Capsaicin ameliorates glycemic levels via gut microbiota-derived 5-aminolevulinic acid in mice

Abstract

Background

Capsaicin, a natural alkaloid in chili peppers, regulates glycemic levels; however, its mechanisms and therapeutic potential remain unclear. This study aimed to elucidate the role of gut microbiota and their metabolites in mediating capsaicin’s glycemic regulatory effects. We conducted experiments in specific pathogen-free (SPF) and germ-free (GF) mice, transient receptor potential vanilloid 1 (TRPV1) receptor ablation studies, and fecal microbiota transplantation (FMT) to demonstrate the involvement of gut microbiota in capsaicin-mediated glycemic control. Metagenomics and metabolomics analyses were employed to identify key microbial strains and metabolic pathways. Keystone strains and metabolites were supplemented in GF mice without capsaicin intervention to validate their effects on glycemic regulation. In vitro co-culture experiments were performed to investigate the mutualistic relationships among keystone strains under capsaicin treatment.

Results

Gut microbiota constitute an important component of capsaicin-mediated glycemic regulation, acting in concert with but not solely dependent on TRPV1 signaling. Gut microbiota altered by capsaicin promote the production of 5-aminolevulinic acid (5-ALA), which contributes to heme synthesis and enhances glycemic control. Supplementation with Akkermansia muciniphila, Ligilactobacillus murinus, or 5-ALA in GF mice recapitulates the glycemic benefits of capsaicin. Furthermore, capsaicin enriches Akkermansia muciniphila, which in turn supports the growth of Ligilactobacillus murinus.

Conclusion

Capsaicin-induced changes in the gut microbiota promote 5-ALA synthesis, leading to improved glycemic control. These findings suggest that dietary or probiotic interventions targeting gut microbiota, particularly Akkermansia muciniphila and 5-ALA, may offer promising strategies for managing glycemic disorders, including type 2 diabetes (T2D).

Download video file ^{(46.2MB, mp4)}

Video Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s40168-026-02415-8.

Background

Hyperglycemia is a major risk factor for severe metabolic disorders, including T2D and cardiovascular diseases, affecting millions globally and posing substantial healthcare challenges [1, 2]. Despite advances in glycemic management, effective, accessible interventions remain urgently needed. Studies have elucidated that widely prescribed glycemic control agents, including metformin and acarbose, elicit profound alterations in the composition and functionality of the gut microbiota [3, 4]. However, the comprehensive health implications of these microbiota-associated changes remain inadequately characterized and warrant further investigation. Notably, metformin has been implicated in adverse clinical outcomes when utilized for glycemic regulation alongside malignancies [5], and the hypoglycemic efficacy of acarbose has been found to be inhibited by certain gut microorganisms, such as Klebsiella grimontii TD1 [6], highlighting potential therapeutic complexities.

Diet is a major determinant of glycemic control and gut microbiota composition, with dietary interventions influencing glycemic metabolism through coordinated modulation of microbial composition and function [7]. While dietary fibers and polyphenols have been extensively studied for their roles in glycemic regulation, the mechanisms by which bioactive compounds like capsaicin modulate glycemic levels via gut microbiota remain poorly understood [8–11]. A large-scale nationwide study encompassing 200 million individuals showed that the intake of spicy foods was negatively correlated with the risk of diabetes [12], highlighting the link between peppers and glycemic levels.

Capsaicin, the main active constituent in chili peppers, has a molecular structure that includes a van der Waals tail and a polar amide section, which allows it to bind to receptors in nerve cells (TRPV1) to convey pain and heat sensations [13, 14]. Capsaicin plays a role in glycemic metabolism by activating the TRPV1 receptor, including calcium influx into islet β cells, stimulating insulin release [15], promoting glucagon-like peptide-1 (GLP-1) and insulin secretion [16]; and improving insulin sensitivity by activating the adenosine monophosphate-activated protein kinase (AMPK) pathway and increasing glucose transporter type 2 (GLUT2) transport [17]. Notably, these TRPV1-mediated effects on glycolipid metabolism still persist in TRPV1 knockout mice, suggesting that missing pieces have been explored further [18].

The gut microbiota plays a key role in glycemic regulation with specific gut bacteria, including Akkermansia muciniphila, Lactobacillus spp., Bifidobacterium spp., Faecalibacterium prausnitzii, and Roseburia spp. Among these, Akkermansia muciniphila and Lactobacillus spp. have been extensively implicated in glycemic regulation, with documented roles in metabolic homeostasis and host–microbiota interactions [19, 20]. Capsaicin consumption has been shown to promote the proliferation of gut microbes associated with glycemic regulation, including Akkermansia, Lactobacillus, Faecalibacterium, and Bifidobacterium, and decrease the abundance of inflammation-associated bacteria within the Firmicutes and Bacteroidetes, notably Escherichia and Bacteroides [21]. However, the existing studies lacked detailed insights into keystone species changes responding to capsaicin and metabolism by which capsaicin influences glycemic regulation. Here, we hypothesize that, beyond the TRPV1 receptor, gut microbiota alterations mediate the effects of capsaicin intervention on controlling glycemic levels, though the underlying mechanisms remain poorly understood.

In this study, using GF mice, TRPV1-ablated mice, and FMT, we demonstrated that capsaicin ameliorates glycemic levels in a gut microbiota-dependent manner. The overall alterations in the gut microbiota and metabolites under capsaicin intervention were further identified with metagenomics and untargeted metabolomics. Remarkably, capsaicin alters the abundance of Akkermansia muciniphila and Ligilactobacillus murinus, resulting in the accumulation of 5-ALA in the host and the regulation of glycemic levels through the heme pathway. Mechanistically, capsaicin promotes mucus layer expansion, supporting the growth of Akkermansia muciniphila and its symbiont Ligilactobacillus murinus, which enhances host glycine levels and stimulates the synthesis of 5-ALA. This novel mechanism links capsaicin-induced gut microbiota changes to improved glycemic levels through heme biosynthesis.

Material and methods

Animal models

Male SPF mice (6–8 weeks old, 18–20 g; Beijing Vital River Laboratory Animal Technology Co., Ltd.) and GF mice (6–8 weeks old, 18–20 g; Laboratory Animal Center, Jiangnan University) were utilized for the experiments. The mice were acclimated to standard environmental conditions (21–25 °C temperature; 40–60% humidity) under a 12-h light-dark cycle for 1 week. During the adaptation period, standard commercial rodent chow and sterile water were provided ad libitum. The experimental protocols were approved by the Ethics Committee of Jiangnan University and the Ethics Committee of the Jiangsu Institute of Parasitic Diseases (Wuxi, China). All procedures complied with the European Community Directive 2010/63/EU.

Animal experiment 1: SPF mice with capsaicin intervention under a H(S)FD

To evaluate the regulatory effects of capsaicin on glycemic metabolism disorders, 60 male SPF C57BL/6J mice were randomly assigned to six experimental groups: the placebo obese group (obesity/P), low-dose capsaicin obese group (obesity/L), high-dose capsaicin obese group (obesity/H), placebo diabetic group (T2D/P), low-dose capsaicin diabetic group (T2D/L), and high-dose capsaicin diabetic group (T2D/H). Each group included 10 mice. The study spanned a total duration of 12 weeks. Week 0 was designated as an adaptation period, during which mice were provided with a standard diet and ad libitum access to water. Placebo or capsaicin was administered daily from weeks 1 to 12. From weeks 5 to 12, dietary interventions were introduced to induce obesity using a (high-fat diet) HFD or T2D using a (high-sugar diet) HSD. At week 10, mice in the diabetic groups received intraperitoneal injections of streptozotocin (STZ). All mice were provided with ad libitum access to food and water throughout the experiment, with weekly monitoring of dietary intake, water consumption, and weight (JN.No20211030c1240125[408]).

