Strain-level screening of human gut microbes identifies Blautia producta as a new anti-hyperlipidemic probiotic

ABSTRACT

Compelling evidence has tightly linked gut microbiota with host metabolism homeostasis and inspired novel therapeutic potentials against metabolic diseases (e.g., hyperlipidemia). However, the regulatory profile of individual bacterial species and strain on lipid homeostasis remains largely unknown. Herein, we performed a large-scale screening of 2250 human gut bacterial strains (186 species) for the lipid-decreasing activity. Different strains in the same species usually displayed distinct lipid-modulatory actions, showing evident strain-specificity. Among the tested strains, Blautia producta exhibited the most potency to suppress cellular lipid accumulation and effectively ameliorated hyperlipidemia in high fat diet (HFD)-feeding mice. Taking a joint comparative approach of pharmacology, genomics and metabolomics, we identified an anteiso-fatty acid, 12-methylmyristic acid (12-MMA), as the key active metabolite of Bl. Producta. In vivo experiment confirmed that 12-MMA could exert potent hyperlipidemia-ameliorating efficacy and improve glucose metabolism via activating G protein-coupled receptor 120 (GPR120). Altogether, our work reveals a previously unreported large-scale lipid-modulatory profile of gut microbes at the strain level, emphasizes the strain-specific function of gut bacteria, and provides a possibility to develop microbial therapeutics against hyperlipidemia based on Bl. producta and its metabolite.

KEYWORDS:

• Functional strain screening
• gut microbiota
• hyperlipidemia
• Blautia producta
• 12-methylmyristic acid (12-MMA)
• G protein-coupled receptor 120 (GPR120)

Introduction

Hyperlipidemia is emerging as an extremely common metabolic disorder all over the world, with the major characteristics of high level of lipids circulating in the blood.¹ It does not typically lead to critical symptoms, yet the underlying pathology can progress into serious illnesses, such as cardiovascular diseases and nonalcoholic fatty liver disease (NAFLD), which may ultimately result in death if few managements were applied. Over the last two decades, mounting research has acknowledged the inherent relationship between gut microbiota and human health, not only revealing microbial insights into the etiology of metabolic diseases, but also opening up new avenues for the development of therapeutic strategies based on gut microbes.^2–4 However, candidate bacteria and derived metabolites beneficial to lipid metabolism homeostasis are still in shortage, particularly for those with verified physiological function and elucidated mechanisms. Hence, studies to address these knowledge gaps would facilitate the identification of novel microbial targets to treat hyperlipidemia.

In recent years, joint applications of next-generation sequencing and high-throughput multi-omics data analysis have revealed several taxa contributing to the maintenance of host metabolic health, other than the well-known probiotics-containing genera Lactobacillus and Bifidobacterium. For instance, Akkermansia muciniphila and Faecalibacterium prausnitzii are two species recognized as promising next-generation probiotics, due to their therapeutic efficacy on metabolic disorders (e.g., obesity and type 2 diabetes) and inflammatory bowel diseases (IBD), respectively.⁵ Blautia has also been noticed as a new functional genus with potential probiotic properties.⁶ The abundance of Blautia spp. was closely correlated with the amelioration of type 2 diabetes or hypercholesterolemia by drugs (e.g., metformin and berberine).^7–9 Despite these, only handful of gut bacterial species were pinpointed in previous studies, most of which mainly followed the regular practices to identify beneficial gut microbe(s) in certain contexts, including cohort construction, correlation determination, causality analysis, and functional validation. In fact, a large majority of microbial species in human gut microbiome remain poorly understood regarding their respective regulatory roles in metabolism. As such, it is of significant meaning to conduct the systematic investigation encompassing extensive species to provide a comprehensive understanding of human gut microbiota.

Besides, it is widely accepted that a remarkable degree of strain-level diversity exists in microbial communities residing in human gut, due to the internal genetic variation.¹⁰ Such diversity consequently embodies differences in various aspects such as colonization, pathogenicity and function.¹¹ Increasing evidence has recognized that different strains of the same species can impact host health disparately. Taking Escherichia coli for an instance, some strains are prevalent in the healthy human gut or exert probiotic-like actions (e.g., Nissle1917,¹² while others could cause life-threatening infections.¹³ Importantly, functional specificity of bacterial strains also gives a caution that disclosure of probiotic or therapeutic microbes should build upon rigorous assessments at the strain level and narrow down to the specific strain, rather than the species or genus level. Moreover, recent advances in sequencing technology and bioinformatics algorithms are triggering microbial studies down to the strain level.¹⁴ Therefore, it is warrant to perform the strain-level massive investigation of human gut microbes to assess respective regulatory roles in metabolism (e.g., lipid metabolism), which would not only provide a comprehensive understanding regarding the lipid-modulatory landscapes of human gut microbiota but also present plenty potential microbial candidates favoring the development of new microbial interventions against hyperlipidemia.

In this study, to present a previously unreported lipid-modulatory profile of human gut microbiota at the strain level, we systemically evaluated the lipid-decreasing activities of 2250 human gut bacterial strains, covering 186 species, using a large-scale cellular screening, and a mucosal bacterium strain Blautia producta was found to show the most potent efficacy. We then confirmed its lipid-lowering effect in high fat-diet (HFD)-induced hyperlipidemic mice, and identified an anteiso-fatty acid, 12-methyl myristic acid (12-MMA), to be the important active metabolite using a joint approach of pharmacology, comparative genomics, and metabolomics. Furthermore, cellular and animal experiments were performed to evaluate the anti-hyperlipidemic activity of 12-MMA and explore its potential mechanism.

Results

Gut microbiota exhibits strain-specific effects on intracellular lipid accumulation

To get a comprehensive insight into the roles of gut microbes on lipid metabolism at the strain level, we set up a cell-based high-throughput screening platform to assess the lipid-modulatory activities of gut bacteria using the conditioned culture media (Figure 1a). In our study, a total of 2250 strains were screened to evaluate their effects on oleic acid-elicited intracellular lipid accumulation. These strains belonged to 5 phyla, 12 classes, 48 genera, and 186 species (Figure 1b; Table S1). After three runs of screening, 388 strains showed relatively steady lipid-lowering effects, which were regarded as positive strain in the context. Of them, 294 strains were from Firmicutes, 76 from Actinobacteria, 13 from Proteobacteria, and 5 from Bacteroidetes (Figure 1b,c; Table S1). At the genus level, 8 genera provided more than 25 strains for screening and their pos rates were ranked as Lactobacillus (24.48%, 141/576), Clostridium (19.44%, 14/72), Lysinibacillus (17.83%, 28/157), Bifidobacterium (16.63%, 75/451), Bacillus (14.41%, 84/583), Enterococcus (9.85%, 13/132), Escherichia (8.00%, 2/25), and Streptococcus (5.88%, 2/34) (Figure 1d; Tables S1 and S2). Down to the species level, especially for individual species with ≥25 tested strains, Lactobacillus gasseri (43.88%), Bifidobacterium pseudolongum (28%) and Bifidobacterium catenulatum (25%) showed higher positive rates, while Enterococcus faecails (8.82%) and Escherichia coli (8%) contained less lipid-lowering strains (Figure 1e, Table S3). These data provided several valuable bacterial pools for further identification of strains beneficial to lipid metabolism.

