2.1 Chemicals and regents

D-gal was purchased from Sigma-Aldrich (MO, USA). D₂O was purchased from Norell-Landisville Co. (NJ, USA). The superoxide dismutase (SOD), total antioxidant capacity (T-AOC), malondialdehyde (MDA), glutathione peroxidase (GSH-Px) and catalase (CAT) assay kits were purchased from Jiancheng Bioengineering Institute (Nanjing, China). The kits for determination of BCA protein concentration were purchased from Beyotime Institute of Biotechnology (Shanghai, China).

The reference standards of quercetin, quercetin-3-O-robinobioside (Rob), rutin, quercetin-3-O-β-D-glucoside (Glu), quercetin-3-rhamnoside (Rha), and quercetin-3-O-β-D-galactose (Gal), were purchased from the Herbest Bio-technology Co. (Shaanxi, China).

LZJS was collected from Yuci, Shanxi, China in June 2017, and was authenticated by Prof. Chenhui Du (Shanxi University of Chinese Medicine, Shanxi, China). LZJSF was prepared according to the report by our group. The preparation method of LZJSF was listed in Supplementary File 1. The content test of LZJSF was performed using our previous method (Yan et al., 2020) and the detailed conditions were listed in Supplementary File 2.

2.2 Animal experimental design and sample collection scheme

A total of 36 adult male Sprague-Dawley rats (7 weeks old, weighing 180 ± 20 g) were procured from Beijing Vital River Laboratory Animal Technology Co., Ltd. with License No. SCXK 2016–0006 (Beijing, China). The rats were housed under controlled conditions of temperature (22 ± 2 °C), light (12 h light/dark cycle) and humidity (50 ± 10%). All animal care and experimental procedures were in compliance with the Guidelines for Care and Use of Laboratory Animals of Shanxi University (Approval number. SXULL 2020029; Time: 03/01/2020).

Following one-week acclimatization to their home cages, the rats were randomly divided into 4 groups (9 rats per group) and treated with normal saline (Control group), 300 mg/kg of D-gal (D-gal group), 300 mg/kg of D-gal plus 100 mg/kg of LZJSF (the low-dose LZJSF, named as LLZJSF) group and 300 mg/kg of D-gal plus 200 mg/kg of LZJSF (the high-dose LZJSF, named as HLZJSF) group. The D-gal group was intraperitoneally injected (i. p.) with D-gal once daily for 36 consecutive days, while the Control group received the same dosage of normal saline. Then, the two drug groups were administered with LZJSF which was dissolved in distilled water by intragastric administration (i. g.) 1 h after D-gal was injected. After 30 days of feeding, Morris Water Maze (MWM) tests were conducted in each group as previously described to determine each rat's memory capacity (Mifflin et al., 2021). The detailed processes of behavioral analysis via the MWM are described in the Supplementary File 3. Animal experimental design and sample collection scheme are shown in Fig. 1A. The rats were weighed weekly. After 35 days of LZJSF administration, the urine samples and the feces samples were collected individually and stored at −80 °C for further use. The probe test was conducted on the 36 th day, and all rats were sacrificed 1 h after the completion of the probe test. Rat blood obtained from the abdominal aorta was centrifuged at 1,123 × g (4 ℃) for 15 min in order to obtain serum samples. Whole brain tissues were collected, immediately weighed and stored at −80 °C for further analysis.

Close

Fig. 1

(A) Animal experimental design and sample collection scheme. (B) Precise and well-defined research strategy for screening different metabolites for serum samples.

Export to PPT

2.3 Measurement of antioxidant indexes in serum and brain

Brain tissues samples were prepared by homogenate and centrifugal treatment at 367 × g (4 ℃) for 15 min. The GSH-Px, SOD, MDA, CAT, and T-AOC levels in serum and brain samples were measured with ELISA according to instructions, respectively.

