UPLC content determination result of THSWD

The content of six components in THSWD water extract was determined by UPLC. The chromatographic results of the sample showed that the retention times of rehmannia glycoside D, amygdalin, hydroxysafflor yellow A, paeoniflorin, ferulic acid and ligustilide were 3.70, 19.36, 21.42, 24.36, 31.65, 41.35 min, respectively. The content determination results were 4.23, 0.11, 1.95, 2.43, 0.57, and 0.076 mg/g, respectively.

The body weight of rats and the results of HE staining of the spleen

The rats in the model group were induced to show signs of blood stasis after bathing in ice water, showing dark lips, dark red around the ears, dark purple nails, and dark red tongues. After the injection of cyclophosphamide, they moved slowly and shrank, the hair was erect and less shiny, and their face, eyes, ears, and tail were pale. Through the growth rate of body weight and spleen index before and after modeling (Fig. 1A, B), compared with the control group, the growth rate of body weight of the model group was small, and there was a significant difference (p < 0.01). The different dosage groups of THSWD were used for modeling. The rate of weight change before and after was higher than that of the model group, and the middle and high dose groups had a significant difference (p < 0.05). Compared with the control group, the spleen index of the model group was significantly lower (p ≤ 0.01). Compared with the model group, the different dosage groups of THSWD can improve the reduction of the spleen index caused by the modeling. The medium and high-dose groups of THSWD had significant differences compared with the model group (p ≤ 0.05).

The results of HE staining sections of the spleen of each group can be shown in Fig. 1C. Compared with the control group, the white pulp volume of the model group is reduced, the marginal band is not obvious (black arrow), the red pulp lymphocytes are greatly reduced, hemorrhage is widely seen (red arrow), and more brown-yellow pigmentation is seen (yellow arrow). Compared with the model group, the white pulp volume, marginal zone, red pulp lymphocytes, hemorrhage, and brown pigmentation in the THSWD administration groups were greatly improved.

Fig. 1

Changes in body weight, growth rate (A), and spleen index of rats in each group before and after modeling (B); STAINING section, results of spleen (C);WBC, PLT, RBC, and HGB levels of rats in each group (D); ≤ Serum hematopoietic related factors in each group of rats (E); Blood coagulation function index of each group (F). Compared with the model group, *p < 0.05, **p < 0.01; compared with the control group, #p < 0.05, ##p < 0.01

Test of blood routine and hematopoietic related factors in rats

The blood routine results of rats in each group are shown in Fig. 1D. The experimental results showed that, compared with the control group, the number of PLT and WBC in the model group was significantly reduced (p ≤ 0.01), an indication of blood deficiency. Compared with the model group, the different dose groups of THSWD can significantly improve the thrombocytopenia caused by modeling (p ≤ 0.05), and can improve the reduction of WBC count caused by modeling to some extent, but there was no significant difference.

The results of rat serum hematopoietic related factors GM-CSF, M-CSF, IL-3, and IL-6 are shown in Fig. 1E. Compared with the control group, the GM-CSF, M-CSF, and IL-3 in the serum of the model group are all significantly decreased (p ≤ 0.05 or p ≤ 0.01); compared with the model group, each group of THSWD can significantly increase GM-CSF, M-CSF, IL-3 in rat serum (p ≤ 0.05 or p ≤ 0.01). Compared with the control group, the rat serum IL-6 in the model group increased significantly (p ≤ 0.01); compared with the model group, each group of THSWD could significantly reduce the IL-6 in the rat serum (p ≤ 0.01).

Detection of hemorheology and coagulation function indexes in rats

The results of plasma and whole blood viscosity of each group are shown in Tables 1 and 2. Compared with the control group, the model group has a significant increase in the viscosity of rat plasma and whole blood. Compared with the model group, each dose group of THSWD had a significant improvement effect (p ≤ 0.05). The coagulation function index results of each group are shown in Fig. 1F. Compared with the control group, the model group can significantly increase the TT, PT, and FIB and significantly reduce the APTT (p ≤ 0.01). It was speculated that the increase of TT and PT may be caused by the decrease of platelet content caused by modeling, while the increase of FIB and decrease of APTT may be caused by the combination of blood deficiency and cold coagulation blood stasis in the rat blood stasis model, caused by changes in its coagulation function. Compared with the model group, the THSWD group with different doses has a significant improvement effect (p ≤ 0.05).

