 418 metabolites increased, and 143 metabolites decreased in BL21 group. During the baseline period, there were a total of 164 metabolites with significant differences between the two groups, while the placebo group had only 173 metabolites with significant differences before and after intervention. Besides, it seems that the serum metabolic profile of the BL21 group was different from the other three groups (Fig. 4C). We further compared the top 50 metabolites with the most significant differences (Fig. 4D). O-Acetylcarnitine, 1-stearoyl-2-linoleoyl-sn-glycero-3-phosphatidylcholine, and some lipid metabolites, including Pc (20:3(5z,8z,11z)/18:0), GPCho (20:3/20:3), LysoPC (18:3(6Z,9Z,12Z)/0:0), etc. The altered metabolites were enriched or decreased in different metabolomic pathways (Fig. 4E-H). For probiotic group, the abundance of protein digestion and absorption, D-amino acid metabolism and ABC transporters pathway was increased significantly. For placebo group, the abundance of Retrograde endocannabinoid signaling pathway, glyocylate and dicarboxylate metabolism and phenylalanine, tyrosine and tryptophan biosynthesis pathway was increased significantly. Compared with placebo group, the sphingolipid metabolism, biotin metabolism and protein digestion and absorption, and alpha-Linoolenic acid metabolism were the top altered pathway in BL21 group compared with placebo group after intervention.

Fig. 4

Analysis of serum metabolite composition during probiotic intervention. Pro, probiotic group; Con, placebo group; ProB/ConB, before intervention; ProA/ConA, after intervention

Correlation analysis of microbiome and metabolome

The correlation analysis of differential bacteria and metabolites was performed to explore the interaction between microbiome and metabolome in different group (Fig. 5). The genera Ruminococcus_torques_group and Hungatella, which were enriched in ProA group, were positively significantly positively correlated with Pc (20:3(5z,8z,11z)/18:0), Pc (18:1(9z)/20:5(5z,8z,11z,14z,17z)), etc. However, these metabolites were negatively correlated with genera of Romboutsia and Turicibacter, which were enriched in ConA group.

Fig. 5

Correlation analysis of microbiome and metabolome

Safety outcomes

Safety outcomes were within the normal range both before and after the study. In the placebo group, 2 (6%) participants had 2 adverse events (enteritis [n = 1], corona virus [n = 1]). In the BL21 group, 2 (6%) participants experienced 2 adverse events (periodontitis [n = 1], corona virus [n = 1]), with no significant differences between groups. No adverse events were related to study medication and no serious adverse events occurred.