2.4 ¹H NMR-based metabolomics analysis

Each serum sample (200 μL) was mixed with 400 μL of D₂O and centrifuged (15, 490 × g, 4 °C, 20 min), and then the supernatant was used for NMR analysis on a Bruker 600 MHz AVANCE III NMR spectrometer (Karlsruhe, Germany). Each urine sample was mixed with 100 μL of D₂O and centrifuged (13,200 × g, 4 ℃, 10 min) to obtain supernatants for NMR analysis. The serum and urine samples were analyzed using the Carr-Purcell-Meiboom-Gill (CPMG) spinecho pulse sequence consisting of 64 scans with the following parameters: spectral width of 12,019.23 Hz, spectral size of 32,000 data points, and relaxation time of 1.0 s. The specific methods for the ¹H NMR spectra processing are in Supplementary File 4.

2.5 Multivariate analysis

For multivariate analysis, all preprocessed NMR data were imported into the Simca-P14.0 software (Umetrics, Sweden). Partial least squares discriminant analysis (PLS-DA) model was applied to examine the trend of changes in serum and urine samples among four groups. For urine sample, the orthogonal projection to partial least square (OPLS-DA) model qualities were determined by examining the total explained variables (R²X) and predictability (Q²) values, followed by rigorous permutations (number: 200).

2.6 Precise and well-defined research strategy for screening different metabolites for serum samples

For serum samples, the different metabolites were screened between Control and D-gal group using a precise and well-defined research strategy based on the previous method with minor modification (Li et al., 2019a). In brief, the detailed chart is shown in Fig. 1B and the strategy contained four steps: (1) the variables which met CV (coefficient of variation) ≤ 30 % and the FDR (false discovery rate) ≤ 0.05 were considered as preliminary variables (N₀); (2) the regression equation for N_a/N₀ = f(a), N_b/N₀ = f(b), N_c/N₀ = f(c) and N_d/N₀ = f(d) was conducted, respectively, using the values for four factors including VIP (Variable importance in the projection), FC (Fold Change), -LNP and |Pcorr|. The thresholds of four factors were set as a, b, c and d, respectively. According to VIP > 1, |Pcorr| > 0.58, FC > 1 and -LNP > 3, the minimum values of a, b, c and d were set as a_min = 1, b_min = 1, c_min = 3 and d_min = 0.58, respectively. The maximum value of a, b, c, d was set as a_max = f⁻¹(N_a/N₀ = 0.01), b_max = f⁻¹(N_b/N₀ = 0.01), c_max = f⁻¹(N_c/N₀ = 0.01), d_max = f⁻¹(N_d/N₀ = 0.01), respectively. Each factor was divided equally into 5 thresholds (a₁, a₂, …a₅; b₁, b₂, …b₅; c₁, c₂, …c₅; d₁, d₂, …d₅). The different metabolites were identified based on thresholds from each factor; (3) the different metabolites were further verified by performing PLS-DA analysis. The R²X, R²Y and Q² were used to evaluate interpreting, grouping and predicting abilities of the model, respectively and the optimal threshold was selected; (4) the above four factors were randomly combined to obtain 11 variable sets which was verified by PLS-DA analysis using the values of R²X, R²Y and Q², respectively. Finally, the significant different metabolites were obtained.

2.7 Gut microbiota sequencing analysis

16S rRNA gene sequencing analysis of feces samples was performed by Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China) using the Illumina Miseq platform (2 × 300 bp paired end runs). Sequence reads was processed using QIIME (version 1.8.0). The data analysis was conducted using the Genes Cloud platform (https://www.genescloud.cn).

2.8 Statistical analysis

All experimental values represented the mean ± standard deviation (SD). The illustrations were created using GraphPad Prism version 8.0 (Graph Pad). Statistical differences between two groups were performed using t-test on SPSS 19.0 (New York, USA). Spearman’s correlation analysis and visualization were performed using the Chiplot platform (https://www.chiplot.online). The generation of a network graph was facilitated through the utilization of Cytoscape 3.7.2. Metabolic pathways and related metabolites were analyzed using the KEGG database (https://www.kegg.com). Potential markers and ¹H NMR data were identified by using BMRB (https://bmrb.io) and HMDB (https://www.hmdb.ca).

3.1 Determination contents of six component in LZJSF

The contents of quercetin-3-O-Rob, rutin, quercetin-3-O-β-D-Gal, quercetin-3-O-β-D-Glu, quercetin-3-Rha, and quercetin of LZJSF were measured by LC-MS/MS according to the method reported in our group (Yan et al., 2020), with 0.83%, 0.010%, 0.030%, 0.020%, 0.00050%, and 1.00%, respectively (Fig. S1).