Animal experiment 2: GF mice with capsaicin intervention under a H(S)FD

Since the effect of high dose of capsaicin on glycemic regulation was observed more significantly in SPF mice, in the subsequent experiments, we focused exclusively on the effects of high-dose capsaicin. To investigate the role of gut microbiota in the regulation of glycemic levels by capsaicin in mice, 16 male GF C3H mice were randomly assigned to four experimental groups: the placebo obese group (obesity/P), high-dose capsaicin obese group (obesity/H), placebo diabetic group (T2D/P), and high-dose capsaicin diabetic group (T2D/H). Each group included 4 mice. The study spanned a total duration of 8 weeks. Week 0 was designated as an adaptation period, during which mice were provided with a standard diet and ad libitum access to water. From weeks 1 to 8, dietary interventions were introduced to induce obesity using a HFD or T2D using a HSD. At week 6, mice in the diabetic groups received intraperitoneal injections of STZ (JN.No20230515c0300730[206]).

Animal experiment 3: TRPV1 receptor-ablated mice with capsaicin intervention under a H(S)FD

Given the reported severe disruption of gut microbiota by resiniferatoxin (RTX) [22], we first performed TRPV1 receptor ablation experiments in GF mice. Subsequently, fecal samples were collected from 10 male SPF C3H mice, and gut microbiota was transplanted into TRPV1 receptor-ablated GF mice via FMT. These mice were then subjected to further modeling. Twenty-four male GF C3H mice were randomly assigned to four experimental groups: the placebo obese group (obesity/P), high-dose capsaicin obese group (obesity/H), placebo diabetic group (T2D/P), and high-dose capsaicin diabetic group (T2D/H). Each group included 6 mice. The study spanned a total duration of 12 weeks. Week 0 was designated as an adaptation period, during which mice were provided with a standard diet and ad libitum access to water. Resiniferatoxin was administered intraperitoneally during the first 3 days of week 1, and FMT was performed on the first 3 days of week 3. From weeks 5 to 12, dietary interventions were introduced to induce obesity using a HFD or T2D using a HSD. At week 10, mice in the diabetic groups received intraperitoneal injections of STZ (JN.No20230830i0401020[343]).

Animal experiment 4: FMT mice with capsaicin-induced gut microbiota under a H(S)FD

To assess whether capsaicin-induced gut microbiota can regulate glycemic levels, 16 male GF C3H mice were randomly assigned to four experimental groups: the placebo obese group (obesity/P), high-dose capsaicin obese group (obesity/H), placebo diabetic group (T2D/P), and high-dose capsaicin diabetic group (T2D/H). Each group included 4 mice. The study spanned a total duration of 10 weeks. Week 0 was designated as an adaptation period, during which mice were provided with a standard diet and ad libitum access to water. FMT (fecal samples from SPF mice with 4 weeks of capsaicin intervention) was performed on the first 3 days of week 1. From weeks 3 to 10, dietary interventions were introduced to induce obesity using a HFD or T2D using a HSD. At week 8, mice in the diabetic groups received intraperitoneal injections of STZ (JN.No20231007c0161220[469]).

Animal experiment 5: key bacteria colonized GF mice under a H(S)FD

To assess whether capsaicin-induced key bacteria can regulate glycemic levels, 24 male GF C3H mice were randomly assigned to six experimental groups: the obese group (obesity), Akkermansia muciniphila and Ligilactobacillus murinus colonized obese group (obesity/AKK-LM), Lactobacillus johnsonii colonized obese group (obesity/LJ), diabetic group (T2D), Akkermansia muciniphila and Ligilactobacillus murinus colonized diabetic group (T2D/AKK-LM), and Lactobacillus johnsonii colonized diabetic group (T2D/LJ). Each group included 4 mice. The study spanned a total duration of 10 weeks. Week 0 was designated as an adaptation period, during which mice were provided with a standard diet and ad libitum access to water. Monocolonization was performed on the first 3 days of week 1. From weeks 3 to 10, dietary interventions were introduced to induce obesity using a HFD or T2D using a HSD. At week 8, mice in the diabetic groups received intraperitoneal injections of STZ (JN.No20231215c0160224[606]).

Animal experiment 6: supplementation of 5-ALA in GF mice under a H(S)FD

To assess the 5-ALA’s regulatory effects on glycemic levels, 16 male GF C3H mice were randomly assigned to four experimental groups: the obese group (obesity), 5-ALA obese group (obesity/5-ALA), diabetic group (T2D), and 5-ALA diabetic group (T2D/5-ALA). Each group included 4 mice. The study spanned a total duration of 8 weeks. Week 0 was designated as an adaptation period, during which mice were provided with a standard diet and ad libitum access to water. 5-ALA was administered daily from weeks 1 to 8. From weeks 1 to 8, dietary interventions were introduced to induce obesity using a HFD or T2D using a HSD. At week 6, mice in the diabetic groups received intraperitoneal injections of STZ (JN.No20231215c0160224[606]).

Animal experiment 7: supplementation of capsaicin in GF mice under a normal diet

Given that the observed capsaicin-induced increases in the intestinal mucus layer were primarily inferred from experiments in SPF mice, FMT, or gnotobiotic colonization models, an additional experiment was performed to directly assess the effect of capsaicin on mucus layer thickness while minimizing microbial interference. Germ-free mice were subjected to capsaicin intervention following an adaptation period, receiving oral gavage at a dose of 30 mg/kg. Six male GF C57BL/6J mice were randomly assigned to two experimental groups: the placebo group and the high-dose capsaicin group. Each group included 3 mice. Thirty minutes after oral gavage, mice were sacrificed [30], and colonic tissues were collected and processed for Alcian blue–periodic acid–Schiff (AB–PAS) staining to enable direct morphological measurement of mucus layer thickness (GPTAP20260127-3).

GF mice and treatment

The GF mice were housed in sterile isolators. The isolators and all internal equipment (such as cages, racks, scissors, and forceps) were sterilized via high-pressure hydrogen peroxide and acetic acid (5:4) spraying before being transferred into the isolator. After 48 h, sterile air was introduced through a filtration system. Feed, bedding, and drinking water were sterilized by gamma irradiation at 100 kGy, while electronic devices (e.g., electronic scales, glucose meters, and gavage equipment) were sterilized using ethylene oxide. Experimental reagents were filtered through a 0.22-μm sterile filter in a laminar flow hood. After sterilization, all materials were passed through a transfer chamber, sprayed, and left to stand for 24 h before being transferred for use in the experiment.

Culture and administration of strains

Akkermansia muciniphila (CCFM, FSDLZ36M5) was cultured in BHI medium containing mucin, and Ligilactobacillus murinus (CCFM, FNXYCHL79T3) and Lactobacillus johnsonii (CCFM, FJLHD57L1) were cultured in MRS medium (Supplemental Table S2) at 37 °C in an anaerobic workstation with a gas mix containing 10% hydrogen, 10% carbon dioxide, and 80% nitrogen. Any bacterial stock obtained from a −80 °C freezer (cryopreserved in 30% glycerol PBS) was first activated through two generations of liquid culture, with the bacterial density reaching 10⁹ CFU/mL before being used in subsequent experiments. The concentration of strains was calculated by measuring the absorbance at the wavelength of 600 nm.

Metagenomic sequencing

Genomic DNA was extracted from fecal samples of SPF mice after 4 weeks of capsaicin intervention using a fecal DNA extraction kit. The purity and integrity of the DNA were analyzed by 1% agarose gel electrophoresis (AGE), while DNA quantification was performed using the Qubit® dsDNA Assay Kit with a Qubit® 2.0 Fluorometer (Life Technologies, CA, USA). Appropriate sample volumes were transferred to centrifuge tubes and diluted with sterile water to achieve an OD value between 1.8 and 2.0. A total of 1 μg of genomic DNA was used for library construction with the NEBNext® Ultra™ DNA Library Prep Kit for Illumina (New England Biolabs). DNA was randomly fragmented to approximately 350 bp using a Covaris ultrasonicator, followed by end repair, A-tailing, adapter ligation, purification, and PCR amplification to complete library preparation. The constructed library was initially quantified with a Qubit® 2.0 Fluorometer and diluted to 2 ng/μL. The insert size was assessed using an Agilent 2100 Bioanalyzer, and once the insert size met expectations, the library’s effective concentration was accurately quantified by qPCR (effective concentration > 3 nM) to ensure quality. Libraries passing quality control were pooled based on effective concentrations and target sequencing data requirements, followed by Illumina PE150 sequencing. Library preparation and sequencing were conducted by Novogene Co., Ltd. (Beijing, China).

