Portulaca oleracea polysaccharide was prepared by Lanzhou Wotelaisi Biotechnology Co., Ltd. using a method modified from Ning et al. (18) with further optimization (e.g., a solid-liquid ratio of 1:7 and extraction at 90–95 °C for 2 h). Animal feeds MD12032 (45% kcal fat), MD12031 (10% kcal fat), and MD12032 containing 3.2% (wt/wt) POP were prepared by Jiangsu Meidison Biomedical Co., Ltd., based on our previous research (4). The dietary components are listed in Supplementary Table S1.

Thirty 4/5 week-old male C57BL/6J mice were purchased from Hunan Silaike Jingda Experimental Animal Co., Ltd. Mice were housed in the Experimental Animal Center of Jinggangshan University (SYXK<Jiangxi>2023-0009) under specific pathogen-free (SPF) conditions at 23 ± 1 °C, 50–60% relative humidity, with a 12/12 light/dark cycle, and free access to food and water. After one week of adaptive feeding, all animals were randomly divided into three groups (n = 10 each): normal diet group (Con) fed MD12031, HFD group fed MD12032, and POP group fed MD12032 containing 3.2% POP, for 17 weeks. Body weight and food intake were recorded weekly. At the end of the experiment, mice were anesthetized with isoflurane by inhalation using a small animal ventilator (oxygen flow rate: 0.5–0.7 L/min; isoflurane concentration: 1.0–1.5%), followed by blood collection via orbital enucleation. Liver, perirenal fat, epididymal fat, inguinal fat, and brown fat tissue were dissected, weighed, and stored at −80 °C. All experimental protocols were approved by the Ethics Committee of Jinggangshan University Experimental Animal Center [(2023) Ethical Approval No. (159)]. All procedures involving animals were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals (8th edition), and the Guidelines for the Care and Use of Laboratory Animals issued by the Ministry of Science and Technology of China.

At week 16, mice were fasted for 12 h and then orally administered glucose at a dose of 2 g/kg body weight. Glucose concentrations in tail blood were measured at 0, 15, 30, 60, 90, and 120 min using a handheld blood glucose meter (Yuyue 580, Yuyue Medical Equipment Co., Ltd., China).

Blood samples were collected by orbital exsanguination, left at room temperature for 4 h, and then centrifuged to obtain serum. Levels of TC, TG, high-density lipoprotein cholesterol (HDL-C), and LDL-C in serum were measured using an automatic biochemical analyzer (Dimension RxL Max, Siemens, United States) according to the manufacturer’s instructions.

Liver, inguinal white adipose tissue (iWAT), and brown adipose tissue fixed in 4% paraformaldehyde were dehydrated in gradient ethanol, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) (19). Fresh liver tissue was frozen-sectioned and stained with Oil Red O according to the kit instructions (Beyotime C0158S#, China). Images were acquired using an Olympus BX53 microscope (Olympus, Tokyo, Japan).

Fecal samples collected from the colon were sent to OE Biotech Co., Ltd. (Shanghai, China) for 16S rRNA gene sequencing. Briefly, total genomic DNA was extracted using the MagPure Soil DNA LQ Kit (Magan) according to the manufacturer’s instructions. The V3-V4 hypervariable regions (343F 5′-TACGGRAGGCAGCAG-3′ and 798R 5′-AGGGTATCTAATCCT-3′) of the prokaryotic 16S rRNA were amplified for bacterial diversity analysis. Raw data were processed by cutting off primer sequences using Cutadapt software. Qualified paired-end raw data were subjected to quality filtering, denoising, splicing, and chimera removal using the DADA2 algorithm with default parameters in QIIME 2 (V2020.11) to obtain final valid data for bioinformatics analysis.