3.2 Effect of LZJSF on cognitive dysfunction in D-gal-induced aging rats

To address whether LZJSF displayed aging-induced cognitive dysfunction, D-gal-induced aging rats in the present study were used as aging models. Rats were tested for spatial learning and hippocampal-dependent memory using MWM tests. As shown in Fig. 2A, all groups gradually decreased their escape latency. Specially, rats in the D-gal group found the platform significantly slower than those in the Control group on day 5 ((26.00 ± 1.12) s vs. (20.00 ± 5.99) s, P < 0.001). Meanwhile, the HLZJSF group shortened significantly escape latency in the spatial learning-memory test compared to the D-gal group on day 5 ((20.44 ± 2.56) s vs. (26.00 ± 1.12) s, P < 0.05), while the escape latency of LLZJSF group tended to shorten with no significant difference ((26.00 ± 1.12) s vs. (22.61 ± 5.29) s, P > 0.05). In addition, the swimming speed of HLZJSF group rats was significantly accelerated (Fig. 2B). In the spatial probe test, the HLZJSF significantly increased the distance and time in the target quadrant as well as the number of platform crossings in the aging rats compared to the D-gal group. (Fig. 2C, D). These results indicated HLZJSF had the protective effect on age-induced decline in spatial memory capacity.

Close

Fig. 2

Effects of LZJSF on cognitive impairment and oxidative stress in D-gal-induced aging rats. (A) Escape latency and (B) Swimming speed of each group of rats; (C) Swimming paths of representative rats in each group. (D) Time spent in target quadrant; distance in the target quadrant; and the number of platform crossings in 60 s in the spatial probe test. (E) Effects of LZJSF on the levels of SOD, MDA, GSH-Px, CAT and T-AOC in brain and serum. Data represent the mean ± SD (n = 9/group). ^#P < 0.05, ^##P < 0.01, ^###P < 0.005, compared to the Control group; *P < 0.05, **P < 0.01, ***P < 0.001, compared to the D-gal group.

Export to PPT

3.3 Effect of LZJSF on oxidative stress in D-gal induced aging rats

In comparison to the Control group, the D-gal group rats had a significant decrease in SOD (P < 0.001), GSH-Px (P < 0.001), T-AOC (P < 0.05) and CAT (P < 0.01) contents in the brain, while a significant increase in the MDA levels (P < 0.01), indicating that the D-gal group rats had oxidative damage in the brain (Fig. 2E). Meanwhile, a protective effect of HLZJSF was observed against D-gal-induced brain damage, supported by significantly higher SOD (P < 0.05), GSH-Px (P < 0.01), T-AOC (P < 0.001) and CAT (P < 0.05) level and lower MDA in brain (P < 0.001). LLZJSF group exhibited a recovery trend in SOD, GSH-Px and CAT level, but only the effects on MDA (P < 0.01) and T-AOC (P < 0.001) had significant difference. Except for MDA and GSH-Px, the profiles of antioxidant-related indicators in the serum were consistent with those in the brain after administration of HLZJSF. These results suggested that the intake of LZJSF significantly improved the antioxidant ability in D-gal-induced aging mice. Notably, stronger effects at the HLZJSF group than at LLZJSF were observed, which was consistent with the behavioral results above. Thus, HLZJSF group was selected for further mechanism study.

3.4 Serum and urine metabolomics study

3.4.1

3.4.1 Effect of HLZJSF on serum and urine metabolic profiles

The urine samples on the 35th and serum samples on the 36th collected from each group were analyzed by ¹H NMR, respectively. The typical ¹H NMR spectra of the serum and the urine of Control group, D-gal group and HLZJSF group rats are shown in Fig. S2A and S2B, respectively. The chromatograms of the three serum groups displayed negligible disparity, despite variations in peak intensities. Similar results were found in three groups of urine samples. Based on the unique chemical shifts and the coupled constants of the compounds, 30 serum metabolites and 30 urine metabolites were tentatively identified through the integration references, HMDB (Human Metabolome Database) and RMRB (Biological Magnetic Resonance Bank) databases, respectively. The detailed information is listed in Table S1 and Table S2.