Figure 1. Strain-level screening of gut microbes for lipid-lowering activity in HepG2 cells. (a) Flowchart of cell-based screening for gut bacterial strains with lipid-lowering activities. (b) Numbers of tested strains in each taxon. The total screened and positive strains numbers at phylum (c) and genus (d) levels. (e) the rank of the positive rate for species containing ≥ 25 tested strains in this study. (f) the probability of species to provide at least one positive strains. The black value on the top of each bar indicates the number of species with equal or more number of tested strains in this study than that shown on the x-axis. The percentage in red indicates the rate of species providing one or more positive strains. Positive strains in the context are those showing relatively steady lipid-lowering effects in three rounds of screening.

Additionally, we observed that the probability of species containing lipid-decreasing strains rose sharply as the strain numbers increased. All species with ≥14 tested strains contained at least one positive strain, independent of their taxa, while many species with <5 tested strains did not exhibit lipid-decreasing effect in these screenings (Figure 1f; Table S1). Notably, some species, such as B. adolescentis, L. gasseri, and B. licheniformis, contained lipid-increasing, lipid-lowering, and neutral strains in parallel (Table S2), exhibiting the strain-specific regulations on lipid accumulation. Besides from these probiotic bacteria, we also noted a beneficial species, Blautia producta, whose all tested strains showed prominent lipid-decreasing activities, with comparable efficiency to the medication fenofibrate, which is used to lower high lipids levels in the blood.

Bl. producta exerts potent anti-hyperlipidemic action

Although the emerging benefits of Blautia producta have been captured by researchers, it remains insufficiently understood; so we focused on the species to carry out in-depth investigations. To systemically assess the lipid-lowering effect and identified the active metabolite(s) of Bl. producta, we selected one Bl. producta strain (No. 501 in Table S2) with the strongest lipid-reducing effect to perform investigations as depicted in Figure 2.

Figure 2. Graphical outline for assessment of the lipid-lowering effect and identification of the potential active metabolite of Blautia producta. Bl. producta strain with the best efficacy in cell-based screening was re-assessed for the lipid-lowering effect in HepG2 cells and high-fat diet-induced hyperlipidemic mice. Pan-genomics and comparative metabolomics analysis were performed to identify the potential active metabolites of Bl. producta, then the anti-hyperlipidemic effect and mechanism of the candidate metabolite were investigated in mice.

We first performed a series of cell-based examinations to confirm the effect of Bl. producta (referring to the No. 501 strain) on lipid accumulation. The conditioned culture medium of Bl. producta exerted lipid-lowering capacity at a low concentration of 5% (v/v) and reached a potent efficiency at 30% (v/v), as high as that of fenofibrate (Figure 3a). OD358 values of oil O red staining (Figure 3b) and BODIPY staining of lipid droplets (Figure 3c) also showed that Bl. producta dose-dependently inhibited lipid accumulation. Subsequent measurements of intracellular triglyceride content corroborated the lipid-decreasing effect in HepG2 cells (Figure 3d). To assess its beneficial role in vivo, we treated HFD-fed mice with Bl. producta via continuous gavage of 10⁹ CFU per animal per day for 4 weeks (Figure 3e). Live Bl. producta significantly suppressed HFD-induced body weight gain (Figure 3f,g) and decreased fat deposition (Figure 3h). High levels of lipids in the blood (TG, TC, and LDL-c) and liver (TG and TC) were reduced by Bl. producta treatment (Figure 3i,j); consistently, there were less lipid droplets in liver tissues of HFD mice with Bl. producta, as compared with HFD mice (Figure 3k), thereby presenting the improved hyperlipidemia and liver steatosis. To further consolidate the anti-hyperlipidemic effect of Bl. producta, we first applied the antibiotics cocktail to construct pseudo-germ-free mice and then orally administrated Bl. producta to antibiotics-treated HFD mice (Figure S1a). The bodyweight of ‘germ-free’ mice gavaged Bl. producta (Abs+HFD+Bl. producta) was clearly lighter than control mice (Abs+HFD) (Figure S1b,c), as well as the weight of liver and fat tissues (Figure S1d–g). Other hyperlipidemia-associated phenotypes such as serum and liver lipids were also improved (Figure S1h-j), in accordance with the performances of Bl. producta in conventional HFD mice. Hence, our results collectively corroborate the hypolipidemic effect of Bl. producta.

Figure 3. Blautia producta ameliorates HFD-induce hyperlipidemia. The in vitro lipid-lowering effect of Bl. producta conditioned culture medium was evaluated by oil red O (a and b) and BODIPY staining, scale bars, 200 μm (c), and intracellular TG quantification (d), with Fenobirate (10 μM) as a positive control. (e) Schematic of animal experiment showing the assessment of anti-hyperlipidemic effect of Bl. producta in HFD mice. (f) Bodyweight curve. (g) Body weight gain. (h) Weight measurement of retropertoneal fat and epididymal fat. (i) Serum lipid levels of TG, TC, LDL-c and HDL-c. (j) Lipid contents of TG and TC in liver tissue. (k) Oil red O staining of liver tissue, scale bars, 100 μm. n = 8 for each group. *P < .05, **P < .01, ***P < .001.

Bl. producta can exert lipid-lowering effect without colonization in mice gut

We also evaluated the impact of Bl. producta administration on the gut microbiota of host mice by performing shotgun sequencing. Few influences were imposed on the α-diversity of intestinal microbial community by Bl. producta administration (Figure S2a,b), but the compositional structure was shifted (Figure S2c-d). Bl. producta could restore the relative abundance of Akkermansia compared to HFD mice (Figure S2e). At the species level, A. muciniphila as well as some other probiotic mucosal bacteria were significantly increased, while opportunistic pathogens Desulfovibrio piger and Desulfovibrio sp. G11 were declined (Figure S2f). However, the gavage of live Bl. producta did not increase its relative abundance in feces rather than decreased after the 4-week treatment (Figure S2f). To figure it out, we performed an extra dynamic detection by repeating the animal experiment of Bl. producta and sequentially collecting feces on Day 1, 3, 7, 14 and 28. Oral administration of Bl. producta dynamically shifted the overall structure of gut microbiota (Figure S3a). The impact of Bl. producta on gut microbiota was the greatest on Day 1, then turned weaker from Day 3 to Day 14, yet remained evident on the 28^th day after treatment (Figure S3a). Similarly, fecal Bl. producta in mice with exogenous live Bl. producta were significantly higher than HFD mice on Days 1 and 7, while this stimulation gradually disappeared afterward. On Day 28, fecal Bl. producta in Bl. producta-treated mice was less than that in control mice (Figure S3b). At the same time, Bl. producta showed the same modulatory trends on Akkermansia spp. and other probiotics in both experiments (Figures S2f and S3c,d). These data indicate that Bl. producta is a beneficial microorganism with relatively poor colonizing capacity.