Metagenomic data analysis

Following high-throughput sequencing, the raw sequences were processed as follows: quality-filtered metagenomic data were subjected to taxonomic profiling using MetaPhlAn2 v4.0.2. Functional analysis was subsequently performed with HUMAnN3 v3.0.0 under default settings in UniRef90 mode, yielding the relative abundances of gut microbiota, genes, and metabolic pathways for each metagenome. The MetaCyc results were further mapped to corresponding pathway information in the KEGG database.

Abundance ratio plots were generated using R version 4.4.2, visualizing the relative abundances of taxa at various classification levels on a circular axis by wrapping the coordinates into a ring format. Stamp plot was applied to perform ANOVA and multiple comparison corrections, identifying gut microbes with significant differences across samples. PCA was conducted to assess the distribution of gut microbiota among groups, aiming to reveal potential ecological patterns and community structures. Sankey diagrams were used to depict the flow relationships between gut microbial genes and metabolic pathways influenced by capsaicin. For all differential analyses, significance was determined by p < 0.05, q < 0.05, VIP > 1, and consideration of 95% confidence intervals (CIs).

Untargeted metabolomics with LC-MS

Samples for untargeted metabolomics included fecal and colonic tissue samples from SPF mice after 4 weeks of capsaicin intervention. Liquid medium samples or homogenized fecal and tissue samples were centrifuged at 4 °C for 5 min at 10,000 × g. A 100 μL aliquot of the supernatant was mixed with 400 μL of methanol (1:1, v/v) for protein precipitation. Samples were vortexed, ultrasonicated on ice, and incubated at −20 °C for 1 h to enhance protein precipitation efficiency. Following centrifugation at 15,000 rpm for 15 min at 4 °C, the supernatant was evaporated to dryness using a rotary evaporator. Before analysis, the residue was reconstituted in 100 μL of acetonitrile (1:1, v/v) and centrifuged at 15,000 rpm for 15 min at 4 °C. An appropriate volume of the supernatant was transferred into vials for analysis. Quality control (QC) samples were prepared by pooling equal volumes of all test samples into a new vial and mixing thoroughly. Blank controls were prepared by adding the reconstitution solvent (acetonitrile = 1:1, v/v) to a new vial.

Chromatographic conditions: A Waters ACQUITY UPLC HSS T3 column (1.8 μm, 2.1 × 100 mm) was used with a flow rate of 0.3 mL/min. The mobile phase for positive ion mode consisted of 0.1% formic acid in water (phase A) and acetonitrile (phase B). For negative ion mode, phase A contained 5 mM ammonium acetate, and phase B was acetonitrile. The gradient elution program was as follows: 0–1 min, 2% B; 1–10 min, 2–98% B; 10–12 min, 98% B; 12–12.1 min, 98–2% B; 12.1–15 min, 2% B. The injection volume was 2 μL. Samples were maintained at 15 °C in the autosampler, and the column temperature was set to 35 °C. Mass spectrometry conditions: Both positive and negative ion modes used an electrospray ionization (ESI) source with polarity switching. The spray voltage was set to 3.5 kV, the capillary temperature to 320 °C, sheath gas flow rate to 35 arb, auxiliary gas flow rate to 15 arb, and drying gas temperature to 320 °C. The acquisition mode was data-dependent full-scan MS/dd-MS2.

Primary MS parameters: mass scan range of m/z 70–1050, resolution of 70,000, automatic gain control (AGC) target of 1 × 10⁶, and maximum ion injection time (IT) of 100 ms. Secondary MS parameters: resolution of 17,500, normalized collision energy (NCE) at 20, 40, and 60 eV, AGC target of 5 × 10⁴, maximum IT of 50 ms, dynamic exclusion time of 10 s, and selection of the top 7 most intense ions for fragmentation.

Metabolomic data analysis

MS data were processed by Compound Discovery 3.1 (Thermo Fisher Scientific, MA, USA) using untargeted metabolic workflow with default parameters, which included data extraction, background features filtered out, peak identification, deconvolution, alignment and integration, and metabolite identification assignment via mzCloud, HMDB, KEGG, and ChemSpider databases. The data processing filtrated out metabolic features that appeared <50% of the QC samples and that >30% RSD of the QC samples, and the mass tolerance was 5 ppm. Metabolic features whose m/z ratios could not match the masses in the above databases were filtrated out. The processed data, including peak area, retention time (RT), molecular weight (MW), and identified or unknown compounds, were integrated. Multivariate analysis was performed in SIMCA software (Umetrics, Umea, Sweden). Metabolic pathway analysis was performed in MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/).

Metabolomic and metagenomic data correlation analysis

A correlation analysis was performed between metagenomic and metabolomic data, focusing on genomes and metabolites that were identified as significantly changed in SPF mice under 4 weeks of capsaicin intervention. The genome relative abundance and raw metabolite values were log-transformed, and Spearman’s correlation coefficients assessment was conducted using R version 4.4.2. Correlation analysis of the metagenomic (p < 0.05) and metabolomic data (p < 0.05, q < 0.05, VIP > 1) was analyzed.

Capsaicin treatment

In animal experiments 1, 2, and 3, mice received daily oral gavage of capsaicin. Due to the poor water solubility of capsaicin, a preliminary study was conducted to determine the minimal volume of ethanol required to dissolve capsaicin. Tween 80 was also added as a solubilizing agent to prepare the gavage solution. The preparation details are as follows: placebo solution, a mixture of 1% ethanol and 2% Tween 80 in 0.9% saline; low-dose capsaicin solution: capsaicin at 6 mg/kg was dissolved in 1% ethanol and 2% Tween 80 in 0.9% saline; high-dose capsaicin solution: capsaicin at 30 mg/kg was dissolved similarly in 1% ethanol and 2% Tween 80 in 0.9% saline. The gavage solutions were freshly prepared before use, with dosing adjusted based on body weight (0.1 mL per 10 g).

For the induction of T2D using STZ, the solution was prepared by dissolving STZ in 0.10 mol/L sodium citrate buffer (pH 4.2–4.5) to achieve a concentration of 10.00 mg/mL. The solution was sterilized using a 0.22-μm filter, prepared under light protection, stored at low temperature, and used immediately after preparation.

Ablation of TRPV1 receptor

The TRPV1 receptor ablation in GF mice was performed following a previously reported protocol [22]. RTX was dissolved in 0.9% saline containing 1% ethanol and 2% Tween-80, freshly prepared before use. Intraperitoneal injections were administered over three consecutive days with escalating doses: 30 μg/kg on the first day, 70 μg/kg on the second day, and 100 μg/kg on the third day.

Immunofluorescence

Following euthanasia, the mouse was placed in a prone position on a dissection board, with its limbs secured to allow access to the dorsal spinal region. Using dissecting scissors, an incision was made along the midline of the spine, extending from the neck to the tail (approximately 1.5–2 cm) to expose the underlying muscle layer. The muscle surrounding the spine was gently separated using forceps and scissors, progressively exposing the vertebral column. Once the spine was fully exposed, bone-cutting scissors or fine surgical scissors were used to sever the vertebrae bilaterally, enabling the removal of the entire spinal column. The overlying lamina was carefully excised using micro scissors and forceps to expose the spinal cord and its dorsal root nerves. The L5–L6 dorsal root nerve ganglia (DRG) were precisely located and severed from their connecting dorsal roots and spinal cord. Fine forceps were then used to delicately extract the DRG from the spinal column. The isolated DRG tissues were immediately transferred to pre-chilled saline or an appropriate culture medium to preserve tissue viability for subsequent immunofluorescence analysis.

