About 100 g CGLC contains 16 g fruiting body powder of C. guangdongensis, 16 g CLGE, 16 g L-arabinose, 14 g Phaseolus vulgaris Linn. powder, 11 g lotus leaves powder, 8 g root powder of Pueraria Lobata, 6 g maltodextrin, 5 g instant green tea powder, 4.3 g microcrystalline cellulose, 2 g magnesium stearate, and 1.7 g coating powder. C. guangdongensis fruiting bodies were provided by Guangdong Kehua Huitai Biotechnology Co., LTD, dried in a rotary drier at 55°C, and were ground using ultra-low temperature pulverizer at −20°C. L-arabinose and conjugated linoleic acid glycerides were purchased in Zhengzhou Best Health Technology Co., LTD. Microcrystalline cellulose and magnesium stearate were purchased in Henan Sugar Cabinet Food Co., LTD. Coating powder was purchased in Jiangsu Xinyu Pharmaceutical Co., LTD. The remaining were purchased in Shaanxi Guosheng Biotechnology Co., LTD. All the prepared materials were directly and thoroughly mixed to create CGLC.

Forty-five male C57BL/6J mice (6-week-old, specific pathogen-free SPF), purchased from Shenzhen Saibenuo Biological Technology Co., Ltd (China), were housed in a SPF animal room on 12-h light–dark cycle with 22 ± 1°C temperature and 55 ± 10% relative humidity. Meanwhile, these mice were provided with food and water ad libitum. After 1 week of adaptation to regular food and water, all the mice were randomly divided into three groups (n = 15): control check group (CK), high-fat diet group (HFD), and CGLC group (CGLC). Mice of CK group were fed with a regular chow diet (Jiangsu Syony Pharmaceutical Bio-engineering Co. LTD, China), while mice in HFD and CGLC groups were fed with a 60% HFD (D12492, SYSE Bio-tech Co., LTD., China). After 8 weeks, mice in the CGLC group received CGLC solution in dose of 1.5 g/kg/d for 14 weeks, whereas mice from CK and HFD group daily were gavaged using 0.8% saline of the same volume as the CGLC group. During the process of gavage experiments, food intake and mouse body weights were weekly documented, and energy intake was calculated by multiplying the energy by grams of food consumed. The calorie values of regular chow diet and HFD were 3.53 and 5.24 Kcal/g. At last, all mice were fasted for 12 h and then were sacrificed by anesthesia. All animal experiments were approved by Animal Care and Use Committee of Institute of Microbiology, Guangdong Academy of Science (No. GT-IACUC20210406).

At 21 weeks, all mice were fasted for 12 h and orally administered glucose (1.0 g/kg body weight) to perform the oral glucose tolerance test (OGTT). After glucose administration, the blood glucose levels at 0, 15, 30, 60, 90, and 120 min were monitored using a glucometer (HGM-121, Omron, China) and calculated the area under the curve (AUC). At the end of the experiment, body weights and lengths (from nose apex to anus) were measured to calculate the rate of body weight increase and Lee’s efficiency on the basis of the description in a previous study (30). In addition, the epididymis adipose tissue, inguinal adipose tissue, brown adipose tissue, and liver tissue were weighed and used to calculate the fat mass-related index (tissue weight/body mass weight × 100%).

At the end of the experiment, blood samples of 15 mice per group were collected from thorax after CO₂ anesthesia. Then, those blood samples were equilibrated at room temperature for 1 h and were centrifuged at 3,000 revolution per minute (rpm) for 15 min at 4°C. The serum was collected and stored at −80°C. Total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), and the activity of alanine aminotransferase (ALT) in serum were measured using a commercial kit (Nanjing Jiancheng Bioengineering Institute, China). The epididymal adipose and liver tissues were washed with 0.8% saline and fixed in a 4% paraformaldehyde solution. All the tissues were embedded in paraffin and cut into tissue slices with a thickness of 4 μm. Then, these tissues were stained with oil red O and hematoxylin and eosin (HE). Finally, all the slides were observed and photographed using an optical microscope (DM6M, Leica microsystems, Germany).