To decipher whether Bl. producta could exert lipid-lowering effect with insufficient colonization, we conducted an additional animal experiment by treating HFD mice with live and pasteurized Bl. producta and prolonged the administration duration from 4 weeks to 8 weeks (Figure 4a). After gavage for 8 weeks, both live and pasteurized Bl. producta (BP) remarkably inhibited HFD-induced body weight gain (Figure 4b,c). Moreover, the two forms of BP also significantly decreased weight of subcutanenous fat tissue and liver (Figure 4d,e), as well as lipids levels in serum and liver (Figure 4f,g). OGTT curve and the accompanied AUC statistics showed that the glucose tolerance in HFD mice was improved by both BP treatment (Figure 4h,i). According to oil O red and HE staining, less lipid droplets and accumulation in mice gavaged live and pasteurized BP (Figure 4j,k), thus presenting improved liver steatosis. Besides, we observed that the anti-hyperlipidemic efficacy of pasteurized BP was slightly weaker than the live form. Even so, our results evidenced that the lipid-lowering action of Bl. producta can be achieved without colonization.

Figure 4. Both live and pasteurized Blautia producta exhibit potent lipid-lowering effect in HFD-fed mice. C57BL/6 mice were orally administrated via gavage of live or pasteurized Bl. producta for 8 weeks. (a) Schematic of animal experiment to assess the anti-hyperlipidemic effect of pasteurized Bl. producta. (b) Body weight curve. (c) Body weight gain. (d) Subcutanenous fat weight. (e) Liver weight. (f) Serum levels of TG, TC, LDL-c and HDL-c. (g) Liver lipid contents of TG and TC. (h and i) OGTT curve (h) and area under the curve (AUC) of blood glucose level (i). (j) Oil red O staining of liver tissue. Scale bars = 100 μm. (k) H&E staining of liver tissue. n = 8 for each animal group. Scale bars = 50 μm. *P < .05, **P < .01, ***P < .001. BP indicates Blautia producta or Bl. producta.

To obtain more accurate regulation of Bl. producta on mice gut microbiota, we collected fecal samples and performed shotgun sequencing at the depth of 5Gb (Figure 5). Similar with 1 G data analysis, treatment with live Bl. producta imposed little influence on the α-diversity of gut microbiota (Figure 5a), but shifted its overall structure and composition as depicted by PCoA/NMDS analysis (Figure 5b,c) and the relative abundance of top 20 genera (Figure 5d). Of note, gavage of live Bl. producta for 8 weeks resulted in a significant increase of fecal Bl. producta. As expected, we did not detect Bl. producta in mice treated with pasteurized BP (Figure 5e). For bacterial species with relative abundance over 0.5% in live Bl. producta-treated mice, Akkermansia muciniphila and Staphylococcus sciuri were significantly increased while Enterorhabdus caecimuris and Lactobacillus reuteri were decreased (Figure 5e). Oral administration of pasteurized Bl. producta also significantly increased Akkermansia muciniphila and reduced Enterorhabdus caecimuris but it did not decrease Lactobacillus reuteri (Figure 5e). Pasteurized Bl. producta further reduced the abundance of Lactobacillus murinus, Bifidobacterium mongoliense, Leuconostoc mesenteroides and Streptococcus thermophilus while live Bl. producta showed no significant impact (Figure 5f). LefSe and LDA analysis also displayed that live Bl. producta enriched beneficial bacteria such as Akkermansia spp., Bl. producta, Erysipelotrichaceae and Corynebacteriaceae (Figures 5h and S4). Collectively, Bl. producta displays a poor capacity in mice gut, but can exert lipid-lowering effect in a pasteurized form independent of colonization. We thus guess that Bl. producta might exert lipid-lowering action via generating active compounds other than stimulating the growth of intestinal Bl. producta.

Figure 5. Blautia producta modulates gut microbiota but shows a poor colonizing capacity. The alpha diversity of gut microbiota was assessed by Shannon and Simpson indices (a). The separation of the gut microbial community of different groups was assessed by principal coordinates analysis (PCoA) (b) and non-metric multi-dimensional scaling (NMDS) (c). (d) the profile of dominant genera in each group. (e) the relative abundances of Bl. producta and significantly altered species by live Bl. producta. (f) the relative abundances of significantly altered species by pasteurized Bl. producta. (g) LEfSe Cladogram of gut microbiota community. n = 8 for each group. *q < .1, **q < .05, ***q < .01.

12-Methyl myristic acid is an important active metabolite of Bl. producta.

Microbiota-derived metabolites are vital mediators to maintain host-commensal interactions, we then conducted pan-genomic analysis and metabolomics to identify the active metabolite(s) mediating the beneficial effect of Bl. producta. Five new genomes of Bl. producta by sequencing and de novo assembly (namely I2DA, ID8A, ID9B, I24C and I31A), and 7 reference genomes from public RefSeq databases were collected and analyzed. The five lipid-lowering Bl. producta strains tended to cluster in two neighbor branches (Figure S5a) and genes involve in lipid metabolism were uniformly enriched (Figure S5b), implying that the cellular activities of lipid metabolism may be active in these effective Blautia strains. To find the metabolites preferentially produced by Bl. producta, we originally planned to perform the untargeted metabolomics analysis to compare Bl. producta with all screened bacteria. Given that it would be rather laborious and repetitive to analyze all strains, we figured out to select 500 out of 2250 bacterial strains as the representative, covering all bacterial genera tested in our study. Their individual monocultures were pooled together with equal weight to make an entirety mixture, then the signal intensity of each recognized metabolite produced by Bl. producta was compared with those of the entirety mixture to assess the priority of Bl. producta to produce each metabolite. A total of 614 metabolites were identified and 13 ones were highly (>2.8 fold) generated by Bl. producta over the entirety (Figure 6a; Table S4). Of them, 12-methyl myristic acid (12-MMA) showed the most priority in Bl. producta culture, with its relative intensity 26.4 times greater than that of the whole. These findings suggest that 12-MMA is likely to be the most dominant microbial metabolite of Bl. producta, at least among the 614 identified ones. Moreover, we also detected a considerable amount of 12-MMA in the conditioned medium of Bl. producta (~58.3 μmol/L) by gas chromatography-mass spectrometry (GC-MS) analysis (Figure 6b).