The collected DRG tissues were fixed in 4% paraformaldehyde phosphate-buffered saline (PBS) at 4 °C for 4 h, followed by three washes in PBS. The tissues were then dehydrated overnight at 4 °C in 30% sucrose dissolved in PBS. After dehydration, the tissues were embedded in Tissue-Tek OCT compound (Sakura), flash-frozen, and stored at −80 °C until sectioning. Using a cryostat (Leica Instruments), 10-μm thick sections were prepared and mounted onto SuperFrost Plus slides (VWR). The slides were stored at −20 °C until immunostaining. For staining, the slides were brought to room temperature and rinsed with PBS to remove residual OCT. The tissue sections were blocked for 30 min in PBS containing 5% normal goat serum (Jackson Immunoresearch), 5% normal donkey serum (Jackson Immunoresearch), and 0.1% Triton X-100 (Sigma-Aldrich). Primary antibody staining was performed overnight at 4 °C using rabbit anti-phospho-p44/42 MAPK (Thr202/Tyr204) monoclonal antibody (CST), diluted in blocking buffer. Following three PBS washes, the sections were incubated at room temperature for 1 h with TRPV1 Rabbit pAb. After another three PBS washes, the sections were counterstained with DAPI (Invitrogen) for 5 min and rinsed with PBS. The stained sections were imaged using a Nikon Eclipse Ti inverted microscope. The immunofluorescence slide preparation and antibody incubation were conducted by Servicebio Technology Co., Ltd. (Wuhan, China).

Growth assay of key strains

To assess the impact of capsaicin on the growth of key strains, capsaicin (0.02 g/L) was added to the respective liquid media, with media lacking capsaicin serving as controls. The wells of a 96-well plate were filled with these mixtures and sealed with Microseal B plate-sealing film (Bio-Rad, USA). The plate was maintained at 37 °C, and the OD600 was monitored every 2 h using a Powerscan HT plate reader (Dainippon Sumitomo Pharma, Japan).

Determination of glycemic levels

Fasting blood glucose (FBG) levels at specific time points in mice were measured using a glucometer with corresponding test strips. Mice were fasted for 12 h prior to measurement. Blood was collected from the tail of SPF mice using a lancet, while tail clipping was employed for GF mice due to handling challenges. To minimize stress-induced variability in glucose levels, mice were handled carefully during routine procedures. Blood glucose measurements were taken using subsequent drops of blood rather than the first drop whenever possible.

Mice were fasted for 12–16 h prior to the oral glucose tolerance test (OGTT). A glucose solution (2 g/kg) was prepared and administered via gavage according to the body weight. Blood glucose levels were measured at baseline, 0 min, 30 min, 60 min, 90 min, and 120 min. Glucose tolerance was assessed by calculating the area under the curve (AUC).

Enzyme-linked immunosorbent assay (ELISA)

Five centimeters of distal colons were harvested from mice and homogenized by sonication in PBS on ice. After sonication, supernatants were purified by centrifugation and used for ELISA analysis. FINS (fasting insulin), TNF-α (tumor necrosis factor-alpha), IL-1β (interleukin-1 beta), IL-6 (interleukin-6), MUC2 (mucin 2), GlyT1 (glycine transporter 1), and HO-1 (heme oxygenase-1) were detected using ELISA kits (SenBeiJia Biological Technology Co., Ltd., Nanjing, China). All procedures were carried out according to the manufacturer’s instructions.

Biochemical assays in mice

TC (total cholesterol), TG (triglycerides), LDL-C (low-density lipoprotein cholesterol), and HDL-C (high-density lipoprotein cholesterol) levels in mouse serum were measured using endpoint analysis on a Mindray BS-360S automated biochemical analyzer. After centrifugation, mouse serum was diluted at least three times, and corresponding reagent kits were used for analysis on the automated analyzer.

Histological examination

During mouse dissection, approximately 0.5 cm of distal colon tissue was collected and fixed in 4% paraformaldehyde solution in the dark for 24–48 h. Subsequent sectioning and staining procedures were performed by Wuhan Servicebio Technology Co., Ltd. The prepared tissue sections were scanned and imaged using a Pannoramic MIDI II digital slide scanner, and images were processed with SlideViewer software.

Determination of content of amino acid

The Agilent 1100 HPLC system (Agilent Technologies, USA) was used to determine the levels of free amino acids in mouse tissues, feces, and key strains in vitro culture media. A sample weighing 0.5–0.7 g was combined with 1 mL of ultrapure water and homogenized using magnetic beads. Subsequently, 5% trichloroacetic acid was added to precipitate proteins. The supernatant was collected and transferred into sample vials for analysis.

The chromatographic conditions were as follows: mobile phase A (pH 7.2) consisted of 27.6 mmol/L sodium acetate-triethylamine-tetrahydrofuran (volume ratio 500:0.11:2.5), and mobile phase B (pH 7.2) consisted of 80.9 mmol/L sodium acetate-methanol-ethyl acetate (volume ratio 1:2:2). The column used was an Agilent Hypersil ODS (5 μm, 4.0 mm × 250 mm). A gradient elution was applied with the following program: 0 min, 8% B; 17 min, 50% B; 20 min, 100% B; 24 min, 0% B. The flow rate was set at 1.0 mL/min, and the column temperature was maintained at 40 °C. Detection was performed using a Variable Wavelength Detector (VWD) at 338 nm, with proline detected at 262 nm. Amino acid concentrations were quantified using an external standard method.

Determination of capsaicin with HPLC

To investigate whether capsaicin is metabolized into other components in the gut, we measured the capsaicin content in the feces of SPF mice after gavage of capsaicin 2 h. Preparation of the standard curve: capsaicin standard (HPLC grade, purity ≥ 99.0%) was used to prepare a stock solution at 1 mg/mL. Gradient concentrations of working solutions (0, 2.5, 5, 10, 12.5, 25, and 50 μg/mL) were freshly prepared for immediate use. Sample pretreatment: a total of 2 g of mouse feces was mixed with 15 mL of methanol, and capsaicin was extracted using ultrasound-assisted extraction. The ultrasound conditions were set as follows: temperature 50 °C, amplitude 65% (200 W), and extraction time 15 min. The extract was filtered through a 0.45-μm membrane before analysis. Liquid chromatography conditions: capsaicin quantification was performed using a HPLC system (Agilent 1200, Agilent Technologies, USA) with a reversed-phase column (Agilent TC-C18, 4.6 mm × 250 mm, 5 μm). The injection volume was 20 μL, and detection was conducted at 280 nm using a UV detector. A gradient elution program was employed with water (mobile phase A) and acetonitrile (mobile phase B) at a flow rate of 1.0 mL/min at 30 °C. The gradient settings were as follows: 0 min, 45% B; 25 min, 65% B; 30 min, 100% B; and 35 min, 45% B. Capsaicin was identified and quantified using external standards. Each extract was analyzed in triplicate.

The quantification of SCFAs with GC-MS

Fecal samples from SPF mice following 4 weeks of capsaicin intervention were used to determine the concentration of SCFAs using the GC-MS platform. A total of 20–50 mg of freeze-dried fecal sample was weighed and mixed with 1 mL of saturated sodium chloride solution and two grinding beads. The mixture was homogenized at 65 Hz for 40 s using a tissue grinder. After centrifugation at 12,000 rpm for 10 min, the supernatant was collected and acidified with 20 μL of sulfuric acid. For SCFA extraction, 1 mL of diethyl ether was added, followed by vortexing for 30 s. The mixture was centrifuged at 14,000 × g for 15 min, and the supernatant was mixed with 0.25 g of anhydrous sodium sulfate and left to stand for 30 min. The upper organic phase was collected, centrifuged at 12,000 rpm, and transferred to an injection vial for immediate analysis on the GC-MS platform.

Chromatographic conditions: A GC-MS system (QP2010, Shimadzu, Japan) equipped with an Rtx-5MS capillary column (0.25 μm, 30 m × 0.25 mm) was used to quantify SCFAs in mouse fecal samples. The oven temperature was programmed as follows: initial temperature of 100 °C, ramped at 7.5 °C/min to 140 °C, then at 60 °C/min to 200 °C, and held for 3 min. High-purity helium was used as the carrier gas at a flow rate of 0.89 mL/min with a split ratio of 10:1. The ion source and transfer line temperatures were maintained at 220 °C and 250 °C, respectively. SCFAs were quantified using an external standard method with a mixture of standard compounds, including acetic acid, propionic acid, butyric acid, valeric acid, isobutyric acid, and isovaleric acid.