Table 1 Whole blood viscosity results of each group Table 2 Plasma viscosity results of each group Results of plasma metabolomics

When conducting metabolomics research based on mass spectrometry technology, in order to obtain reliable and high-quality metabolomics data, quality control (QC) is usually required. The dense distribution of QC samples on the PCA analysis chart shows that the data is reliable (Additional file 1: Fig. S1). The results of mass spectrometry data prove that the UHPLC-MS technology has good reproducibility. 13,008 precursor molecules were obtained in positive ion mode, and 12,912 precursor molecules were obtained in negative ion mode. The typical basic peak intensity chromatograms of the control group, model group, and THSWD group are shown in Fig. 2A and B. The difference between the groups was analyzed by PLS-DA, and the results showed that the classification effect was significant, and each group was separated from the other (ESI + R2Y = 0.995, Q2 = 0.863: ESI − R2Y = 0.988, Q2 = 0.912) as shown in Fig. 2C and D, revealed the obvious difference between the model group and control group. The THSWD group was close to the control group, which indicates that THSWD can significantly improve the plasma metabolism of rats with blood deficiency and blood stasis. For the identified metabolites, based on the screening indicators of p ≤ 0.05 and VIP ≥ 1, twenty-three different metabolites were identified (Additional file 2: Fig. S2, Additional file 3: Fig. S3), and the heat map of differential metabolites is shown in Fig. 3A.

Next, the identification of metabolites was launched using an online metabolite database and self-built database to identify and screen metabolites through its accurate m/z fragment and MS/MS spectrum. Taking interesting variables (tR/s-m/z 85.9595–90.0554) to form the ESI+ data set as an example, the quasi-molecular ion of m/z 90.0554 [M + H]+. The molecular formula for the interesting variable was speculated to be C3H7NO2 by elemental composition analysis using Masslynx 4.1. The main fragment ions of C3H7NO2 were observed at m/z 72.06, 91.06, 90.06. In these fragments, the m/z 72.06 [M + H-H2O]+ and 91.06 [M + H]+ are dehydrated and hydrolyzed products.

Twenty-three metabolites related to THSWD for improving blood deficiency and blood stasis syndrome were screened and identified. The detailed information is shown in Table 3. THSWD can regulate the metabolic disorder induced by blood deficiency and blood stasis syndrome, and all potential biomarkers after THSWD administration intervention tend to restore the control group.

After modeling, (S)-Methylmalonic acid, a semialdehyde, which is involved in Propanoate metabolism, was significantly increased and beta-Alanine was significantly decreased in rats. Tetracosanoic acid and Docosapentaenoic acid (22n-3), which are involved in the biosynthesis of the unsaturated fatty acid metabolism pathway, were significantly reduced. Deoxyuridine and β-Alanine, which are involved in Pyrimidine metabolism, were significantly reduced. Participate in Synthesis and degradation of ketone bodies (R)-3-Hydroxybutyric acid is significantly increased. Gentiolic acid involved in tyrosine metabolism was significantly reduced. Cyclophosphamide, which is involved in drug metabolism cytochrome P450 and drug metabolism and other enzymes metabolic pathways, was significantly increased, and 6-Methylmercaptopurine was significantly decreased. Cyclic AMP, a different metabolite involved in purine metabolism, was significantly increased. 5-Aminopentanoate, which is involved in the metabolic pathways of Lysine degradation, Arginine, and proline metabolism, was significantly increased. After treatment with THSWD, these indicators can be significantly adjusted.

The MetaPA database was used to analyze the related metabolic pathways of the differential metabolites in the model group and the THSWD administration group. The results showed that THSWD had a significant regulatory effect on the eight imbalanced metabolic pathways caused by blood deficiency and blood stasis syndrome (Fig. 3B), which were phenylalanine, tyrosine and tryptophan biosynthesis, taurine and hypotaurine metabolism, ascorbic acid and alginate metabolism, riboflavin metabolism, biotin metabolism, arginine and proline metabolism, phenylalanine metabolism, pyrimidine metabolism.