Figure 6. Metabolomic analysis identifies 12-methylmyristic acid (12-MMA) as a key active metabolite of Bl. producta. (a) the top 13 metabolites that are markedly enriched in Bl. producta. (b) Mass spectrums of 12-MMA detected in the conditioned culture medium of Bl. producta. RT: retention time. (c-d) PLS-DA analysis (Partial least squares Discriminant Analysis) and Clustering analysis of the metabolomics data of 11 strains including 10 Bacilli licheniformis strains and one Bl. producta (labeled as 1K83 for metabolomics analysis). (e) Volcano plot of metabolites of effective and ineffective groups. p: p value, FC: fold change. 3 metabolite(s) with the absolute value of FC >2 and p value < 0.05 were selected. (f) the abundance of 3 key metabolites in effective/ineffective stains identified from volcano plot. (g) Random forest analysis showing the top 15 metabolites that contribute to the discrimination of effective/ineffective stains. n = 6 for effective strains, n = 5 for ineffective strains. (h,i) Cecal and serum levels of 12-MMA (h) and acetate (i) in mice treated with live or pasteurized Bl. producta^.*P < .05, **P < .01, ***P < .001. Effective group indicates strains with lipid-decreasing effect and ineffective group comprises lipid-increasing strains.

Next, we set about to investigate whether 12-MMA could act as the functional component of Bl. producta to convey the hypolipidemic effect. The routine approach identifying active metabolites of a bacterial species is to analyze the metabolomics of strains with distinct pharmacological efficiencies, followed by the correlation construction of metabolites with differential efficacies. In our work, all tested Bl. producta strains were highly positive to suppress lipid accumulation, with no negative effects, so that it is hard to unravel key active metabolites following this strategy. As such, we modified the method by selecting another gut bacterial species as a reference, which should not only harbor both lipid-decreasing and increasing strains but also largely resemble the microbial metabolites profile of Bl. producta. The species Bacilli licheniformis was found to virtually meet the above requirements, so 10 strains of Ba. licheniformis, together with the Bl. producta strain were analyzed by dividing them into two groups based on their effects on lipid accumulation: the effective (lipid-decreasing) and ineffective (lipid-increasing) groups. To evaluate whether metabolites of the two groups could be distinguished just as their distinct functions, we performed partial least squares discriminant analysis (PLS-DA) using their metabolomics data. Metabolites of both bacteria groups were markedly separated (green ellipse and pink ellipse) (Figure 6c). Consistently, the hierarchical clustering analysis clearly displayed their gathering and shifting (Figure 6d), suggesting there would be certain metabolites differentially existing in them. According to Volcano plot, several metabolites were differently distributed among 11 bacteria (Figure 6e,f), for instance, 12-MMA and anandamide were significantly enriched in effective strains, whilst pyrrolidonecarboxylic acid was higher in the ineffective group (Figure 6f). Moreover, random forest analysis also exhibited that 12-MMA was the most critical metabolite to distinguish the effective bacteria from the ineffective ones (Figure 6g; Table S5). Altogether, we speculated that 12-MMA may act as the potential active metabolite of Bl. producta to mediate its beneficial efficacy.

We then measured the cecal and serum contents of 12-MMA in mice administrated with live or dead Bl. producta. Given that previous studies once reported that a marked increase in acetate production by Bl. producta administration upon inulin consumption contributed to the lipid-lowering effect of inulin,¹⁵ so we analyzed the acetate level in serum and cecum in parallel. Treatment of live Bl. producta elevated both serum and cecal levels of 12-MMA and acetate, but pasteurized Bl. producta only significantly increased 12-MMA not acetate (Figure 6h,i). As acetate is main produced by live bacteria while 12-MMA is a compositional lipid of bacterial membrane and can persist in dead cells, these results suggest that 12-MMA may be an important metabolite over acetate for the lipid-lowering efficacy of Bl. producta.

12-MMA effectively alleviates HFD-induced hyperlipidemia in mice

We then treated HFD mice with 12-MMA to assess its in vivo anti-hyperlipidemic effect (Figure 7a). Oral administration of 12-MMA showed few impacts on the food and water intake of HFD mice (Figure 7b,c), but could significantly inhibit HFD-induced body weight gain during the whole experiment period (Figure 7d). Mice with 12-MMA treatment also had less fat accumulation compared with HFD control mice (Figure 7e). High serum levels of lipids in HFD mice were reduced by 12-MMA, including total glyceride (TG), total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-c), as compared to the control group (Figure 7f). Moreover, liver steatosis in HFD mice was clearly alleviated upon 12-MMA treatment, according to lipid quantification and oil-red O staining (Figure 7g,h). All these results indicate that 12-MMA could effectively ameliorate HFD-induced hyperlipidemia.

Figure 7. 12-MMA effectively alleviates HFD-induced hyperlipidemia and activates GPR120. (a) Schematic diagram of animal experiment. 12-MMA was administrated to HFD mice for 4 weeks to evaluate its anti-hyperlipidemic effect. (b-c) Food and water intake of mice with 12-MMA treatment. (d) Body weight gain of HFD-mice treated with 12-MMA or vehicle control. (e) Fat weight. (f) Serum lipids levels of TG, TC, and LDL-c. (g) Lipids contents of TG and TC in liver. (h) Oil red O staining of liver tissue, scale bars, 100 μm, 50 μm for inlay. (i) H&E staining of inguinal white adipose tissue (iWAT), scale bars, 200 μm. (j) mRNA levels of fat browning-related maker genes in iWAT. (k) Immunostaining of Ucp1 in iWAT, scale bars, 500 μm, 200 μm for inlay. (l) Oral glucose tolerance test (OGTT). (m) Serum levels of insulin and GLP-1 in mice with 12-MM. n = 8 for each animal group. 12-MMA was suspended in solvent (1% tween 80 + 0.5% carboxymethyl cellulose sodium (CMC-Na)) and the HFD group was treated with an equal volume of solvent. (n) Fluorescence-based GPR120 activation in response to 12-MMA. (o) Fluo-4-AM intensity and GLP-1 secretion (p) in STC-1 cells with 12-MMA treatment. n = 6 for each cell group^.*P < .05, **P < .01.