Thirty milligrams of fecal samples were weighed out and mixed with 400 μL of methanol-water (v:v = 4:1, containing L-2-chlorophenylalanine at 4 μg/mL). The mixture was precooled at −40 °C for 2 min, ground for 2 min (60 Hz), subjected to ultrasonic extraction in an ice-water bath for 10 min, and then incubated at −40 °C for 30 min, followed by centrifugation at 12,000 rpm at 4 °C for 10 min. Three hundred microliters of the supernatant was collected and dried under nitrogen. The residue was reconstituted with 300 μL of methanol-water (v:v = 1 : 4), vortexed for 30 s, subjected to ultrasonic treatment in an ice-water bath for 3 min, and incubated at −40 °C for 2 h. The reconstituted extract was centrifuged at 12,000 rpm at 4 °C for 10 min. One hundred and fifty microliters of the supernatant was filtered through a 0.22 μm organic phase syringe filter and stored at −80 °C until LC-MS/MS analysis. Quality control (QC) samples were prepared by mixing equal volumes of extracts from all fecal samples. The insertion frequency of QC samples was set as follows: once before sample detection, once every 7 samples during the detection process, and twice consecutively after the completion of all sample detection.

The UPLC-MS/MS analysis was performed using an ACQUITY UPLC I-Class plus (Waters Corporation, Milford, United States)/Thermo QE HF (Thermo Fisher Scientific, Waltham, MA, United States) system with an ACQUITY UPLC HSS T3 (100 mm × 2.1 mm, 1.8 μm) column. It was equipped with a heated electrospray ionization (ESI) source (Thermo Fisher Scientific, Waltham, MA, United States), which was used for analyzing metabolic profiles in both positive and negative ESI modes. The gradient elution system consisted of (A) water (containing 0.1% formic acid, v/v) and (B) acetonitrile (containing 0.1% formic acid, v/v), with the following gradient program: 5% B at 0.01 min; 5% B at 2 min; 30% B at 4 min; 50% B at 8 min; 80% B at 10 min; 100% B at 14 min; 100% B at 15 min; 5% B at 15.1 min; and 5% B at 16 min. The flow rate was 0.35 mL/min, the column temperature was maintained at 45 °C, and the injection volume was 2 μL. The mass range was set from 70 m/z to 1,050 m/z, with a resolution of 70,000 for full-scan MS (MS¹) and 17,500 for MS/MS (MS²) scans. Collision energies of 10, 20, and 40 eV were applied for MS/MS fragmentation. The operating parameters of the mass spectrometer were as follows: spray voltage: 3800 V (+) and 3,000 V (−); sheath gas flow rate: 35 Arb; auxiliary gas flow rate: 8 Arb; capillary temperature: 320 °C; auxiliary gas heater temperature: 350 °C; and S-lens RF level: 50 (20). Information of internal standards, solvents, and reagents are listed in Supplementary Table S2.

Raw data were processed for baseline filtering, peak identification, integration, retention time correction, peak alignment, and normalization using Progenesis QI V2.3 (Nonlinear Dynamics, Newcastle, United Kingdom), with key parameters set as follows: precursor ion tolerance of 5 ppm, product ion tolerance of 10 ppm, and product ion threshold of 5%. Compound identification was based on precise mass-to-charge ratio (m/z), secondary fragments, and isotopic distribution using The Human Metabolome Database (HMDB), Lipidmaps (V2.3), Metlin, and self-built databases (LuMet-Animal3.0). Preprocessed data were analyzed on the Oebiotech Cloud platform¹. The matrix was imported into R for principal component analysis (PCA) to observe the overall distribution of samples and the stability of the analysis process. Orthogonal partial least squares-discriminant analysis (OPLS-DA) was used to identify differential metabolites between groups. Metabolites with p < 0.05 and VIP > 1 were selected as differential metabolites. Differential metabolites were further subjected to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway² enrichment analysis.

Data are presented as means ± SEM and analyzed using GraphPad Prism 8.1.0 (GraphPad Software, United States). One-way analysis of variance (ANOVA) was used to determine significant differences. A p-value <0.05 or <0.01 was considered statistically significant between groups. Correlation analysis of obesity-related parameters with intestinal flora and metabolites was performed using Spearman correlation analysis and Euclidean distances-based redundancy analysis (RDA).

To evaluate the anti-obesity effect of POP, C57BL/6J mice were fed a normal diet (Con), HFD, or HFD with POP (POP) for 17 weeks. Notably, body weight in the HFD group was significantly higher than that in the Con group after 6 weeks of HFD feeding and remained higher until the end of the experiment (Figure 1A). However, POP supplementation significantly reduced body weight gain in HFD-fed mice (Figure 1A). Additionally, POP supplementation decreased total body weight gain induced by HFD (Figure 1B), independent of food intake (Figure 1C). These results indicate that the effect of POP on body weight was not related to changes in food consumption. Representative images of the mice’s overall appearance, liver, and iWAT are shown in Figure 1D, demonstrating significant improvement in liver steatosis and fat accumulation.