By firstly using of PLS-DA 3D score plot (Fig. 3A), the metabolic profile of rats in the D-gal group deviated from that of the Control group, indicating that D-gal induced significant biochemical changes in the rats. The metabolic profiles in the HLZJSF group exhibited significant differences compared to those in the D-gal group, but were similar to those in the Control group. For urine sample, the similar results were observed in HLZJSF treated group, which was close to the Control group with apparent classification from D-gal group (Fig. 3B).

Close

Fig. 3

Multiple statistical analysis from ¹H NMR spectra. PLS-DA 3D score plots of the serum samples (A) and the urine samples (B). The regression curves for N/N₀ (Ratio of the number of variables satisfying the factor threshold to the number of preliminary variables) and factor thresholds (VIP, FC, -LNP and |Pcorr|) in serum samples (C). Variables were screened by single factor in serum samples: the R²X, R²Y, and Q² of the PLS-DA model were performed on the variable sets selected by screening the thresholds of VIP, FC, -LNP and |Pcorr|, respectively (D). The Venn plots of the variable meeting the requirements of combined factors in serum. The optimal combination of factors for screening biomarkers was marked in red (E). The scores plot of PLS-DA models performed by the best combination of factors in serum samples(F). The best combination was tested by PLS-DA model permutation (G). (n = 6).

Export to PPT

3.4.3

3.4.3 Effect of HLZJSF on the altered serum and urine metabolites and metabolic pathway in D-gal-induced aging rats

The variation tendency of the identified potential serum metabolites was compared among the three experimental groups using a one-way ANOVA. As shown in Fig. 4A, a significant increase in leucine, valine, isoleucine, arginine, lysine and glutamate was observed in the D-gal group compared to the Control group. Conversely, HLZJSF could remarkably reverse the abnormities of those six serum metabolites. For urine sample, the D-gal group had significant reductions in leucine, creatinine and glutamine, while a significant increase in the xanthine, alanine and hippurate levels compared to the Control group (Fig. S4D). The deviations induced by D-gal of xanthine, alanine and hippurate were corrected with HLZJSF treatment.

Subsequently, perturbed metabolic pathways in response to D-gal-induced-aging were mapped on KEGG (impact values > 0.1). As shown in Fig. 4B, six serum metabolites were involved into four different metabolic pathways in D-gal-induced aging rats, namely D-Glutamine and D-glutamate, alanine, aspartate and glutamate, arginine and proline metabolism, as well as arginine biosynthesis. Six potential urine metabolites in D-gal-induced aging rats involved in D-Glutamine and D-glutamate, alanine, aspartate and glutamate metabolism and arginine biosynthesis (Fig. 4C). Meanwhile, the analysis revealed that HLZJSF intervention in the D-gal-induced-aging model rats targeted four serum metabolic pathways and three urine metabolic pathways, respectively (Table S3).

Close

Fig. 4

(A) Comparison of the relative intensity of potential biomarkers in the aging rats associated with HLLZSF treatment. ^#P < 0.05, ^##P < 0.01, ^###P < 0.005, compared to the Control group; *P < 0.05, **P < 0.01, ***P < 0.001, compared to the D-gal group. (B) Serum metabolic pathways of _D-galactose rats compared with control rats as visualized by bubble plots. Bubble size is proportional to the impact of each pathway and bubble color denoting the significance from highest in red to lowest in white. (C) Urine metabolic pathway analysis of _D-galactose rats compared with control rats as visualized by bubble plots. (D) The holistic metabolic network map of the differential metabolites in serum and urine according to the KEGG PATHWAY database (n = 6). Metabolite names in green, red and blue mean they were detected in both serum and urine, only in serum and only in urine, respectively.

Export to PPT

Finally, an integrated serum and urine metabolic pathways in aging rats associated with HLZJSF were further investigated. As shown in Fig. 4D, a holistic network map was constructed according to the KEGG (https://www.kegg.com) and references (Zhao et al., 2020). Notably, treatment with HLZJSF regulated nine differential metabolites involved in the four altered metabolic pathways.