At 23 weeks, 12 fresh fecal samples for each group were randomly collected, divided into six mixture samples, and immediately stored in a −80°C freezer to analyze gut microbial composition and SCFA contents in Wuhan Metware Biotechnology Co., Ltd. (Wuhan, China). The concentration and integrity of the extracted DNA by CTAB method were detected using 1.5% (w/v) agarose gel electrophoresis. Then, the V3-V4 region of 16S rRNA (ribosomal RNA) gene was amplified with 515F/806R primers. The purified PCR amplicons were used for library construction using TruSeq^® DNA PCR-Free Sample Preparation Kit (FC-121-3001, Illumina, USA) and sequenced on an Illumina Novaseq6000 platform. Paired-end reads were assembled into raw tags using FLASH (Version 1.2.7), and raw tags were filtered using QIIME (V1.9.1) and mapped to species annotation database to obtain effective tags, which were clustered into operational taxonomical units (OUTs) at 97% identity using Uparse algorithm (31).

About 20 mg fecal samples from each mixture were dissolved in 1 mL phosphoric acid (0.5% v/v) solution with a small steel ball and were grinded for 10 min and ultrasonicated for 5 min. After 10 min of centrifugation at 12,000 rpm and 4°C, methyl tert-butyl ether (contain internal standard) solution was mixed with the same volume of supernatant, and the mixture was vortexed for 3 min and ultrasonicated for 5 min. The mixture was then centrifuged for 10 min at 12,000 rpm at 4°C. The supernatant was collected and used for the detection of SCFA contents on the Agilent 7890B-7000D GC-MS/MS platform (32). Every group included six replicates.

Total RNAs of hepatic samples were extracted using a TRIzol Kit (Invitrogen, Dalian, China) according to the manufacturer’s instruction. Every group had four replicates. The qualitative and quantitative analyses of all RNAs were performed on an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). Then, all the RNAs were used for library construction and were sequenced on an Illumina HiSeq platform. After quality trimming and adapter clipping, all the reads were mapped to the reference genome using Hisat2 v2.0.4. DESeq2 was adopted to identify differential expression genes (DEGs) between groups (33). DEGs involved in cholesterol metabolism, fat metabolism, hepatitis B, non-alcoholic fatty liver, lipolysis in adipocytes, and TNF signal pathway were selected for further analysis.

Statistical differences among CK, HFD, and CGLC groups were determined by one-way ANOVA using the Statistical Package for the Social Sciences program (SPSS 22.0). Error bars represent the standard deviation from the mean of six independent replications in every group. Association analyses of many indexes and heat-map figures were performed using SupCorrPlot and Heat-map illustrator packages of TBtools software (34). TBtools software and GraphPad Prism 8 were used for figure construction. Asterisks were used to represent significant differences between each group.

At 8th week of the experiment, the obvious differences of body weights were found between CK and another two groups (Figure 1A). After 14-week gavage of CGLC, a highly significant difference of body weights emerged between HFD and CGLC groups (34.43 ± 2.01 vs 30.73 ± 1.29 g). During the process of gavage, the average energy intakes between HFD and CGLC groups had no obvious difference (Figure 1B). Compared with mice in the HFD group, mice in the CGLC group showed a lower increase in body weights (Figure 1C). In addition, Lee’s coefficient of mice in the HFD group (3.63 ± 0.13) was higher than those of CK and CGLC groups (3.40 ± 0.16 vs 3.41 ± 0.09, Figure 1D).

From the perspective of adipose weights, compared with mice in the CK group, mice in the HFD group showed a markedly increase in the inguinal and epididymis adipose indexes (Figures 1E,F), whereas an obvious decrease was found for the brown adipose index at the end of the experiment (Figure 1G). After CGLC administration for 14 weeks, the inguinal and epididymis adipose indexes significantly decreased (P < 0.05 and P < 0.01), and brown adipose index increased (P < 0.05). In terms of morphological characteristics of adipose tissues, the sizes of inguinal and epididymis adipose exhibited 30.99 and 38.10% decreases, respectively, but a 26.47% increase was found for the brown adipose size in width. Depending on microscopic observation of epididymis adipose tissue, compared with mice in the CK group, epididymis adipose cells in the HFD group were irregular, enlarged in diameter (blue arrows), had more cytoplasmic fractures (red arrows), and significantly decreased for cell number in the same field of viewpoint (Figure 1H). However, after 14-week administration of CGLC, epididymis adipose cells were more tightly arranged, showed less cytoplasmic fractures, reduced in diameter, and had an obvious increase in cell number. These results indicated that CGLC can reduce body weight gain and adipose accumulation in HFD mice without suppressing energy intake.