Previous studies have shown that long-term HFD feeding may cause histopathological changes in adipose tissue (15). In this study, we found hypertrophy in liver and adipose tissues in HFD-fed mice. However, POP supplementation significantly reduced the weights of liver, white adipose tissue (epididymal fat, eWAT; inguinal fat, iWAT; and perirenal fat, pWAT) in HFD-fed mice. Meanwhile, H&E staining analysis showed that brown adipocytes in HFD mice supplemented with POP were smaller, with reduced lipid droplet size (Figure 2G). Additionally, POP supplementation significantly reduced the cell size of iWAT in HFD-fed mice (Figure 2H).

Abnormalities in glucose and lipid homeostasis are common metabolic disorders in obese patients (10). We further analyzed oral glucose tolerance (OGTT) and serum lipid levels. Results showed that the blood glucose response in the HFD group exceeded that in the Con group throughout the OGTT, while POP treatment improved glucose tolerance in OGTT. Serum biochemical indices including TG, TC, LDL-C, and HDL-C were measured. As shown in Figure 3A, HFD significantly increased TG, TC, LDL-C, and HDL-C levels. POP intervention reversed this trend, indicating that POP can improve HFD-induced changes in lipid levels. Additionally, histological evaluation of liver tissue showed hepatocyte disorder and numerous lipid droplet vacuoles in HFD-fed mice, which were reversed by POP intake (Figures 3B,C).

Considering that HFD feeding leads to intestinal dysbiosis, we detected gut microbiota structure by 16S rRNA analysis. Figure 4A shows that the α-diversity indices Chao1 and ACE were significantly lower in the HFD group compared to the Con and POP groups, while the Shannon index in the POP group was significantly higher than in the HFD group. Principal coordinate analysis (PCoA) showed significant differences in gut microbiota composition among different diet groups (Figure 4B), indicating important changes in gut microbial profiles after POP supplementation. Unweighted pair group method with arithmetic mean (UPGMA) analysis based on the binary Jaccard algorithm also confirmed these differences between groups (Figure 4C). The number of unique and shared bacteria at the OTU level is listed in Figure 4D. The Con group had 226 unique bacteria, while the HFD and POP groups had 177 and 203, respectively, indicating that HFD and POP altered gut microbiota.

Next, we further studied the composition and changes of gut microbiota at the phylum and genus levels. Results showed five dominant bacterial phyla at the phylum level (Figures 5A,B). HFD increased the relative proportions of Firmicutes, Desulfobacterota, Deferribacterota, and Campilobacterota, while decreasing the relative proportion of Bacteroidota. Conversely, these changes were reversed by POP intervention. Additionally, HFD led to an increase in the F/B ratio (Figure 5C), a common gut microbiota feature in insulin-resistant and obese individuals. Interestingly, long-term POP supplementation reduced the F/B ratio in HFD-fed mice to a level similar to the Con group.

At the genus level, community bar charts showed the top 15 gut microbiota genera in each group, with Muribaculaceae, Helicobacter, Bacteroides, Parabacteroides, and Mucispirillum being the most dominant (Figure 5D). In particular, compared to the Con group, HFD significantly increased the proportions of Helicobacter and Mucispirillum, while significantly decreasing the abundances of Muribaculaceae and Parabacteroides. Notably, POP supplementation enriched the abundances of Muribaculaceae and Parabacteroides, while reducing those of Helicobacter and Mucispirillum. Figure 5E shows the 15 bacteria with the highest relative abundances at the species level. Compared to the HFD group, POP significantly reduced the abundances of Helicobacter typhlonius, Bacteroides sartorii, Bacteroides acidifaciens, and Mucispirillum sp., while increasing the abundance of Parabacteroides distasonis. Collectively, these findings provide complementary evidence for the effects of POP on the community and composition of gut microbiota in HFD-induced obese mice.