3.5 Effect of HLZJSF on the regulation of gut microbiota structure

Accumulating reports pointed that gut microbiota dysbiosis and aging were interconnected (Chen et al., 2018). To further demonstrate the crucial role of the gut microbiota in mediating aging induced cognitive dysfunction, the study of the intake of HLZJSF on the gut microbiome in rats with cognitive dysfunction was performed. A total of 908,263 sequences were gained among three group by high-throughput sequencing. The results demonstrated that both the Rarefaction and Shannon index curves approached the saturation plateau, revealing that high coverage of approximately 99% was attained in all samples (Fig. S5A). The ACE, Chao1, Shannon and Simpson indexes were used to analyze the alpha diversity of gut microbiota. Compared with the Control group, ACE, Chao1, Shannon and Simpson index were significantly decreased in D-gal group, indicating a lower microbiota community diversity induced by D-gal (Fig. 5A). The intake of HLZJSF largely restored the gut microbiota α-diversity. To better compare microbiota community differentiation among the three groups, PLS-DA score plot was performed on the OTU abundance. As shown in Fig. S5B, the Control and D-gal groups were clearly separated. The HLZJSF group was positioned between Control and D-gal groups, in close proximity to the Control group, suggesting that HLZJSF could restore gut microbiome disorders in D-gal-induced aging rats. These findings implied that HLZJSF could improve microbial diversity and uniformity.

Close

Fig. 5

Analysis of the diversity, composition and differences of gut microbiota in feces samples. (A) Alpha diversity indexes in each group. (B) Circos plot showing gut microbiota taxonomic profiling at the phylum level in three groups. (C) The relative abundance of major gut microbiota at the phylum level; F/B represents the ratio of Firmicutes-to-Bacteroidetes. ^#P < 0.05, ^##P < 0.01, ^###P < 0.005, compared to the Control group; *P < 0.05, **P < 0.01, ***P < 0.001, compared to the D-gal group. (D) LEfSe analysis of bacterial taxa (n = 4). Histogram of LDA displayed the most enriched bacterial taxa in each group (LDA score threshold > 2.0).

Export to PPT

At the phylum level, Firmicutes, Bacteroidetes, Verrucomicrobia, Actinobacteriota and Proteobacteria were the dominant microbiota in the rat feces of three groups (Fig. 5B). HLZJSF administration resulted in the reduction of the relative abundance of Firmicutes, Actinobacteriota and an increase in Bacteroidetes and Proteobacteria abundance compared to D-gal group, as shown in Fig. 5C. In addition, the HLZJSF group exhibited a significant increase in Firmicutes/Bacteroidetes (F/B) ratios compared to the D-gal group. At the family level, the abundant microbiota taxa among the three groups included Lactobacillaceae, Unclassified_Clostridiales, S24-7, Ruminococcaceae, Prevotellaceae, and Verrucomicrobiaceae (Fig. S6A). Meantime, we observed that a greater relative abundance of Lactobacillaceae and a lower relative abundance of S24-7 in the D-gal group compared to the Control group. With HLZJSF administration, the abundances of Lactobacillaceae and S24-7 were dramatically reduced and enhanced, respectively (Fig. S6B).

Furthermore, LEfSe analysis was also used to analyze gut microbiota diversity among different groups. The histogram of LDA revealed that the presence of 12 distinct gut microbiota, including p_Firmicutes, p_Bacteroidetes, p_Actinobacteria, p_Proteobacteria, f_Coriobacteriaceae, f_Lachnospiraceae, f_Turicibacteraceae, g_Unclassified_Peptostreptococcaceae, g_Clostridium, g_Unclassified_Lachnospiraceae, g_Adlercreutzia and g_Unclassified_RF39 (Fig. 5D).

3.6 Correlation analysis between key bacterial taxa and host parameters

Spearman correlation analysis was conducted to examine the relationship between changes in gut microbiota composition and age-related characteristics, using a significance threshold of |r| > 0.5 and P < 0.05. As shown in Fig. 6A, the main altered genera belonging to the phylum Firmicutes were found to be highly associated with swimming speed of aging rats in MWM test. More specially, a significantly negative correlation was observed between Clostridium (members of phylum Firmicutes), Actinobacteria and swimming speed.