OGTT results showed that 12-week intervention of CGLC reduced blood glucose levels of HFD mice at 15, 30, 60, 90, and 120 min (Figure 2A, P < 0.01), and the area under the curve (AUC) of the CGLC group exhibited 20.66% decrease (Figure 2B, P < 0.01), indicating that CGLC improved the glucose tolerance of high-fat diet mice. From the perspective of lipid metabolism, we found that 14-week administration of CGLC significantly decreased the levels of TG, TC, and LDL-C in serum (Figures 2C–E, P < 0.01) compared to the HFD group. Among the above three indexes, the levels of TG and LDL-C in the CGLC group had no obvious differences with the CK group and showed, respectively, 21.87 and 52.56% decreases compared to the HFD group (Figures 2C,E). However, the levels of TC in the CGLC group reduced 32.91% of that in the HFD group, whereas the levels were higher than that in the CK group (Figure 2D). These results demonstrated that CGLC can improve blood glucose tolerance and hyperlipidemic level in high-fat diet mice by decreasing the TG, TC, and LDL-C levels in serum (especially TG and LDL-C).

As shown in Figure 2F, the 14-week administration of CGLC reduced the activity of alanine transaminase (ALT) in liver tissue, and a 42.20% decrease level was obtained in the comparison between the CP and HFD groups. At the same time, we found that the liver index in the CGLC group showed a 18.35% increase compared with that in the HFD group (Figure 2G). Based on the observation of liver histological sections stained by HE (Figure 2H), we observed that plenty of fat vacuoles (green arrow) were found in the liver cells of the HFD mice, and their frontier and morphology were obscure; nevertheless, frontier and morphology of liver cells in the CGLC mice were clear and regular, and they had few fat vacuoles. Oil red O staining observation showed that a large amount of lipid drops (red arrows) were deposited in the liver tissue of HFD group mice, and most of the liver tissue area was dark red (Figure 2I). However, CGLC significantly decreased the lipid accumulation compared with the HFD group, and liver tissues in the CGLC group were very slight red. The aforementioned findings revealed that CGLC alleviated lipid accumulation in the liver and further protected the liver cells from fat damage.

A total of 122,468 effective tags were obtained from 18 fecal samples (Supplementary Table 1). On the basis of 97% sequence similarity criterion, a total of 416 OTUs were generated, 353 were in the CK group, 351 were in the HFD group, and 344 were in the CGLC group (Figure 3A), including 31 unique OTUs in the CK group, 17 unique OTUs in the HFD group, and 26 unique OTUs in the CGLC group. PCoA result of the weighted UniFrac distance showed that the clustering of microbiota compositions for the three groups was significantly distinct (Figure 3B). The abundance coverage-based estimator (ACE) and Chao1 indexes in the HFD were larger than those in the CK and CGLC groups (P < 0.05), indicating a higher microbial richness in the HFD group. However, compared with the two other groups, the Shannon and Simpson indexes of the HFD showed a marked decrease (P < 0.01), suggesting a lower microbial diversity in the HFD group (Figure 3C).

To further assess the effect of CGLC on the overall bacterial community overall structure, the degree analyses were further performed at the phylum, genus, and species levels. In terms of phylum level (Figure 4A), the predominant bacterial phylum in the CK group was Bacteroidota, followed by Firmicutes and Verrucomicrobiota; however, Firmicutes was the predominant bacterial phylum in the HFD group, followed by Bacteroidota and Verrucomicrobiota (Figures 4A,B). Compared to the HFD group, after 14-week intervention of CGLC, the relative abundance of Bacteroidota in the CGLC group showed a dramatical increase, whereas Firmicutes was downregulated. In addition, the ratio between Firmicutes and Bacteroidota in the CGLC group decreased 37.32% of that in the HFD group (Figure 4B). These results suggested that high-fat diet changed dominant taxa from Bacteroidota to Firmicutes, and CGLC alleviated the shift.