In addition, we found that 12-MMA-administrated mice displayed plentiful multilocular adipocytes in inguinal white fat tissue (iWAT), presenting a typical characteristic of WAT browning (Figure 7i), then the transcription of crucial thermogenic genes were evaluated, including Pgc1α, Ucp1, Cidea and Cpt1b. 12-MMA remarkably up-regulated their relative mRNA levels in iWAT (Figure 7j). The protein expression of Ucp1 was also consequently elevated in 12-MMA-treated mice, compared with HFD control mice (Figure 7k). Besides, 12-MMA supplementation improved the oral glucose tolerance induced by HFD feeding (Figure 7l). Serum levels of insulin and glucagon-like peptide 1 (GLP-1) were obviously promoted in response to 12-MMA treatment, which are two crucial hormones to regulate host glucose homeostasis (Figure 7m). Based on these, it is plausible that 12-MMA could play beneficial roles in regulating both lipid and glucose metabolism. GPR120, also referred as FFAR4 (free fatty acid receptor 4), is a critical hub in modulating WAT browning and glucose consumption via binding with its ligands LCFAs to fulfill activation. As a typical readout of GPR120 activation, the secreted GLP-1 can trigger insulin release to maintain glucose homeostasis,¹⁶ thus providing a hint that 12-MMA might exert the beneficial efficacy via activating GPR120. Subsequently, we examined the influence of 12-MMA on GPR120 activation using a fluorescence-based assay in STC-1 enteroendocrine cells. It was found that 12-MMA drastically strengthened GPR120 activity in a dose-dependent manner (Figure 7n). Since GPR120 activation can boost the intracellular Ca²⁺, which is pivotal for the release of GLP-1 incretin, we then evaluated whether 12-MMA could elicit the rise-up of Ca²⁺ in STC-1 cells. As expected, the fluorescent signaling of Ca²⁺, as well as GLP-1 secretion was gradually elevated with the increase of 12-MMA concentration (Figure 7o,p), further confirming the activation of GPR120 in response to 12-MMA. Collectively, we reason that Bl. producta-produced active metabolite, 12-MMA could exert beneficial effects on lipid/glucose metabolism via stimulating GPR120.

Discussion

Over the past two decades, it has been well-acknowledged that the gut microbiota plays a critical role in the regulation of human lipid metabolism, which contributes to elucidate the etiology of dyslipidemia in a microbial perspective.^4,17 Intensive studies provide essential rationale for the development of novel therapeutic approaches from the vast bacteria reservoir. However, the exact regulatory effect of individual species/strain in human gut on lipid homeostasis remains largely unknown, hindering the comprehensive understanding of gut microbiota and its actual exploitation. Here, we, for the first time, exhibited a considerable functional profile regarding lipid-modulatory capacities of 2250 human gut microbial strains, also identified a potential bacterium Bl. producta and its active microbial metabolite, 12-MMA, which were both demonstrated to ameliorate hyperlipidemia in HFD mice.

With the development of multi-omics approaches and computational analysis, microbiome researches have moved from descriptive analysis to cause-and-effect studies and can taxonomically pinpoint from the phylum, down to the genus, species, and strain level. Culture-based screening investigations on gut microbiota have been conducted to understand its functions or association with hosts, such as antibiotic resistance,¹⁸ probiotic properties,¹⁹ and drug metabolism.²⁰ However, the microbial community in human gut is enormous and most residents are strictly anaerobic and nutrients-demanding, thus restraining the actual scales of quantity and species/strain in screening studies. In our work, we set up a cell-based high-throughput screening platform to seek the functional human gut bacteria with satisfying lipid-decreasing effects. The examined 2250 strains covered 186 bacterial species and represented an essential part of human gut microbiota. 388 out of the 2250 strains (about 17%) steadily suppressed the lipid accumulation in three rounds of evaluations. Firmicutes (17.81%) and Actinobacteria (16.59%) possess a larger portion of lipid-decreasing bacteria than Bacteroidetes (13.16%) and Proteobacteria (12.87%). At the genus level, Blautia (100%), Providencia (33.33%), Pseudomonas (28.57%), Lactobacillus (24.48%), Bacteroides (22.22%), Mitsuokella (20.00%), Clostridium (19.44%), Lysinibacillus (17.83%), Bifidobacterium (16.63%) and Bacillus (14.41%) were the top 10 genera containing lipid-lowering strains. Among them, Lactobacillus and Bifidobacterium are two renowned genera containing multiple species beneficial to hyperlipidemia and NAFLD.^21,22 More than 60.99% of positive Lactobacillus strains were from L. gasseri, a probiotic species with great potential to facilitate weight loss in humans and rodents.^23,24 Another interesting species is Lysinibacillus sphaericus, 16.83% (17/101) of its tested strains stably inhibited lipid accumulation though the biological activity of this species is rarely reported. Therefore, our screenings not only showed a large-scale functional profile of intestinal bacteria at the strain level, but also provided valuable bacteria pools with massive potentials for therapeutic exploitation in future. Taking L. gasseri for example, we have further selected 166 strains to assess their lipid-lowering effects and identified a special strain L. gasseri BDUP with potent anti-hyperlipidemic activity in vitro and in vivo, ²⁵ which could be therapeutically or commercially developed. In addition, we observed obvious strain-specific regulations on cellular lipid accumulation, so close attention should be paid when it comes to the exploration of active functional intestinal bacteria or probiotics. It would be more accurate to set the investigation scale down to the strain level, rather than the upper species.

One important finding of this work is the identification of Bl. producta, an intestinal bacterium belonging to Blautia, which is regarded as a new functional genus with potential probiotic characteristics.⁶ Although accumulating evidence documented that Blautia spp. were implicated in the amelioration of human type 2 diabetes or hypercholesterolemia by drugs (e.g., metformin and berberine),^7–9 the specific species were less mentioned in most of the descriptive results. Recently, more species-targeting researches on Blautia have been released. Decreased Blautia luti was detected in the gut ecosystem of obese children and related to the worsening of intestinal inflammation and metabolic phenotype.²⁶ Another cross-sectional study of Japanese adults identified Blautia wexlerae could ameliorate obesity and type 2 diabetes.²⁷ Two species, Blautia hansenii and Blautia producta were reported to significantly and negatively associated with visceral fat accumulation using a one-year longitudinal study.^28,29 Bl. producta has also been closely correlated with the amelioration of human type 2 diabetes or hypercholesterolemia by drugs (e.g. metformin and berberine),^7,9 though earlier studies merely found its essential role in the colonization resistance to the invasion of Vancomycin Resistant Enterococci (VRE)^13,30 and anti-inflammatory effect in HT-29 intestinal epithelial cells.³¹ As such, the favorable roles of Blautia, particularly the individual species, are emerging. In our screening, Bl. producta was of the most attraction species, because all tested strains robustly decreased lipid accumulation with similar efficacy to the positive drug fenofibrate. Furthermore, gavage of live Bl. producta to HFD mice was effective to suppress HFD-induced body weight gain, decrease serum levels of lipids, and alleviate liver steatosis in both conventional mice and antibiotics-induced pseudo-germ-free mice, exhibiting potent efficacy against hyperlipidemia. Hence, the demonstrated anti-hyperlipidemic effect of Bl. producta would contribute to pave a path for the development of Bl. producta as a next-generation probiotic.

