Effect of Tanshinone IIA on Gut Microbiome in Diabetes-Induced Cognitive Impairment

Abstract

Diabetes-induced cognitive impairment (DCI) presents a major public health risk among the aging population. Previous clinical attempts on known therapeutic targets for DCI, such as depleted insulin secretion, insulin resistance, and hyperglycaemia have delivered poor patient outcomes. However, recent evidence has demonstrated that the gut microbiome plays an important role in DCI by modulating cognitive function through the gut-brain crosstalk. The bioactive compound tanshinone IIA (TAN) has shown to improve cognitive and memory function in diabetes mellitus models, though the pharmacological actions are not fully understood. This study aims to investigate the effect and underlying mechanism of TAN in attenuating DCI in relation to regulating the gut microbiome. Metagenomic sequencing analyses were performed on a group of control rats, rats with diabetes induced by a high-fat/high-glucose diet (HFD) and streptozotocin (STZ) (model group) and TAN-treated diabetic rats (TAN group). Cognitive and memory function were assessed by the Morris water maze test, histopathological assessment of brain tissues, and immunoblotting of neurological biomarkers. The fasting blood glucose (FBG) level was monitored throughout the experiments. The levels of serum lipopolysaccharide (LPS) and tumor necrosis factor-α (TNF-α) were measured by enzyme-linked immunoassays to reflect the circulatory inflammation level. The morphology of the colon barrier was observed by histopathological staining. Our study confirmed that TAN reduced the FBG level and improved the cognitive and memory function against HFD- and STZ-induced diabetes. TAN protected the endothelial tight junction in the hippocampus and colon, regulated neuronal biomarkers, and lowered the serum levels of LPS and TNF-α. TAN corrected the reduced abundance of Bacteroidetes in diabetic rats. At the species level, TAN regulated the abundance of B. dorei, Lachnoclostridium sp. YL32 and Clostridiodes difficile. TAN modulated the lipid metabolism and biosynthesis of fatty acids in related pathways as the main functional components. TAN significantly restored the reduced levels of isobutyric acid and butyric acid. Our results supported the use of TAN as a promising therapeutic agent for DCI, in which the underlying mechanism may be associated with gut microbiome regulation.

Keywords: diabetes-induced cognitive impairment; gut microbiome; inflammation; lipid metabolism; short-chain fatty acid; tanshinone IIA.

FIGURE 1

The changes in blood glucose level and cognitive function in response to TAN treatment (n = 7 rats per group). (A) TAN time-dependently reversed the HFD and STZ-induced high blood glucose level (mmol/L) in diabetic rats from week 6 to week 14. ^#### p < 0.0001 vs. CON group at the same time point. *p < 0.05, **p < 0.01 vs. MOD group at the same time point as analysed by two-way repeated measures ANOVA analysis. (B) TAN treatment exhibited time-dependent effect in reducing escape latency time (s) in the training phase from day 1 to day 4. The escape latency time was compared among the groups at the same time point by two-way repeated measures ANOVA. ^#### p < 0.0001 vs. CON group at the same time point. *p < 0.05, **p < 0.01 vs. MOD group at the same time point. (C) Trajectory chart (green color) represents occupancy across the entire 90 s trial in the probe test (day 5). The green circles on the northwest platform refer to the target platform. (D) Statistical analysis by one-way ANOVA showed that TAN treatment significantly increased the time spent in the target quadrant and (E) platform-crossing times compared with the MOD group (F) TAN significantly restored the total distance travelled in 90 s (m) compared with the MOD group. Results are shown as the mean ± standard deviation. ^## p < 0.01, ^### p < 0.001 vs. CON group and *p < 0.05, **p < 0.01 vs. MOD group. TAN, tanshinone IIA; CON, control; MOD, model.

FIGURE 2

Effect of TAN on histopathologic changes in the rat hippocampal CA1 region and colon tight junction. (A) Representative sections of H&E staining (magnification 200×). Red arrow points to the package of neurons. (B) Statistical analysis of the number of neurons in the hippocampal CA1 area (per view, n = 3). ^#### p < 0.0001 vs. CON group and *p < 0.05 vs. MOD group by one-way ANOVA. (C) Representative sections of H&E staining (magnification 200×) on colon tight-junction area in the CON, MOD, and TAN groups. Red arrow represents the damaged villi. TAN treatment reduced plasma levels of LPS (D) and TNF-α (E) in the HFD and STZ-induced diabetic rats, n = 7 rats in each group. ^# p < 0.05, ^## p < 0.01 vs. CON group and *p < 0.05 vs. MOD group by one-way ANOVA analysis. TAN, tanshinone IIA; CON, control; MOD, model; H&E, hematoxylin and eosin; ANOVA, analysis of variance.