Additionally, certain differential genera assigning to the phylum Firmicutes (member of genus Clostridium) and Actinobacteria exhibited significant negative correlation with brain oxidative indexes (GSH-Px, CAT and T-AOC). In contrast, positive correlations were presented with brain MDA. In addition, the relative abundance of Bacteroidetes were positively correlated with brain SOD, GSH-Px and T-AOC, while negative correlations were observed with brain MDA. Bacteroidetes, Proteobacteria and Unclassified_Lachnospiraceae showed strong positive associations with differential metabolites (such as leucine, valine, isoleucine, arginine, lysine and glutamate), whereas Firmicutes and Unclassified_RF39 showed the opposite correlation. Notably, glutamate was significantly associated with the eleven key bacterial taxa except for Turicibacteraceae (members of phylum Firmicutes). Twelve key bacterial taxa, two phenotypes, five brain oxidative indexes and nine differential metabolites were integrated into the network, respectively (Fig. 6B).

4.1 LZJSF mitigated cognitive dysfunction in D-gal-induced aging rats

Numerous studies had demonstrated that aging was accompanied by a decline in memory and cognitive abilities (Ruan et al., 2013). Consistent with a previous study (Li et al., 2023a), we observed that D-gal-induced aging rats had a significant reduction of the swimming speed, time and distance in the target quadrant and the number of platform crossings in the MWM test, indicating that D-gal impaired memory and cognition in aging rats. HLZJSF significantly increased all the aforementioned indexes and showed beneficial effects in mitigating cognitive dysfunction, which provided new evidence to reveal candidate bioactive compounds by which LZJS regulated aging. Previous studies had shown that aging was closely related to oxidative stress (Ionescu-Tucker and Cotman, 2021). Emerging evidence suggested that accumulated oxidative stress may be one of the key mechanisms causing cognitive dysfunction and neurodegenerative diseases (Birla et al., 2020). Our previous studies showed that LZJSF exhibited antioxidant activity in vitro through decreasing 1,1-Diphenyl-2-picrylhydrazyl free radical (DPPH^·) free radical, superoxide anion (O₂⁻) radical scavenging, as well as ferric reducing antioxidant power (FRAP) at chemical levels and reducing MDA and reactive oxygen species (ROS) at PC12 cell level (Yan et al., 2020). Consistent with our previous results, we also found that HLZJSF effectively reduced the high levels of oxidative stress in D-gal-induced aging rats by increasing the activity of antioxidant enzymes SOD, CAT, T-AOC and GSH-Px in the brain and decreasing the accumulation of toxic oxidation products MDA, which was helpful for delaying the progress of cognitive ability and preventing neurodegenerative diseases. In addition, the effects of HLZJSF on improving learning-memory and regulating oxidative stress indexes were superior to those of LLZJSF. Thus, the mechanism study of HLZJSF on delaying the progress of cognitive ability was further analyzed.

4.2 HLZJSF regulated the metabolic abnormalities of serum and urine in D-gal-induced aging rats

Cognitive dysfunction caused by aging was closely related to the body's metabolism (Kondoh et al, 2020). In the present study, D-Glutamine and D-glutamate, alanine, aspartate and glutamate, arginine and proline metabolism, and arginine biosynthesis were considered as the significant pathways associated with the regulatory effects of HLZJSF based on the results of the serum and urine metabolomics.

D-Glutamine and D-glutamate metabolism made an important impact on the process of aging (Zou et al., 2010). Glutamate, an important excitatory neurotransmitter, was involved in a variety of age-related neurodegenerative diseases (Cox et al., 2022). In addition, the dysregulation of glutamate signaling could lead to cognitive ability (Chang et al., 2020). An increasing body of research indicated the significant involvement of arginine and proline metabolism in the aging process (Phang, 2019, Zhao, et al., 2020). In the present study, three important metabolites, including arginine, glutamate and creatinine, were involved in the disturbances of those pathways induced by D-gal. Arginine served as the central substrate in the pathway and acted as the exclusive precursor of nitric oxide (NO). NO led to the generation of reactive oxygen species and was closely associated with arginine content during the natural aging process (Shannon et al., 2022). In our study, arginine and glutamate levels were elevated in the serum of D-gal-induced aging rats compared to the Control group. Meanwhile, treatment with HLZJSF significantly reduced serum arginine and glutamate levels.