Gut microbiota-derived metabolites are critical factors bridging the commensal community with various physiological and pathological conditions in the host, thus identifying functional metabolites of Bl. producta would be a key step to elucidate the underlying mechanism of Bl. producta. Our results showed that Bl. producta was of poor colonization in mouse intestine but heat-kill Bl. producta exhibited comparable anti-hyperlipidemic effect as live cells. Therefore, it is reasonable that this bacterium may benefit lipid homeostasis via producing active compounds. According to metabolomics analysis, we determined a fatty acid, 12-MMA, as an important metabolite of Bl. producta and subsequently verified its marked efficacy against hyperlipidemia in HFD mice, thereby broadening our knowledge of Bl. producta and providing a prerequisite for further mechanistic research. Referring to microbial metabolites, free fatty acids (FFAs) are one of the best-characterized microbial metabolites types to act as the proxy of intestinal bacteria to modulate host physiology, e.g., lipid/glucose metabolism.²¹ Elevating the production of short-chain fatty acids (SCFAs) or direct dietary supplementation of SCFAs could significantly ameliorate dyslipidemia and NAFLD.³² Unlike the renowned SCFAs, long-chain fatty acids (LCFAs) were just been observed to maintain the balance of energy metabolism in recent years.^33,34 12-MMA found in our work is also named as 12-methyltetradecanoic acid or anteiso-C15:0 fatty acid, belonging to LCFAs family. Our results of 12-MMA again highlights the importance of LCFAs in metabolism, together with another LCFA, myristoleic acid (MA), which was previously revealed to reduce obesity through brown fat activation.³⁵ Moreover, other than the identity of LCFA, 12-MMA is, to be exact, an anteiso-fatty acid (ai-FA), which is a type of fatty acid containing an alkyl branch at the ante-penultimate carbon atom of a chain. The branched-chain fatty acids, ai-FAs, have been recognized as a main category of microbial metabolites.³⁶ Compared with the straight-chain saturated counterparts, ai-FAs are featured with evidently decreased melting point,³⁷ which may greatly increase the membrane fluidity of their producers and thus change the physiological functions, just like their straight-chain unsaturated counterparts. Nevertheless, studies on the physiological functions of microbial ai-FAs are rather scarce. Colosimo et al. once reported that several ai-FAs could act as the polyunsaturated FAs do,³⁸ suggesting the functional resemblance of ai-FAs with unsaturated FAs. A lately study published in Nature revealed a profound finding that a diacyl phosphatidylethanolamine (a15:0-i15:0 PE) from A. muciniphila’s cell membrane, also a derivative of anteiso-C15:0 fatty acid, can induce homeostatic immune responses,³⁹ indicating lipids are another critical components to mediate its functions, other than proteins (e.g. Amuc_1100 and P9). Hence, the identification of 12-MMA as an active metabolite to improve hyperlipidemia broadens our knowledge of ai-FAs and LCFAs on their physiological functions.

Furthermore, LCFAs can activate the G protein-coupled receptor 120 (GPR120/FFAR4), a versatile regulator for lipid and glucose metabolism,^16,40 such as the ω-3 fatty acids including docosahexaenoic acid (DHA), α-linolenic acid (ALA), and eicosapentaenoic acid (EPA).^16,41 Once activated, GPR120 can not only effectively provoke WAT browning and brown fat mobilization but also trigger GLP-1 release to modulate glucose metabolism.^42,43 In our study, oral administration of 12-MMA remarkably up-regulated key thermogenic factors and elicited many multilocular adipocytes in iWAT, exhibiting the typical phenotypes of WAT browning. Together with the improved oral glucose tolerance and elevated insulin and GLP-1 in 12-MMA-treated mice, we naturally associated this fatty acid with GPR120 activation and thus conjectured that 12-MMA, as a LCFA, might be a ligand for GPR120. As speculated, we found 12-MMA potently stimulated GPR120 in a dosage-dependent way. Colosimo and colleagues also revealed the strong response of GPR120 to 12-methyltetradecanoic acid (referred to as 12-MMA in the context),⁴⁴ rightly supporting our findings. In addition, due to the analogous characteristics of anteiso- and unsaturated-FAs counterparts (12-MMA/MA), there is a possibility that MA may activate GPR120 as well to result in the previously reported BAT activation and obesity improvement.³⁵ Actually, we observed the activating effect of MA on GPR120 (data not shown), providing a possible mechanistic explanation to the anti-obesity effect of MA. It is worth to noting that the transmembrane family, GPCRs, to date, is a quite successful type of therapeutic targets for various diseases,^45,46 thus 12-MMA might be a potential candidate agent against hyperlipidemia via targeting GPR120, yet massive sophisticated studies are awaited to verify this conjecture.

Although our results provide a large-scale lipid regulatory profile of human gut microorganisms and revealed the anti-hyperlipidemic effect as well as potential mechanism of Bl. producta, several limitations exist in this work which need further attention and verification. Firstly, the functional screening was only performed based on immortalized hepatocellular carcinoma cells. A large-scale animal-based functional screening, such as using C. elegans, may provide in vivo activities of different bacterial strains. Secondly, Bl. producta isolated from human is poorly colonized in mouse gut in our work, thus whether mouse-derived Bl. producta could exhibit a better colonizing capacity that needs to be further investigated. Thirdly, 12-MMA was identified as an important active metabolite to mediate the anti-hyperlipidemic effect of Bl. producta, but we cannot exclude the contributions of other microbial metabolites and compositional factors which are not investigated in the present study, future in-depth work will facilitate the thorough dissection of Bl. producta.

In summary, our study first displayed a large-scale lipid-modulatory profile of human gut microbiota at the strain level and demonstrated .Bl producta and its metabolite, 12-MMA, could exert prominent anti-hyperlipidemic effect. Our work opens up new avenues to systematically understand the modulation of gut microbes on host health and lays theoretic basis for the exploitation of Bl. producta and its metabolite as the potential therapeutics to manage hyperlipidemia or related metabolic disorders.

Materials and methods

Preparation of conditioned culture media

All tested bacteria were from the human gut microbial library constructed by Beijing QuantiHealth Technology Co., Ltd. (Beijing, China) (https://microbe.quantibio.com/). Specifically, all strains were isolated from the mixture of fecal samples donated by healthy adults in Hainan province, China. A questionnaire survey was conducted among the population before samples collection to obtain information of age, dietary habits and diseases history. The subjects have all signed a written consent and the project has been approved by the Hainan Branch of the General Hospital of the People’s Liberation Army (PLAGH)’s ethics committee under number 301hn11-2017-03. Each bacterial strain was purely cultivated, identified by mass spectrometry, labeled with a unique code, and then were cryopreserved at −80°C refrigerator.

For screening assay, each bacterial clone was pin transferred from the cryopreserved tube onto a solid yeast casitone fatty acid (YCFA) medium (Solarbio) plate and was grown for 24–48 h at 37°C under anaerobic conditions. A single colony was inoculated into liquid YCFA medium and anaerobically cultured at 37°C for 48 h to obtain the first generation (F1) bacterial solution. 10% (v/v) of F1 bacterial solution was inoculated into fresh liquid YCFA medium and cultured at the same conditions for 48 h to generate F2 solution. Following the similar step as F2 production, the working bacterial solution F3 was further obtained from 10% (v/v) F2. After centrifugation at 1,600×g for 15 min, the supernatant part was collected and then filtered using a Millipore filter (0.22 μm). These sterile conditioned culture media were used as screening samples in the study.