RNA preparation and real-time quantitative polymerase chain reaction

RNA was extracted using an RNA extraction kit following the manufacturer’s protocol (Vazyme Biotech Co., Ltd., Nanjing, China). After obtaining high-quality RNA, its concentration was measured, and total RNA was reverse-transcribed into cDNA using a reverse transcription kit. The cDNA was then diluted appropriately based on the target gene expression levels and subjected to quantitative real-time PCR (qPCR) using SYBR Green Supermix on a BioRad CFX384 system. The qPCR reaction was carried out in a 10 μL total volume consisting of 0.5 μL forward primer, 0.5 μL reverse primer, 5 μL SYBR Green Supermix, 3 μL water, and 1 μL cDNA template. The cycling conditions were as follows: initial denaturation at 95 °C for 30 s, followed by 40 cycles of denaturation at 95 °C for 10 s, annealing at 60 °C for 30 s. A melting curve analysis was performed with the following steps: 95 °C for 5 s, 65 °C for 5 s, and 95 °C with a 0.5 °C increment per cycle while continuously detecting the fluorescence signal. The expression levels of target genes were normalized to the β-actin expression levels within the same samples (Supplemental Table S3).

Fecal microbiota transplantation

FMT and monocolonization were involved in animal experiments 3, 4, and 5. FMT sample collection: fresh feces from mice were collected into sterilized EP tubes (0.25 g per tube) and mixed with 1 mL of 20% glycerol. Samples were stored at −80 °C. For FMT, fecal suspensions were thawed and activated at 37 °C for 10 min, followed by low-speed centrifugation (600 rpm, 5 min) to collect the supernatant. The suspension was resuspended in 0.9% saline and centrifuged repeatedly under the same conditions until a total of 2 mL of supernatant was obtained. This yielded a fecal suspension concentration of approximately 0.125 g/mL. The fecal suspension was orally gavaged into GF mice under sterile conditions for three consecutive days to complete the FMT procedure. For monocolonization, liquid cultures of the bacterial strain grown to the logarithmic phase were centrifuged at 35,000 rpm, and the resulting bacterial pellet was resuspended in sterile saline for oral gavage in monocolonization.

Data visualization

Bioinformatic analysis was partly performed using the OmicStudio tools at https://www.omicstudio.cn/tool. All other figures were generated by R software 4.4.2, using ggplot2 and Complex Heatmap packages, or by GraphPad Prism 10.2.0.

Quantification and statistical analysis

Data were shown as mean ± SEM. Data analysis and visualization were performed using GraphPad Prism 10.2.0. Normality and lognormality tests were conducted to assess data distribution prior to analysis. For data meeting normality and homoscedasticity assumptions, one-way ANOVA followed by Tukey’s multiple comparisons test was used to determine significant differences. For normally distributed data without homoscedasticity, Brown-Forsythe and Welch ANOVA tests with Dunnett’s T3 multiple comparisons test were applied. For non-normally distributed data, Kruskal-Wallis nonparametric analysis followed by Šídák’s test was employed. Significance was defined as p < 0.05, with p < 0.05, p < 0.01, and p < 0.001 denoted as *, **, and ***, respectively. Comparisons among multiple groups were analyzed using one-way ANOVA, while differences between two groups were assessed with t-tests. Longitudinal analysis over time utilized two-way ANOVA.

Comprehensive details regarding reagents, strains, chemicals, animal models, software, and additional materials are available in the supplemental materials.

Results

Capsaicin mitigates glycemic disturbances through gut microbiota modulation

Given the frequent association of capsaicin intake with glycemic regulation rather than direct therapeutic effects, we propose the hypothesis that capsaicin mitigates glycemic disturbances caused by a HFD or a HSD. The obesity and T2D models were established in SPF and GF mice (Supplemental Table S1; Fig. 1A, B). To investigate this effect, capsaicin was administered for 1 month prior to HFD or HSD induction in SPF mice to precondition the gut microbiota to a capsaicin-modulated state.

Fig. 1.

The gut microbiome-dependent effect of capsaicin on glycemic regulation. A After 4 weeks of capsaicin or placebo intervention, HFD or HSD diet was additionally added for the establishment of obesity or T2D models in SPF mice, respectively. B Simultaneously with capsaicin or placebo intervention, HFD or HSD diet was additionally added for the establishment of obesity or T2D models in GF mice, respectively. C, E FBG and OGTT of the obese SPF mice across three treatment groups, including placebo (P), low dose of capsaicin (L), and high dose of capsaicin (H). D, F FBG and OGTT of the obese GF mice across two treatment groups, including placebo (P) and high dose of capsaicin (H). G, I FBG and OGTT of the diabetic SPF mice across three treatment groups, including placebo (P), low dose of capsaicin (L), and high dose of capsaicin (H). H, J FBG and OGTT of the diabetic GF mice across two treatment groups, including placebo (P) and high dose of capsaicin (H). Data are mean with SEM. * < 0.05, ** < 0.01, and *** < 0.001, based on Student’s unpaired t-test, ordinary one-way or two-way ANOVA followed by Tukey’s (Gaussian distributed, equal SDs) post hoc test, and ordinary two-way ANOVA followed by Šídák’s (non-Gaussian distributed) post hoc test. ig, intragastric gavage; H(S)FD, high-fat diet or high-sugar diet; P, placebo (solvent of capsaicin); L, low dose of capsaicin (6 mg/kg); H, high dose of capsaicin (30 mg/kg); ns, not significant

The results showed that the capsaicin intervention significantly ameliorated glycemic levels in SPF mice (Fig. 1C, E, G, I; p < 0.001). This improvement was accompanied by consistent reductions in insulin levels, weight, lipid profiles, and inflammation levels (Supplemental Fig. S1A–L, O–S). Importantly, these effects occurred without any noticeable changes in the food and water consumption (Supplemental Fig. S1M, N). In stark contrast, capsaicin failed to produce any changes in glycemic levels (p = 0.93) in GF mice (Fig. 1D, F, H, J Supplemental Fig. S2A–E). Furthermore, no effects were observed on weight, lipid metabolism, the integrity of the intestinal barrier, and colon inflammation in GF mice (Supplemental Fig. S2F–S). The comparison between SPF mice and GF mice strongly suggests that the ability of capsaicin to regulate the diet-induced glycemic disturbance depends on the presence of gut microbiota. This highlights the interplay between dietary bioactivity and the microbiota in maintaining glycemic homeostasis.

Capsaicin alleviates glycemic levels in the absence of the TRPV1 receptor

Given the established role of TRPV1 receptor in capsaicin-mediated glycemic regulation [23], we sought to determine whether its presence is essential for capsaicin’s effect. The ablation efficiency of TRPV1 was confirmed through the analysis of the DRG, the predominant site of TRPV1 channel localization (Fig. 2B–E). After validating TRPV1 ablation, the obesity and T2D models were established to investigate the specific contribution of the TRPV1 receptor to the regulation of glycemic levels by capsaicin (Fig. 2A).

Fig. 2.

Capsaicin regulates glycemic levels in TRPV1 receptor-ablated mice. A RTX-treated GF mice were daily administered with capsaicin or placebo with HFD (obesity) or HSD (T2D) after colonization of gut microbiota from healthy SPF mice. B Immunofluorescence microscopy of TRPV1 in DRGs isolated from RTX- or its solvent-treated GF mice. C–E DRG nociceptor marker gene expression in DRGs isolated from RTX- or its solvent-treated GF mice. F–H FBG, OGTT, and FINS of the obese mice across two treatment groups, including placebo (P) and high dose of capsaicin (H). I–K FBG, OGTT, and FINS of the diabetic mice across two treatment groups, including placebo (P) and high dose of capsaicin (H). Data are mean with SEM. * < 0.05, ** < 0.01, and *** < 0.001, based on Student’s unpaired t-test followed by Tukey’s (Gaussian distributed) post hoc test, and ordinary two-way ANOVA followed by Šídák’s (non-Gaussian distributed) post hoc test. TRPV1⁻, ablation of TRPV1 receptor; ip, intraperitoneal injection; ig, intragastric gavage; H(S)FD, high-fat diet or high-sugar diet; P, placebo (solvent of capsaicin); H, high dose of capsaicin (30 mg/kg); ns, not significant

Remarkably, in the absence of the TRPV1 receptor, capsaicin still significantly ameliorated the disorder of glycemic levels induced by HFD or HSD (p < 0.05 in the obesity model, p < 0.01 in the T2D model; Fig. 2F–H, I–K; Supplemental Fig. S3M–P). This suggests that the regulation of glycemic levels by capsaicin does not solely rely on the TRPV1 receptor, aligning with certain findings from existing research [18, 24]. Notably, the improvements in weight, lipid profiles, and inflammation levels observed with capsaicin intervention in TRPV1-ablated mice were consistent with those seen in SPF mice (Supplemental Fig. S3A–F). Furthermore, capsaicin intake did not exacerbate colonic inflammation levels in the context of either obesity or diabetes (Supplemental Fig. S3G–L). Collectively, our results provided robust evidence that capsaicin’s beneficial effects on glycemic regulation are not exclusively mediated by the TRPV1 receptor. Instead, our findings suggest that gut microbiota play a key complementary role in capsaicin’s mechanism of action. This highlights the potential for gut microbiota as a critical pathway through which capsaicin exerts its glycemic regulatory effects.