FIGURE 3

TAN restored the neurological and colon tight-junction biomarkers. (A) Representative Western blot images and (B) their statistical analysis of protein expressions of ZO-1, occludin, GFAP, p-Tau/Tau in the hippocampus (n = 3 independent experiments). ^# p < 0.05, ^## p < 0.01 vs. CON group and *p < 0.05 vs. MOD group by one-way ANOVA. (C). Representative Western blot images from the three individual experiments and (D) their statistical analysis of protein expressions of ZO-1 and occludin in colon tissues (n ≥ 3). ^# p < 0.05 vs. CON group and *p < 0.05 vs. MOD group by one-way ANOVA. TAN, tanshinone IIA; CON, control; MOD, model; ANOVA, analysis of variance.

FIGURE 4

TAN restored the impaired diversity of the gut microbiome and regulated the gut microbiome at the phylum level (n = 7 rats in each group). (A) The β-diversity analysed by PCoA suggested that the distribution of the dots in the MOD group was distinguished from that in the CON group, whereas most of the samples in the TAN group were distributed similarly as in the CON group. (B) The relative abundance of the three top-ranked phyla among the three groups. (C) The relative abundance (log10) of Firmicutes and Bacteroidetes and their ratio compared among the three groups. The difference between the groups was assessed by one-way analysis of variance. ^# p < 0.05 vs. CON group and **p < 0.01, ***p < 0.001 vs. MOD group. (D) Feature phyla in the CON, MOD, and TAN groups at the LDA2 level as analysed by LEfSe. (E) Feature phyla in the TAN group at the LDA4 level as analysed by LEfSe. TAN, tanshinone IIA; CON, control; MOD, model; LDA, linear discriminant analysis; LEfSe, LDA effect size.

FIGURE 5

The effect of TAN on the abundance of the gut microbiome at the genus level (n = 7 rats in each group). (A) The relative abundance of the top-ranked genera among the three groups. (B) Hierarchical clustering heat map of the top differentiated taxa at the genus level among the three groups. (C) Feature genera in the MOD and TAN groups at the LDA4 level as analysed by LDA effect size. There were no feature genera detected in the CON group at the LDA4 level. (D) The comparison of the relative abundance (log10) of Lachnoclostridium, Clostridioides, Anaerotignum, and Acetivibrio in the three groups. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001 vs. CON group and *p < 0.05, ***p < 0.001 vs. MOD group. Differences were examined by unpaired t test. TAN, tanshinone IIA; CON, control; MOD, model; LDA, linear discriminant analysis.

FIGURE 6

The effect of TAN on the abundance of the gut microbiome at the species level (n = 7 rats in each group). (A) The relative abundance of the top-ranked species among the three groups. (B) Feature species in the MOD and TAN groups at the LDA4 level as analysed by LEfSe. There was no feature species detected in the CON group at the LDA4 level. (C) The comparison of the relative abundance (log10) of Bacteroidetes dorei, Lachnoclostridium sp. YL32, and Clostridiodes difficile in the three groups. ^# p < 0.05, ^## p < 0.01 vs. CON group and *p < 0.05, **p < 0.01, ****p < 0.0001 vs. MOD group. Differences were examined by unpaired t test. TAN, tanshinone IIA; CON, control; MOD, model; LDA, linear discriminant analysis.

FIGURE 7

Functional components in the gut microbiome that were modulated by TAN against HFD and STZ-induced diabetes using the HUMAnN2 program (n = 7 rats in each group). (A) Altered KEGG pathways (level 1) among the three groups as shown in the Circos plots. (B) Heat map of the altered KEGG pathways (level 2) among the three groups. (C) Heat map of the altered KEGG mapping among the three groups. (D) KEGG map00061 with altered genes in the CON and TAN groups. TAN, tanshinone IIA; KEGG, Kyoto Encyclopedia of Genes and Genomes.

FIGURE 8

TAN restored the fecal SCFAs in diabetic rats, and the correlation between the changed species responding to TAN and the diabetes-related index (n = 7 rats in each group). (A) Comparison of identified SCFAs in the three groups analysed by one-way ANNOVA test. ^# p < 0.05, ^### p < 0.001 vs. CON group and *p < 0.05, **p < 0.01 vs. MOD group. (B) Correlation heat map of changed species responding to TAN and the diabetes-related index, including FBG, MWM (platform-crossing times), LPS, TNF-α, butyric acid, lipid metabolism, and map 00061. Red color represents positive correlation, and blue color represents negative correlation. TAN, tanshinone IIA; SCFA, short-chain fatty acid.