Cellular lipid accumulation assay

HepG2 cells (purchased from National Infrastructure of Cell Line Resource, China) were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific), 100 U/ml penicillin, and 100 μg/L streptomycin incubated at 37°C in a humidified atmosphere containing 5% CO_2. For lipid accumulation assay, HepG2 cells were seeded in 96-well plates containing 100 μl DMEM. When the confluence reached 85%, the medium was replaced with 70 μl fresh DMEM supplemented with 100 μM oleic acid (OA, Sigma) and 30 μl spent culture broth filtrate or YCFA medium. After incubation for 22–24 h, lipid accumulation was evaluated by oil red O staining or TG quantification kit (Solarbio) as described previously.⁴⁷ Each experiment (n = 8 for oil red O staining, n = 4 for TG determination) was repeated in triplicate. Liquid YCFA medium and fenofibrate (10 μM, Sigma) were used for the negative and positive control, respectively. Specifically, (1) for the large-scale bacteria screening, HepG2 cells were first stained with oil red O solution at room temperature for 30 min, and then dimethyl sulfoxide (DMSO) was added to dissolve stains attached to cells, followed by measuring optical density (OD) value at 358 nm in a microplate reader. Lipid-lowering efficiency of each examined bacterial strain was evaluated and ranked according to the calculation of [(YCFA OD₃₅₈ - Sample OD₃₅₈)/YCFA OD₃₅₈]*100%. (2) To confirm the lipid-lowering efficacy of Blautia producta, BODIPY staining and intracellular TG content quantification were performed to corroborate the screening result according to the manufacturers’ instructions.

Animal experiment

All the animal experiments were performed in accordance with the National Institutes of Health regulations for the care and use of animals in research. The protocol was approved by the medical ethics committee of Peking Union Medical College (Nos. YZS201904021; YZS201910013; YZS202105022).

1. To assess the anti-hyperlipidemic effect of live Bl. producta, 24 male C57BL/6J mice (8-week-old, 20–24 g, Vital River Laboratory Animal Technology) were divided into three groups with eight animals in each group. One group was used as blank control and continued to feed on normal chow (Chow group) while other two groups were fed HFD (60% kcal fat as indicated, Beijing HuaFuKang Bioscience). HFD animals were gavaged with Bl. producta (HFD+Bl. producta group, 10⁹ CFU per animal per day) or an equal volume of YCFA medium (HFD group). Bodyweight was assessed weekly. After 4 weeks of treatment, mice were fasted overnight, anesthetized in chambers saturated with isoflurane, then sacrificed by cardiac puncture. Stools for metagenomic analysis were collected on the day before animals were euthanized. Blood samples were collected for estimation of serum levels of total triglyceride (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-c) and high-density lipoprotein cholesterol (HDL-c) by respective kits (Nanjing jiancheng Bioengineering Institute). Liver samples were weighed and snap-frozen in liquid nitrogen for sequential biochemical analysis.

2. To figure out whether dead Bl. producta cells possess lipid-lowering activity, 32 male C57BL/6J mice (8-week-old, 20–24 g, Vital River Laboratory Animal Technology) were divided into four groups with eight animals in each group. One group was used as blank control and continued to feed on normal chow (Chow group) while other groups were fed HFD. HFD animals were gavaged with live or pasteurized Bl. producta or equal volume of YCFA medium (HFD group). Bl. producta was daily administered by oral gavage at the dose 10⁹ CFU/0.2 mL suspended in sterile anaerobic PBS. An identical quantity of Bl. producta was inactivated by pasteurization for 30 min at 70°C and gavaged to mice at 0.2 mL per animal. Bodyweight was assessed weekly. After 4 weeks of treatment, mice were fasted overnight, anesthetized in chambers saturated with isoflurane, then sacrificed by cardiac puncture. Stools for metagenomic analysis were collected on the day before animals were euthanized. Cecal content was obtained from each mouse for 12-MMA and acetate quantification. Blood samples were collected for estimation of serum levels of total triglyceride (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-c) and high-density lipoprotein cholesterol (HDL-c) by respective kits (Nanjing jiancheng Bioengineering Institute) and for serum levels of 12-MMA and acetate.

3. To assess the anti-hyperlipidemic effect of 12-methylmyristic acid (12-MMA), 16 male C57BL/6J mice (8-week-old, 20-24 g) were randomly divided into two groups and fed HFD. The test group (HFD +12-MMA) was gavaged with 12-MMA (40 mg/kg, Sigma) once a day, while the HFD group was given an equal volume of solvent (1% tween 80 + 0.5% carboxymethyl cellulose sodium (CMC-Na, Sigma)) for 4 weeks. At the end of the experiment, mice were fasted overnight, anesthetized in chambers saturated with isoflurane and then sacrificed by cardiac puncture. Blood samples were collected for parameters evaluation via respective kits (kits for TG, TC, LDL-c, insulin, GLP-1 from Nanjing jiancheng Bioengineering Institute). Liver and fat samples were weighed and snap-frozen in liquid nitrogen for sequential biochemical analysis or fixed in 4% paraformaldehyde (Solarbio) for histological analysis.

Histology and immunohistochemistry analysis

The liver and fat tissues of each mouse were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into slides with a thickness of 4 μm. Liver and fat tissue sections were stained with hematoxylin and eosin (H&E) for histological analysis. For oil red O staining, liver tissues from the same liver lobe were cut into small pieces, then the frozen sections were rinsed in distilled water and stained with 0.2% oil red O (Sigma) and 60% 2-propanol (Sigma) for 10 min at 37°C. For immunohistochemistry analysis, fat slides were rinsed in 0.01 mol/L sodium citrate (pH 6.0) and heated for 20 min in a microwave to retrieve antigen. The sections were blocked in blocking buffer containing 5% goat serum, 2% BSA, 0.1% Triton X-100 and 0.1% sodium azide in PBS, then incubated overnight with anti-UCP1 (Cell Signaling) by a dilution of 1:100 at 4°C. After being washed twice in PBS, slices were incubated with secondary antibodies (Cell Signaling) for 1 h at room temperature. Slides were counterstained with H&E. All digital pictures were acquired using an EVOS X1 microscopy (Thermo Fisher Scientific).

Shotgun sequencing

Microbial genomic DNAs of fecal samples were extracted using QIAamp DNA Stool Mini kit (QIAGEN) and subjected to 1% agarose gel electrophoresis for evaluation. Concentration and purity of microbial DNA were determined with NanoDrop 2000 UV-vis spectrophotometer (Thermo Fisher Scientific) and Qubit 3.0 fluorometer (Thermo Fisher Scientific).