Capsaicin-induced gut microbiota enables the alleviation of glycemic levels

One month of capsaicin intervention induced significant alterations in organismal structure of gut microbiota at the phylum and genus level (animal experiment 1). Notable changes included an increase in the abundance of Proteobacteria, Verrucomicrobia, Bacteroidetes, and Actinobacteria (Fig. 3A). We further identified an array of capsaicin-enriched gut microbial strains, including Akkermansia muciniphila (p = 0.024), Ligilactobacillus murinus (p = 0.004), Alistipes sp. (p < 0.001), Muribaculum gordoncarteri (p < 0.001), and Paramuribaculum intestinale (p < 0.001), all of which have been previously reported to exhibit the potential to alleviate hyperglycemia (Fig. 3B) [25–28]. Moreover, in the principal coordinate analysis (PCoA) plot, the gut microbiota were primarily clustered by capsaicin and placebo treatment, suggesting that capsaicin intake may induce the drastic alteration in the gut microbiota (Fig. 3C).

Fig. 3.

Gut microbiota in response to capsaicin intervention ameliorates glycemic levels in GF mice. A Abundance ratio plot at the phylum and genus levels based on capsaicin-induced gut microbiota homeostasis (the time point of capsaicin or placebo administration for 1 month). Each bluish radian segment represents a distinct phylum or genus; reddish radian segments represent the capsaicin or placebo intervention. B Stamp plot shows the strains with significant differences, right lab based on t-test, p < 0.05, 95% CI. C PCoA to investigate the impact on the gut microbiota between 1-month capsaicin and placebo intervention. D Obese and diabetic model established in GF mice after FMT (fecal samples for FMT were collected from SPF mice after 1 month capsaicin or placebo intervention). E, F FBG and OGTT in the obese model in GF mice after FMT. G, H FBG and OGTT in the diabetic model in GF mice after FMT. Data are mean with SEM. * < 0.05, ** < 0.01, and *** < 0.001, based on Student’s unpaired t-test followed by Tukey’s (Gaussian distributed), and ordinary two-way ANOVA followed by Šídák’s (non-Gaussian distributed) post hoc test. H(S)FD, high-fat diet or high-sugar diet; P, placebo (solvent of capsaicin); H, high dose of capsaicin (30 mg/kg); ns, not significant

We employed FMT to further validate the role of capsaicin-induced gut microbiota in glycemic regulation. Germ-free mice were supplemented with fecal suspension from SPF mice subjected to 1 month of capsaicin or placebo treatment, followed by the establishment of established obesity and T2D models (Fig. 3D). The results indicated that, even in the absence of continued capsaicin intervention, the capsaicin-induced gut microbiota changes could still exert a sustained influence on glycemic levels (p < 0.001). Improvements in body weight, lipid profiles, and inflammation levels were consistent with observations in the capsaicin-treated SPF mice (Fig. 3E–H, Supplemental Fig. S4A–F, H–Q). These findings underscore the pivotal role of capsaicin-induced gut microbiota alterations in mitigating glycemic disturbances, highlighting their potential for sustained metabolic benefits independent of ongoing capsaicin intake.

Mutual interactions between capsaicin-induced key bacteria and their roles in glycemic regulation

To investigate the interplay between capsaicin-induced key bacteria, GF mice were supplemented with Akkermansia muciniphila and Ligilactobacillus murinus, with Lactobacillus johnsonii—significantly reduced after capsaicin treatment—used as a negative control (Fig. 3B). The regulatory effects of Akkermansia muciniphila and Ligilactobacillus murinus on glycemic control were systematically evaluated (Fig. 4A). The supplementation with Akkermansia muciniphila and Ligilactobacillus murinus effectively mitigated glycemic disturbance induced by HFD or HSD in GF mice (p < 0.01 in obesity model, p < 0.001 in T2D model). Conversely, Lactobacillus johnsonii showed no significant impact on glycemic levels (p = 0.68 in obesity model, p = 0.20 in T2D model; Fig. 4B–E). Similar trends were observed in weight, insulin, blood lipids, and inflammation levels (Supplemental Fig. S5D–S). We next assessed the antibacterial activity of capsaicin through in vitro culture experiments to explore the underlying mechanisms. Growth curves of the above three strains demonstrated that the supplementation of capsaicin had no significant impact on their proliferation (Fig. 4F; Supplemental Fig. S6E–G). This finding suggests that capsaicin does not exert a direct effect on the proliferation of key gut microbiota.

Fig. 4.

Colonization of GF mice with capsaicin-induced key bacteria, Akkermansia muciniphila and Ligilactobacillus murinus, improves glycemic control and reveals a mutualistic interaction. A Obese and diabetic model established in GF mice with H(S)FD after colonized capsaicin-induced significant increased and decreased bacteria (Akkermansia muciniphila and Ligilactobacillus murinus for positive group, Lactobacillus johnsonii for negative control). B, C FBG and OGTT in the obese model in GF mice after colonized with capsaicin-induced key bacteria. D, E FBG and OGTT in the diabetic model in GF mice after colonized with capsaicin-induced key bacteria. F Determination of growth curves of key bacteria with or without capsaicin in vitro. G AB–PAS staining and mucus thickness of colonic tissues from GF mice orally treated with placebo or 30 mg/kg capsaicin. H Volcano plot based on OPLSDA difference analysis (p < 0.05, q < 0.05, VIP > 1) for untargeted metabolomics of Transwell solution from lower chamber with Ligilactobacillus murinus. I Metabolic pathway enrichment analysis of differential metabolites based on KEGG. Data are mean with SEM. * < 0.05, ** < 0.01, and *** < 0.001, based on ordinary one-way or two-way ANOVA followed by Tukey’s (Gaussian distributed, equal SDs) or Dunnett’s T3 (Gaussian distributed, non-equal SDs) post hoc test. H(S)FD, high-fat diet or high-sugar diet; Akk., Akkermansia muciniphila; L.M., Ligilactobacillus murinus; L.J., Lactobacillus johnsonii; OD_{600 nm}, optical density at 600 nm; BHI, brain heart infusion; MM, minimal medium; ns, not significant

Instead, previous studies have suggested that capsaicin promotes mucin production, thereby accounting for the increased abundance of Akkermansia muciniphila observed following capsaicin treatment [29, 30]. Consistently, we observed enhanced mucin-related gene expression and protein levels in SPF mice, FMT recipients, and key bacteria–colonized mice (Supplemental Fig. S5T, U; Supplemental Fig. S6I, J, N–O), supporting a mucus-associated mechanism underlying Akkermansia muciniphila. Importantly, to determine whether this effect reflects a direct host response to capsaicin rather than a microbiota-dependent process, we further assessed mucus layer morphology in GF mice following capsaicin administration. AB–PAS staining revealed a clear increase in colonic mucus layer thickness in GF mice, demonstrating that capsaicin is sufficient to directly stimulate mucus production in the absence of gut microbiota (Fig. 4G).

A Transwell assay was also performed as a preliminary in vitro approach to assess possible metabolic interactions with Akkermansia muciniphila (Supplemental Fig. S5V). Colony counting and OD_{600 nm} measurements showed that Ligilactobacillus murinus exhibited enhanced growth in the presence of metabolites derived from Akkermansia muciniphila in the upper chamber, whereas no comparable effect was observed for Lactobacillus johnsonii (Supplemental Fig. S5W–Y). Several metabolites potentially associated with this interaction, including guanine, leucylproline, 2′-deoxyadenosine, valylproline, and 4-indolecarbaldehyde, were detected and were mainly enriched in amino acid–related pathways (Fig. 4H, I).