Linear Discriminant Analysis Effect Size (LEfSe) is a method for identifying high-dimensional biomarkers and revealing genomic features. We used Linear Discriminant Analysis (LDA) to estimate the impact of characteristic species. The histogram of LDA values (Figure 6A) shows species with LDA scores >3, which were statistically different between groups. Biomarkers in the Con group mainly included Gammaproteobacteria, Corynebacteriales, Rhizobiales, Bacteroidaceae, Peptostreptococcaceae, Erysipelatoclostridiaceae, and Micrococaceae and their subclasses. The HFD group was dominated by Lachnospirales and Oceanospirillales. The POP group mainly included Burkholderiales, Tannerellaceae, and Sutterellaceae. Representative specific bacteria in all groups are further shown in Figures 6B–G. The relative abundances of Anaerotruncus and Enterorhabdus in the HFD group were significantly higher than those in the Con group. However, POP intervention significantly reduced their abundances (Figures 6C,D). Meanwhile, compared to the Con and HFD groups, the relative abundances of Parabacteroides, Acetatifactor, and Incertae_Sedis in the POP group were significantly increased (Figures 6B,F,G), while Defluviitaleaceas_UCG-011 was significantly decreased (Figure 6E). Additionally, gut microbiota in the HFD group showed significantly more genes related to sulfur relay system, glycerolipid metabolism, porphyrin and chlorophyll metabolism, quorum sensing (QS), and ABC transporters, but fewer genes related to other glycan degradation (Figure 6H).

To reveal metabolic changes induced by POP, we analyzed the metabolic profiles of colonic contents from mice on normal diet, HFD, and HFD with POP using positive/negative ion LC-MS/MS mode. Metabolite annotation showed that all metabolites mainly belonged to nine categories, including Lipids and Lipid-like Molecules, Organic Acids and Derivatives, Organoheterocyclic Compounds, Benzenoids, etc. (Figure 7A). Partial least squares discriminant analysis (PLS-DA) models were used to detect the relationship between metabolite levels and sample types. Comparisons between groups showed that different dietary patterns and POP intervention led to metabolic differences, with mice in each group clustering separately (Figures 7B,C). Differential analysis identified 44 differential metabolites between the HFD and Con groups, including 42 upregulated and 2 downregulated metabolites (Figure 7D and Supplementary Table S3), and 46 differential metabolites between the POP and HFD groups, including 35 upregulated and 11 downregulated metabolites (Figure 7E and Supplementary Table S4). To reveal the potential functions of differential metabolites, KEGG analysis was performed. The main metabolic pathways between the HFD and Con groups included linoleic acid metabolism, glycerophospholipid metabolism, biosynthesis of unsaturated fatty acids, alpha-linolenic acid metabolism, and arachidonic acid metabolism (Figure 7F and Supplementary Table S5). The main metabolic pathways between the POP and HFD groups included glycerolipid metabolism and linoleic acid metabolism (Figure 7G and Supplementary Table S6).

Moreover, the common differential metabolites among the Con, HFD, and POP groups are presented in Figure 8. Compared to the Con group, 9r,10s-Epome, LacCer (d18:1/12:0), and NDC were significantly increased in the HFD group, while succinyladenosine was significantly decreased (Figure 8A). Long-term POP supplementation reversed these metabolic changes compared to the HFD group (Figure 8B).

Spearman’s correlation analysis and Euclidean distances-based RDA were used to further explore the interaction between gut microbiota and obesity-related biochemical parameters. As shown in Figure 9A, Anaerotruncus was positively correlated with body weight, body weight gain, perirenal fat, inguinal fat, epididymal fat, brown fat, and TG, and negatively correlated with food intake. Enterorhabdus was positively correlated with body weight gain, liver weight, HDL-C, and LDL-C. Additionally, the Parabacteroides genus was negatively correlated with liver weight and TG.

The RDA results (Figure 9B) showed that succinyladenosine was negatively correlated with body weight, body weight gain, perirenal fat, inguinal fat, epididymal fat, brown fat, TG, TC, LDL-C, liver weight, and kidney weight, and positively correlated with food intake. In contrast, LacCer (d18:1/12:0) and NDC were positively correlated with body weight, body weight gain, perirenal fat, inguinal fat, epididymal fat, brown fat, TG, TC, LDL-C, HDL-C, liver weight, and kidney weight. Meanwhile, 9r,10s-Epome was positively correlated with inguinal fat.