In addition, isoleucine, leucine and valine were branched-chain amino acids that played a key physiological role in metabolism and aging (Le Couteur et al., 2020). (Trautman et al., 2022) had reported that reducing branched-chain amino acids could delay aging and prevent aging-related diseases. Our study showed that a significant increase in serum isoleucine, leucine, and valine in rats induced by D-gal, aligning with the finding of a previous report (Juricic et al., 2020). HLZJSF treatment returned the levels of branched-chain amino acids. In conclusion, HLZJSF might ameliorate the imbalanced metabolic disorders of amino acids and oxidative stress, which contributed to improve and alleviate memory and cognitive abilities of D-gal-induced aging rats.

4.3 HLZJSF regulated gut microbiota dysbiosis in D-gal-induced aging rats

The gut microbiota had been referred to as the “forgotten organ” due to its crucial role in maintaining bodily homeostasis and contributing to improve cognitive ability (Liu et al., 2022a). In 2018, Lu and his colleagues (Lu et al., 2018) showed significant deficits of memory in germ-free mice, which supported the important role of the microbiota in memory development. An increasing body of research indicated that the age-related changes in the gut microbiota composition include a decline in microbiota diversity, a decrease in the abundance of glycolytic bacteria and an increase in proteolytic bacteria (Vaiserman et al., 2017). In the present study, HLZJSF significantly increased the α-diversity and β-diversity, indicating that HLZJSF could alleviate the disorders of gut microbiota induced by D-gal. Therefore, we also explored the effect of HLZJSF on the gut microbiota to reveal the molecular mechanisms by which HLZJSF mitigated cognitive ability. Firmicutes and Bacteroidetes were enriched in elderly people gut microbiota, and smaller proportions of Proteobacteria, Actinobacteria and Verrucomicrobia were observed (Białecka-Dębek et al., 2021). Bacteroidetes were one of the most abundant in gut bacteria and were closely related to aging (Luo et al., 2020). In this study, HLJZSF could largely restored the gut microbiota structures, decreased the relative abundance of Bacteroidetes and increased the relative abundance of Firmicutes and the ratio of F/B. Notably, we also observed a significantly reduction in the abundance of Proteobacteria in D-gal-induced aging rats, which was consistent with the research report (Shin et al., 2015). Conversely, HLZJSF treatment recovered the abundance of Proteobacteria. Importantly, the six of 12 differential taxa on the family level and five of 12 on the genus level associated with HLZJSF belonged to phylum Firmicutes and Actinobacteria. These results suggested that HLZJSF may play a beneficial role in aging prevention by regulating gut microbiota dysregulation.

Correlation analysis further verified that the changed abundance of gut microbiota was connected with cognitive ability. Previous studies suggested that gut microbiota dysbiosis was associated with cognitive ability during aging (Atienza et al., 2018, Yang et al., 2019). Hoffman et al. found that decreased Bacteroidetes were associated with impaired cognitive ability in aging mice (Hoffman et al., 2017). Consistently, we found that Bacteroidetes showed significant positive associations with swimming speed, suggesting HLZJSF improved cognitive ability by regulating Bacteroidetes. Previous research showed that Clostridium was a harmful bacterium to health (Amrane et al. 2018). In the present study, Clostridium was negatively correlated with swimming speed, which was consistent with previous studies (Xie et al., 2020). These results suggested that HLZJSF improved cognitive ability by regulating gut microbiota.