The identified metabolites were predominantly associated with amino acid–related pathways, which is consistent with the well-established capacity of Lactobacillus spp. to efficiently utilize dipeptides and small peptides to support cellular growth and metabolic activity [31–33]. In comparison to free amino acids alone, dipeptides or small peptides are typically more readily absorbed and metabolized by Lactobacillus, facilitating their metabolic roles [34]. Notably, the physiological relevance of this interaction is further supported by in vivo metabolomic analyses following capsaicin intervention, which underscore its association with glycemic regulation and metabolic homeostasis.

The effects of capsaicin-induced gut microbiota on glycemic levels are associated with 5-ALA

Targeted liquid chromatography analysis confirmed the unchanged structure of capsaicin in mouse feces (Supplemental Fig. S6K) and the lack of significant changes in colon SCFAs content in SPF mice following capsaicin intervention (Supplemental Fig. S4R–T). This finding prompted us to further investigate the underlying mechanisms of gut microbiota action using untargeted metabolomics. By conducting quantitative and normalized data analysis, we identified intestinal metabolites exhibiting significant changes following capsaicin intervention (p < 0.05, q < 0.05, VIP > 1), including increased 5-ALA, isobutyraldehyde, 4-diethylaminoacetanilide, and N-(2-chloroethyl)-benzylamine HCl (Fig. 5A, B), and the redundancy analysis (RDA) revealed a remarkable finding: 5-ALA exhibited the most pronounced impact on the significantly enhanced gut microbiota (Fig. 5C). As a natural precursor for heme synthesis [35, 36], the regulatory effect of 5-ALA on glycemic levels has been demonstrated to be associated with the heme metabolism enzyme HO-1 [37].

Fig. 5.

Capsaicin intake significantly augments the intestinal 5-ALA level responsible for post-capsaicin-intervention gut microbiota alterations. A Correlation analysis of the metagenomic (p < 0.05) and metabolomic data (p < 0.05, q < 0.05, VIP > 1) (feces samples from SPF mice under 4 weeks capsaicin intervention). B Principal component analysis (PCA) based on FactoMineR to investigate the impact on the intestinal metabolites between 1-month capsaicin and placebo intervention. C Redundancy analysis (RDA) based on the metagenomic and metabolomic data (p < 0.05). D Obese and diabetic model established in GF mice with daily 5-ALA intervention (200 mg/kg). E, F FBG and OGTT in the obese model in GF mice with daily 5-ALA intervention. G, H FBG and OGTT in the diabetic model in GF mice with daily 5-ALA intervention. Data are mean with SEM. * < 0.05, ** < 0.01, and *** < 0.001, based on Student’s unpaired t-test followed by Tukey’s (Gaussian distributed), and ordinary two-way ANOVA followed by Šídák’s (non-Gaussian distributed) post hoc test. H(S)FD, high-fat diet or high-sugar diet; ig, intragastric gavage; ns, not significant

To verify that significantly increased 5-ALA in the colon regulates glycemic levels, we supplemented 5-ALA in GF mice and similarly repeated the prevention models for obesity and T2D (Fig. 5D). The results demonstrated that 5-ALA significantly ameliorated the disorder of glycemic levels induced by HFD or HSD in GF mice (Fig. 5E–H; Supplemental Fig. S7A–E; p < 0.05 in obesity model, p < 0.01 in T2D model), which was accompanied by a significantly upregulation in the expression of HO-1 (Supplemental Fig. S7F). The significant improvements in body weight, lipid profiles, and inflammation levels were observed, consistent with previous studies investigating the association between 5-ALA and metabolic syndrome (Supplemental Fig. S7G–R) [38, 39].

The effects of capsaicin-induced gut microbiota on the accumulation of glycine are associated with 5-ALA

The functional analysis of the capsaicin-induced gut microbiota changes reveals significant alterations in the tricarboxylic acid (TCA) cycle and amino acid metabolism, which are associated with 5-ALA synthesis (Fig. 6A). According to the synthesis of 5-ALA, we quantified glycine, the precursor of 5-ALA, in colonic tissue and colonic fecal samples obtained from SPF mice, which demonstrated a significant increase in the levels of glycine following capsaicin intervention (p < 0.001). Furthermore, there was an observed upregulation of GlyT1 in colon tissue samples (p < 0.01 in obesity model, p < 0.001 in T2D model; GlyT1, a member of the solute carrier family 6 (SLC6) responsible for glycine transport in intestinal epithelial cells and maintenance of glycine supply in colon cells) [40], which implies the translocation of glycine from the colon to the host. Moreover, untargeted metabolomics was also performed on colonic tissue samples; 5-ALA was also found to increase significantly (p < 0.001 in obesity model, p < 0.05 in T2D model). This suggests that the process of glycine synthesis of 5-ALA occurs within the host and is facilitated by capsaicin-induced gut microbiota providing the reaction substrate, glycine (Fig. 6B; Supplemental Fig. S5A, B, Supplemental Fig. S6A, B, H, P, Q).

Fig. 6.

Capsaicin-induced gut microbiota increases glycine both intracolonically and extracolonically, thereby promoting the host to synthesize 5-ALA. A Functional analysis of capsaicin-induced gut microbiota based on HUMAnN3, including related KEGG orthology ID and pathway ID. B Schematic diagram of capsaicin-induced gut microbiota promoting 5-ALA synthesis (the light blue region in the upper left represents the intracolonically environment, whereas the corresponding light red region in the lower section signifies the epithelial environment; the dashed black arrow indicates the direction of the transfer; the right axis chart illustrates the levels of glycine and 5-ALA both intracolonically and extracolonically, alongside the expression patterns of ALAS1 and GlyT1 within intestinal epithelial cells). C Kyoto encyclopedia of genes and genomes (KEGG) metabolic pathway of 5-ALA (including related KEGG orthology ID and enzyme codes, content enclosed by the left dashed line), mechanism of HO-1 regulation of glycemic levels (based on current research, content enclosed by the right dashed line), HO-1 content in each group of SPF and FMT mice (upper-left coordinate axis), capsaicin could promote the synthesis of glycine by Ligilactobacillus murinus (lower-left coordinate axis). Data are mean with SEM. * < 0.05, ** < 0.01, and *** < 0.001, based on Student’s unpaired t-test and ordinary one-way ANOVA followed by Tukey’s (Gaussian distributed) post hoc test. ROS, reactive oxygen species; GLUT, glucose transporter; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; PPAR-γ, peroxisome proliferator-activated receptor gamma; Akk., Akkermansia muciniphila; L.M., Ligilactobacillus murinus; L.J., Lactobacillus johnsonii; cap, capsaicin; ns, not significant

Accordingly, the network mechanism on the effects of capsaicin alleviates glycemic levels was depicted (Fig. 6C). Capsaicin-induced gut microbiota changes lead to a substantial enrichment of glycine both the inner of colon to the epithelium, thereby facilitating enhanced synthesis of 5-ALA within the host, subsequently promoting its absorption into the intestine (Supplemental Fig. S5C; Supplemental Fig. S6M). The upregulation of HO-1 exhibited a consistent result in capsaicin-treated SPF mice, 5-ALA supplemented GF mice, FMT-treated GF mice, and key bacteria colonized GF mice (Supplemental Fig. S6L; Supplemental Fig. S7F; Supplemental Fig. S4G; Supplemental Fig. S6C, D). This finding further substantiates the mechanism by which capsaicin enhances 5-ALA content through specific gut microbiota, consequently ameliorating glycemic levels via the heme pathway. Additionally, we confirmed that capsaicin supplementation can enhance glycine production by Ligilactobacillus murinus in vitro (Fig. 6C). Along with the interaction between Akkermansia muciniphila and Ligilactobacillus murinus, we further elucidated the network mechanism underlying the improvement of glycemic levels through gut microbiota modulation under capsaicin intervention.