The gut microbial composition was determined by shotgun sequencing of the fecal samples collected from each mouse. Libraries were prepared using KAPA HyperPlus Library Preparation kit (KAPA Biosystems) and quantified by KAPA Library Quantification Kits (KAPA Biosystems) following the manufacturer’s instructions. Shotgun sequencing was performed on Illumina NovaSeq 6000 System (Illumina) at a depth of 5Gb for the experiment testing live and heat-killed Bl. producta. Cluster generation, template hybridization, isothermal amplification, linearization, blocking, denaturing and hybridization of the sequencing primers were performed according to the workflow indicated by Illumina.

The shotgun sequencing was analyzed by Beijing QuantiHealth Technology Co., Ltd. As described in previous work,⁴⁸ low-quality reads were removed from the raw data using MOCAT2. Sequencing adapters were removed by Cutadapt software (version v1.14, -m 30), then SolexaQA package was used to remove the reads with a threshold of less than 20 or the length of less than 30bp. The reads which could be aligned with the mouse genome (Mus musculus, GRCm38) were cleaned by using SOAP aligner software (v2.21, -M 4 -l 30 -v 10).

The relative abundance of bacteria was obtained using MetaPhlAn3 software. Based on the taxonomy information, α-diversity was calculated by R package vegan (2.5–6) package and presented by Shannon and Simpson indices. The principal coordinate analysis (PCoA) and nonmetric multidimensional scaling (NMDS) were calculated based on the Bray-Curtis distance using vegan (2.5–6). Linear discriminant analysis (LDA) Effect Size (LEfSe) method (https://github.com/biobakery/lefse) was used to identify species that show statistically significant differential abundances among groups.

12-MMA and acetate quantification

The content of 12-MMA in the conditioned culture medium of Bl. producta and in the cecal materials and serum of Bl. producta-treated animals was quantified by GC/MS. Preparation of fatty acid methyl esters and GC/MS analysis were performed as previously reported.⁴⁹ The pure 12-MMA was used as a standard reference. Fatty acid methyl esters were separated using an Elite 5-MS column and monitored by mass spectrometry (Clarus 500-SQ8S, Perkin Elmer, Courtaboeuf, France). The cecal and serum content of acetate was quantified by GC/MS analysis as previously reported.⁵⁰ N-methylbenzylamine-d0/d3 was applied as the chemical derivatization reagent to enhance the sensitivity and accuracy.

Quantitative real-time quantitative PCR (Qrt-PCR)

The inguinal fat tissues of mice used to verify the anti-hyperlipidemia effect of 12-MMA were collected to evaluate the mRNA levels of key genes involved in fat browning. Total RNA extraction, first-chain cDNA synthesis, and quantitative PCR assays were performed following the protocols described in the previous study.⁵¹ The primers used for each gene were listed in Supplementary Table S6.

Oral glucose tolerance test (OGTT)

As previously described,⁴⁷ before the OGTT test, mice were fasted for 6 h and then gavaged of 2 g/kg glucose. The blood glucose concentration in the tail vein was directly monitored at 0, 30, 60, 90 and 120 min after glucose administration (p.o.) using a glucose meter (Roche).

GPR120 activation assay

To evaluate the activating effect of 12-MMA on GPR120, we set up a fluorescence-based GPR120 activation assay using STC-1 intestinal endocrine cells (purchased from ATCC, CRL-3254) stably transfected with a fluorescent protein-based cAMP indicator Flamindo2.⁵² Cells were incubated with different concentrations of 12-MMA (Sigma) or DMSO (as a vehicle) for 20 min after transient transfection of GPR120-expressing plasmid. The fluorescence signal intensity (FL) was recorded by Tecan Infinite M1000Pro Microplate Reader at excitation 485 nm and emission 535 nm. The intensity of GPR120 activation was calculated as (FL_12-MMA-FL_DMSO)/FL_DMSO *100%.

STC-1 culture and cell work

STC-l enteroendocrine cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific), 100 U/ml penicillin and 100 µg/L streptomycin incubated at 37°C, 5% CO₂ following the routine procedures.

For Ca²⁺ analysis, a fluorescent probe Fluo-4-AM (MCE) was used to indicate the Ca²⁺ signal. Briefly, STC-1 cells were plated onto 96-well fluorescent plates and cultured for 18–24 h to reach ~ 90% confluence. Cells were washed three times with sterile PBS and then incubated with serum-free medium containing vehicle (DMSO) or 12-MMA with indicated concentrations for 3 h. After treatment, cells were washed three times with PBS and incubated with Fluo-4-AM (5 μM) at 37°C for 0.5 h. The fluorescent data at excitation 490 nm and emission 525 nm were acquired by Tecan Infinite M Nano Plate reader. The relative Fluo-4-AM intensity was calculated as (FL_12-MMA - FL_DMSO)/FL_DMSO *100%. For GLP-1 secretion studies, STC-1 cells were seeded onto 24-well plates and treated with 12-MMA, then supernatants were collected for evaluating GLP-1 secretion by GLP-1 kits (Nanjing jiancheng Bioengineering Institute) as the manufacturer’s instructions described.

Statistical analysis, graphing, and figure assembly

Data are presented as means ± SEM. SPSS 17.0 and Prism 7 (Graphpad) were used for statistical analysis. The significance of group differences was assessed by one-way ANOVA followed by Newman-Keuls post hoc tests for normally distributed data and Kruskal–Wallis test for non-normally distributed data. The corrected significance of multiple comparisons was assessed by the FDR method of Benjamini-Hochberg. P < 0.05 (pharmacological and metabolomic data) or adjusted P (q) < 0.1 (bacterial abundance) was considered statistically significant. All final figures were assembled using Adobe Illustrator.

Authors’ contributions

CW and BZ conceived the project, designed the experiments. WX and JY performed most of the experiments with the help of YY, ZL, YZ, FZ, YX and QW. CW and WX prepared figures and wrote the manuscript. BZ and ZL provided the gut bacteria samples. All authors discussed the manuscript, commented on the project, and contributed to manuscript preparation.

Disclosure statement

WX, ZL, YZ, QW, and BZ are employees of Beijing QuantiHealth Technology Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data availability statement

The sequencing raw data and genome sequences in the present study have been deposited in the National Center for Biotechnology Information (NCBI) database, and the project numbers are PRJNA771420 (https://www.ncbi.nlm.nih.gov//bioproject/PRJNA771420), PRJNA771488 (https://www.ncbi.nlm.nih.gov//bioproject/PRJNA771488) and PRJNA771491 (https://www.ncbi.nlm.nih.gov//bioproject/PRJNA771491). Other data relevant to the study are included in the article or uploaded as supplementary files.

Supplemental material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19490976.2023.2228045

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work was supported financially by the National Natural Science Foundation of China (Grant 81973217 to CW), and the Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (Grant 2018RC350014 to CW).