The top 20 taxa at the genus level were further analyzed in Figure 4C. Among them, Faecalibaculum, Akkermansia, Dubosiella, Lactobacillus, and Bacteroides were the predominant genera. Compared to the CK group, Dubosiella and Faecalibaculum in the HFD group showed 76.74 and 133.50% increases, respectively; in addition, the relative abundances of Romboutsia, Ligilactobacillus, and Lactobacillus of Firmicutes in the HFD group also showed highly obvious increases (Figure 4D, P < 0.01). However, the relative abundances of Muribaculum and Bacteroides showed 39.85 and 24.85% decreases (Figure 4E). After 14-week intervention of CGLC, the relative abundances of Ligilactobacillus, Romboutsia, Dubosiella, Faecalibaculum, and Lactobacillus of Firmicutes significantly decreased by 66.59, 55.79, 48.47, 35.77, and 35.70%, respectively; nevertheless, a significant increase was found for the relative abundances of Bacteroides, Alistipes, Bifidobacterium, and Muribaculum. Moreover, the top 10 differential taxa at the genus level were analyzed (Figure 4F). In the comparison CK vs HFD, the relative abundances of Faecalibaculum rodentium, Lactobacillus johnsonii, Romboutsia ilealis, and Lactobacillus reuteri in the HFD group showed a highly significant upregulation (P < 0.01), while Bacteroides acidifaciens in the HFD group showed an obvious decrease (32.64%, P < 0.01). After 14-week administration of CGLC, the relative abundances of Bacteroides sartorii, B. acidifaciens, B. caecimuris, B. vulgatus, and Bifidobacterium pseudolongum showed a markedly significant increase (P < 0.01); however, R. ilealis, D. newyorkensis, F. rodentium, Lactobacillus johnsonii, and Lactobacillus reuteri showed obvious decreases. These results indicated that CGLC treatment increased the relative abundances of Bacteroides and Bifidobacterium but decreased Romboutsia and Dubosiella.

Considering that SCFAs can affect the absorption of sugars and energy, the fecal samples from three group mice were collected for SCFA content measurement. As shown in Figure 5A, a high-fat diet significantly reduced the total SCFA production with an approximate 21.06% decrease compared to the CK group (P < 0.01). After 14-week administration of CGLC, the total SCFA content of the CGLC group showed a marked increase (about 26.27%, P < 0.01) and had no obvious difference compared to the CK group. In terms of certain SCFA, the production of propionic acid (PA), butyric acid (BA), valeric acid (VA), and isobutyric acid (IBA) reduced in the HFD group (P < 0.01). Nevertheless, a 14-week gavage of CGLC significantly promoted the production of all the SCFAs of obese mice. Among them, the contents of acetic acid (AA), VA, IBA, isovaleric acid (IVA), and caproic acid (CA) in the CGLC group were markedly higher than those of the CK group (P < 0.01), whereas the butyric acid content in CGLC group showed a significant decrease.

Changes in gut microbiota composition can likewise regulate the production of SCFAs. Therefore, the association analysis of gut microbiota and SCFAs was performed. At the phylum level, Desulfobacterota had an positive correlation with the production of SCFAs except AA; the production of BA, IBA, PA, and VA showed a significantly positive correlation with Deferribacteres (P < 0.01); a markedly positive correlation was found between Actinobacteria and CA (Figure 5B, P < 0.001). At the genus level (Figure 5C), the production of IBA and PA and VA positively correlated with Bacteroides (P < 0.01), and a highly positive correlation existed between CA and Bifidobacterium (P < 0.001). However, Dubosiella markedly negatively correlated with BA, IBA, IVA, PA, and VA (P < 0.001). In addition, the production of IBA and VA showed a significantly negative correlation with Ligilactobacillus and Romboutsia (P < 0.01). At the species level, Bacteroides acidifaciens and B. sartorii and B. vulgatus showed significantly positive correlations with BA, IBA, PA, and VA (P < 0.01), Bifidobacterium pseudolongum positively correlated with BA and CA and IVA (P < 0.01), and a markedly positive correlation was found between Bacteroides caecimuris and IBA (Figure 5D, P < 0.01). Nevertheless, Lactobacillus johnsonii significantly negatively correlated with BA, IBA, IVA, PA, and VA (P < 0.001), and a highly negative correlation was observed between D. newyorkensis and CA and IVA (P < 0.01). In addition, we found that R. ilealis showed a markedly negative correlation with IBA, PA, and VA (P < 0.01). The above findings suggest that CGLC can promote the production of SCFAs by increasing the abundance of Bacteroides and Bifidobacterium or decreasing the abundance of Dubosiella and Ligilactobacillus and Romboutsia, further regulating the absorption of sugars and energy.