Discussion

The inclusion of dietary components, particularly plant extracts, has demonstrated a substantial impact on enhancing glycemic levels [8]. Capsaicin, as a common dietary component, has been reported to possess the ability to alleviate glycemic levels, primarily through its interaction with the TRPV1 nerve receptor [41], but the precise mechanism underlying the hypoglycemic effect of capsaicin remains elusive. The regulation of glycemic levels primarily involves receptor pathways associated with islet cells in the host pancreas, such as the PI3K/Akt pathway, GLP-1, and glucagon-epinephrine related receptors [42]; however, a plethora of studies have demonstrated that the gut microbiota exerts regulatory effects on glycemic levels through mechanisms independent of the direct change on host, including modulation of gut barrier function, production of microbiota-derived metabolites, alteration of bile acid metabolism, and the oxidative stress levels [43]. Our research unveiled the mechanism centered on gut microbiota. Similarly, previous studies have shown that dietary factors such as dietary fiber, polyphenols, and chlorogenic acid can modulate the abundance of Bifidobacterium spp., Lactobacillus spp., Akkermansia muciniphila, and Faecalibacterium prausnitzii to regulate glycemic levels by promoting the production of SCFAs, enhancing gut barrier function, and improving antioxidant capacity [8–11].

Our study highlights the impact of capsaicin on the composition of gut microbiota, including significantly increased Akkermansia muciniphila, Ligilactobacillus murinus, Alistipes sp., Muribaculum gordoncarteri, and Paramuribaculum intestinale, and decreased strains, including Lactobacillus johnsonii, Ileibacterium valens, and Rikenellaceae bacterium. Among them, specific strains, such as Akkermansia muciniphila and Ligilactobacillus murinus, have been reported to exhibit significant glycemic regulatory functions [26, 44–46]. We targeted Akkermansia muciniphila and Ligilactobacillus murinus and analyzed the potential effects of capsaicin on both. Current studies have indicated that capsaicin may promote the secretion of the mucosal layer, which could elucidate the observed increase in the abundance of Akkermansia muciniphila in our results [31, 47]. Additionally, we discovered an undocumented interaction between Akkermansia muciniphila and Ligilactobacillus murinus, specifically that Ligilactobacillus murinus can utilize metabolites from Akkermansia muciniphila in mucin-containing media, including guanine, leucylproline, 2′-deoxyadenosine, valylproline, and 4-indolecarbaldehyde. The mutualistic relationships among microorganisms play a crucial role in maintaining ecosystem stability and biodiversity, as well as supporting gut health, enhancing immune function, and regulating metabolism. Our study provides insights into capsaicin-driven multi-level microbial mutualism in the gut, which may offer new approaches for the prevention and treatment of metabolic diseases. Although direct high-dose capsaicin supplementation may have limited clinical practicality, these findings highlight the therapeutic potential of targeting the identified microbial axis through probiotics, prebiotics, or dietary modulation.

In conjunction with metabolomics measurements, we have identified a pivotal gut microbiota-associated metabolite of capsaicin to alleviate glycemic levels, 5-ALA. Studies have established a relationship between 5-ALA and glycemic regulation, and previous studies have demonstrated the utility of 5-ALA in medical and agricultural domains, encompassing cancer therapy and its application as a growth regulator for crops [48–50]. The synthesis of 5-ALA involves the catalysis of glycine and succinyl-coenzyme A (succinyl-CoA) by 5-aminolevulinate synthase 1 (ALAS1), with ALAS1 serving as the rate-limiting enzyme (K00643, EC 2.3.1.37) [51]. However, significantly increased species in capsaicin-induced gut microbiota were found to lack the gene encoding ALAS1 (Prokka, v.1.14.6). First, based on microbiota analysis, it was observed that capsaicin-induced alterations in the gut microbiota significantly enhance the glycine metabolism. Furthermore, capsaicin-induced key bacteria have the ability to encode glycine synthetase (GlyT1; K00600, EC 2.1.2.1), coupled with the upregulation of ALAS1 expression in the host and the bidirectional transfer of 5-ALA within and beyond the colon [52]; it can be seen that capsaicin-induced gut microbiota leads to a substantial enrichment of glycine, thereby facilitating enhanced synthesis of 5-ALA within the host, subsequently promoting its absorption into the intestine (Fig. 6B). Interestingly, capsaicin-induced alterations in the gut microbiota were also found to significantly enhance amino acid metabolism. Furthermore, we quantified the levels of 5-ALA and glycine in the host, along with glycine concentrations in the colon, to validate the aforementioned network mechanism. The correlation between 5-ALA and glycemic levels has been extensively reported; however, 5-ALA currently lacks specific clinical applications, and our results revealed a potential strategy to indirectly enhance 5-ALA levels by capsaicin, thereby facilitating the regulation of glycemic levels.

Collectively, our proof-of-concept study with multiple systems (i.e., SPF mice, GF mice, TRPV1-ablated mice, and the Transwell culture) and in-depth metagenomic sequencing and metabolome data highlighted the critical yet underexplored roles of gut microbiota in the glycemic regulation by capsaicin. A novel mechanism was elucidated whereby capsaicin-induced gut microbiota changes promote 5-ALA synthesis by increasing colonic glycine content, ultimately leading to improved glycemic levels through heme metabolism. Within this framework, microbiota-targeted interventions—such as specific probiotics or prebiotics involved in this pathway—may act in concert with capsaicin to support its multifaceted glycemic regulatory effects.

Abbreviations

GF

Germ-free

SPF

Specific pathogen-free

TRPV1

Transient receptor potential vanilloid 1

FMT

Fecal microbiota transplantation

5-ALA

5-Aminolevulinic acid

GLP-1

Glucagon-like peptide-1

AMPK

Adenosine monophosphate-activated protein kinase

GLUT2

Glucose transporter type 2

SCFAs

Short-chain fatty acids

HFD

High-fat diet

HSD

High-sugar diet

T2D

Type 2 diabetes

P

Placebo

L

Low-dose capsaicin

H

High-dose capsaicin

STZ

Streptozotocin

RTX

Resiniferatoxin

AKK

Akkermansia muciniphila

LM

Ligilactobacillus murinus

LJ

Lactobacillus johnsonii

DRG

Dorsal root nerve ganglia

HO-1

Heme metabolism enzyme

ALAS1

5-Aminolevulinate synthase 1

GlyT1

Glycine transporter 1

Authors’ contributions

QYF carried out most of the experiments and analyzed the data. MYL, ZY, HG, and MFX helped with experiments. CCZ provided guidance on bioinformatics analysis and data visualization. SHW provided guidance on aseptic techniques and assisted in the husbandry of GF mice. QXZ, HZ, LLY, FWT, WC, and JXZ provided the conception and design for the research. QYF wrote the manuscript. QXZ and SH reviewed and edited the manuscript, while also providing coordination and supervision for the study. All authors made substantial contributions and approved the final version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (No. U23A20259 and 32394051), the Fundamental Research Funds for the Central Universities JUSRP622013, the Collaborative Innovation Center of Food Safety and Quality Control in Jiangsu Province, and the Postdoctoral Fellowship Program and China Postdoctoral Science Foundation (BX20250339).

Data availability

All data relevant to the study are included in the article or uploaded as online supplemental information. This paper does not report the original code. Computational analyses were performed using the bioBakery suite of tools including MetaPhlAn2 (https://github.com/biobakery/MetaPhlAn) for microbiota taxonomic profiling and HUMAnN3 (https://github.com/biobakery/humann) for profiling of functional potential (ECs, pathways and modules). Metagenomic data are publicly available in NCBI Sequence Read Archive under BioProject: PRJNA1171754. The metabolomics data have been deposited to MetaboLights [53] repository with the study identifier MTBLS12575 (https://www.ebi.ac.uk/metabolights/MTBLS12575). Bacteria isolates were deposited in the Culture Collection of Food Microorganisms (CCFM), Jiangnan University (Wuxi, China). This study did not generate new unique reagents. Any additional information required to reanalyze the data reported in this work is available from the lead contact upon request.

Declarations

Ethics approval and consent to participate

The animal experiments 1–6 were approved by the Ethics Committee of Jiangnan University and the Ethics Committee of the Jiangsu Institute of Parasitic Diseases (Wuxi, China), whereas the additional experiment (animal experiment 7) performed during revision was independently approved by the Institutional Animal Care and Use Committee (IACUC) of Gempharmatech Co., Ltd. All procedures complied with the European Community Directive 2010/63/EU.

Consent for publication

All authors have consented to publication.

Competing interests

The authors declare no competing interests.