Experimental evidence suggested that a complex association between commensal gut microbiota and oxidative indexes (Sun et al., 2022). Especially, the main altered microbiota taxa belonging to the phylum level including Firmicutes, Bacteroidetes, Actinobacteria and Proteobacteria, which were highly correlated with the oxidative indexes. Previous studies had reported that the higher abundance of Bacteroidetes could increase intestinal wall thickness to reduce oxidative stress (Xie et al., 2020). In the present study, we found Bacteroidetes exhibited significant positive correlations with brain SOD, GSH-Px and T-AOC, while exhibiting negative correlations with brain MDA. Importantly, the serum SOD, CAT and T-AOC levels were enhanced after administration of HLZJSF. Unclassified_Lachnospiraceae (members of Firmicutes) was one of the important microbiotas at the genus level and was closely related to aging (Sheng et al., 2022). Previous studies had reported that supplementation with unclassified_Lachnospiraceae could effectively mitigate oxidative stress within the body in aging (Li et al., 2019b), and unclassified_Lachnospiraceae was closely related to cognitive impairment (Li et al., 2023b). In consistent, the relative abundance of unclassified_Lachnospiraceae was positively correlated with brain SOD and T-AOC. In addition, previous research had showed that Actinobacteria, closely related to aging, was one of the major gut microbiotas in D-gal-induced mice (Liu et al., 2023b). Recent studies had reported that Actinobacteria were potentially important pathogens in humans and caused oxidative stress to immune function (Barka et al., 2016). Coincidentally, we found that Actinobacteria was negatively correlated with brain oxidative indexes (such as GSH-Px, CAT and T-AOC), while an opposite trend with brain MDA was observed in the present study. These results suggested that HLZJSF may play a beneficial role in anti-aging by regulating gut microbiota dysregulation, thereby reduced the high levels of oxidative stress.

In addition, studies had reported that the homeostasis of the gut microflora affected metabolites and led to cognitive ability (Liu et al., 2023b). Recent studies showed that glutamate was a key signaling metabolite which linked gut microbiota to the brain (Zhao et al., 2023). Kadyan et al. found a significant negative correlation between glutamate and Bacteroides in an aging mouse model (Kadyan et al., 2023). In line with the previous study, we found that glutamate was significantly associated with Bacteroides, Proteobacteria, Firmicutes and Actinobacteria. Previous studies had found that long-term use of probiotics and arginine could treat age-related cognitive ability decline (Joho et al., 2023). In this study, Bacteroidetes and Proteobacteria showed strong positive associations with serum arginine. These results suggested that HLZJSF treatment ameliorated cognitive ability by improving gut microbiota thereby metabolic disorders in D-gal-induced aging rats. Interestingly, we also found that Bacteroidetes and Unclassified_Lachnospiraceae showed strong positive associations with branched-chain amino acids (such as leucine, valine, isoleucine), suggesting HLZJSF could improve cognitive ability in D-gal-induced aging rats by effectively regulating gut microbiota and branched-chain amino acid metabolism. However, the mechanism needed to be studied in the further research.

Moreover, LZJSF extracts were rich in quercetin- and kaempferol- type flavonoids. According to in vitro fecal incubations study, LZJSF were easily metabolized to their aglycones (quercetin and kaempferol) via deglycosylation and dehydroxylation reactions (Yan et al., 2020), which might support the finding here that LZJSF mitigated aging-associated cognitive ability via the microbiota-gut-brain axis. Moreover, a number of metabolites originating from LZJSF have been reported to be have a higher antioxidant activity and better absorption than flavonoid glucosides, suggesting that the LZJSF metabolites may also participate in the anti-aging effect of LZJSF.

Hence, we have preliminarily proven that HLZJSF might improve cognitive impairment and delay brain aging by regulating the disturbance of gut microbiota through the “microbiota-gut-brain” axis to affect amino acid and regulate oxidative stress (Fig. 7). In addition, the current study had some limitations. In gut microbiota study, the group size of rat was small, therefore, the validations at the genes and protein levels were required for a further study. More work (the functions of affected genes) needed to be done to completely clarify the mechanism underlying the antiaging effect of LZJSF.

Close

Fig. 6

(A) Spearman correlation analysis of gut microbiota with aging–related features. Purple and yellow colors represent negative and positive correlations, respectively; P value, the depth of color and the size of the circle indicate the degree of correlation; (B) The correlation network of differential gut bacteria and phenotype, brain oxidative indexes, and differential metabolites. Yellow: phenotype; Green: brain oxidative indexes; Purple: differential gut bacteria; Red: differential metabolites; The blue line: negative correlation; The red line: positive correlation. The strength of correlation increases proportionally with the thickness of the line. Significant correlations were marked by *P < 0.05, **P < 0.01 and ***P < 0.001.

Export to PPT

Close

Fig. 7

The scheme of the anti-aging effects of LZJSF from the perspective of the “microbiota-gut-brain” axis.

Export to PPT