Spearman’s coefficient was calculated to estimate the relationship between the biochemical parameters of serum and the mainly differentially genera and SCFAs. As shown in Figure 6A, four indexes (TC, TG, LDL-C, and ALT) showed positive correlations with Dubosiella (P < 0.05); TC and LDL-C markedly positively correlated with Faecalibaculum, Romboutsia, and Ligilactobacillus (P < 0.05); and TC and Lactobacillus were positively correlated (P < 0.01). However, TC significantly negatively correlated with Bacteroides and Muribaculum (P < 0.01), and a highly markedly negative correlation was obtained between LDL-C and Muribaculum (P < 0.001). At the species level, it was found that TC showed an positive correlation with F. rodentium, Lactobacillus johnsonii, D. newyorkensis, and R. ilealis (P < 0.05); LDL-C highly positively correlated with F. rodentium, Lactobacillus johnsonii, and R. ilealis (P < 0.01); TG only had a positive correlation with D. newyorkensis (P < 0.05); ALT and Lactobacillus johnsonii were significantly positively correlated with each other (Figure 6B, P < 0.01). In addition, negative correlations were found between TC and Bacteroides sartorii, B. vulgatus, B. caecimuris, and B. acidifaciens (P < 0.05). The association analysis of biochemical parameters of serum and SCFAs found that PA and VA had an negative correlation relationship with TC, TG, and LDL-C (P < 0.05); BA markedly negatively correlated with TC and LDL-C; and LDL-C showed a highly significant correlation relationship with IBA and CA (Figure 6C, P < 0.001). Those descriptions indicate that CGLC can improve blood lipid level in serum by the abundance change in gut microbiota and the increase in SCFA production.

To explore the expression pattern of genes involved in lipid and fat metabolism, liver samples from three group were collected and used for transcriptome analysis. Compared to the HFD group, 577 DEGs (297 upregulation and 280 downregulation) and 356 DEGs (302 upregulation and 54 downregulation) were obtained in the CK and CGLC groups, respectively (Figure 7A and Supplementary Tables 2, 3). A total of 123 DEGs were shared in HFD vs CK and HFD vs CGLC (Figure 7B). Among these overlapping DEGs, 20 genes were related to fat and lipid metabolism. According to the RPKM values, it was found that their expression patterns in the CGLC group were similar to those of the CK group (Figure 7C).

Among 10 genes involved to fat metabolism, six genes (colipase pancreatic Clps, phospholipase group IB pancreas Pla2g1b, pancreatic lipase Pnlip, pancreatic lipase-related protein 1 Pnliprp1, Pnliprp2, and peroxisome proliferator activator receptor delta Ppard) were markedly upregulated in the comparison CGLC vs HFD, whereas the expression levels of those genes were obvious downregulation or showed no significant differences for HFD vs CK; nevertheless, the expression levels of cell death-inducing DNA fragmentation factor alpha subunit-like effector A Cidea and cell death-inducing DFFA-like effector c Cidec showed marked upregulation only in the HFD group (Figure 7D and Supplementary Table 4).

In terms of two genes related to cholesterol metabolism, Lrp2 (low-density lipoprotein receptor-related protein 2) was upregulated in the HFD group, whereas the expression level of Star (steroidogenic acute regulatory protein) showed significant upregulation in the CGLC group (Figure 7D). For six genes related to hepatitis B and NAFLD, Cdknla (cyclin-dependent kinase inhibitor 1A), Tlr4 (toll-like receptor 4), Fos (FBJ osteosarcoma oncogene), Jun (jun proto-oncogene), and Irs1 (insulin receptor substrate 1) in the CGLC group were upregulated compared to the HFD group, whereas Cdknla and Fos and Irs1 showed an obvious downregulation in the comparison HFD vs CK (Figure 7D and Supplementary Table 4). In addition, receptor-interacting serine–threonine kinase 3 (Ripnk3) and adrenergic receptor beta 1 (Adrb1) were upregulated only in the comparison CGLC vs HFD (Figure 7D and Supplementary Table 4). These above-mentioned results suggest that CGLC administration can regulate the expression levels of genes related to lipid and fat metabolism in obese mice induced by high-fat diet, further decreasing the fat accumulation in liver.