Fig. 2

Basic peak chromatograms in positive ion mode (A); basic peak chromatogram in negative ion mode (B); PLS-DA score chart and load chart in positive ion mode (C); PLS-DA score chart and load chart in negative ion mode (D)

Table 3 Differential metabolites in different groups (n = 8) Fig. 3

Heatmap showing the abundance of differential metabolites (A); results of metabolic pathway analysis, Main metabolic pathways interfered by THSWD (B) (a. Phenylalanine, tyrosine, and tryptophan biosynthesis; b. Taurine and hypotaurine metabolism; c. Ascorbate and aldarate metabolism; d. Riboflavin metabolism; e. Biotin metabolism; f. Arginine and proline metabolism; g. Phenylalanine metabolism)

High-throughput sequencing of 16S rDNA

The dilution curve verifies the rationality of the amount of sequencing data and indirectly reflects the abundance of species in the sample. The flat curve indicated that the depth of sequencing is enough (Additional file 4: Fig. S4). With the increase of sequencing time, the number of OTUs in each group also increased, but the increase rate showed a downward trend, indicating that all samples in this study had sufficient sequences, and the number of sequencing reads covered most of the microorganisms in the samples, which can be used for data analysis. Species cumulative boxplots were used to judge whether the sample size was sufficient. Under the premise of sufficient sample size, species richness was predicted using the species cumulative boxplot. As can be seen from the species accumulation boxplot, as the sample size increases, if the position of the graph tends to be flat, it means that the species in the environment does not increase significantly with the increase of the sample size, indicating that the sampling is sufficient to carry out data analysis (Additional file 5: Fig. S5).

In order to explore the relationship between THSWD in the treatment of blood deficiency and blood stasis syndrome and gut microbiota, we analyzed the fecal flora of each group of rats. The species composition of each sample was studied, and the effective sequences of all samples were clustered by OTUs (Operational Taxonomic Units) at a similarity level of 97%, and then the species of OTUs were annotated to obtain 3236 OTUs. According to the results of OTUs obtained by clustering and research needs, analyze the common and unique OTUs between different groups and draw the Venn diagram (Fig. 4A). From the figure, the difference in the number of OTUs in the three groups can be initially seen. The model group has 692 unique OTUs, the control group has 255 unique OTUs, and the THSWD group has 178 unique OTUs. It can be seen from the Venn diagram that, compared with the control group, there were 877 unique OTUs in the model group and 363 unique OTUs in the THSWD group, which was smaller than the difference between the model group and the control group. THSWD can regulate the composition and abundance of intestinal microflora in rats with blood deficiency and blood stasis syndrome, which may be the mechanism of its treatment of blood deficiency and blood stasis syndrome.

The diversity of microbial communities in the samples was analyzed by Alpha Diversity. The Chao1 and ACE indices responded to intestinal community richness, and the Shannon and Simpson indices responded to the diversity of the flora (Table 4). The results showed that Chao1, ACE, and Shannon indices were significantly higher in the model group compared with the control group (p < 0.01); and Chao1, ACE, and Shannon indices were significantly lower in the THSWD administration group compared with the model group (p < 0.01, 0.05). These results indicated that the abundance and diversity of intestinal microbiota increased and the intestinal microbiota became disturbed after the rats were modeled through blood deficiency and blood stasis syndrome. THSWD can promote the decrease of gut microbiota richness and diversity, and may promote its return to a normal state.

Table 4 Influence of THSWD on Alpha Diversity Index of gut microbiota (x ± s, n = 8)

PCoA analysis and UPGMA cluster analysis revealed differences in gut microbiota between the three groups. In the PCoA score chart (Fig. 4B), the percentages explained by PC1 and PC2 to the overall variance are 37.05% and 16.37%. PCoA analysis and UPGMA cluster analysis (Fig. 4D) to investigate the similarity between different samples, and can also construct a cluster tree of samples by clustering the samples. The results show that the control group and the THSWD group are clustered into one category, indicating the relative abundance of gut microbiota in the state of blood deficiency and blood stasis syndrome changes greatly, and the intervention of THSWD can effectively alleviate this change.

Fig. 4

OUTs Venn diagram (A); PCoA analysis diagram (B); UPGMA clustering tree (D); histogram of species abundance in each group Phylum level (C); histogram of species abundance in each Genus level (E)

According to the results of species annotations, select the top 10 species with the largest abundance in each group in Phylum and Genus, and generate a columnar cumulative map of the relative abundance of species to visually view the different classification levels of each sample. The proportion of species with higher relative abundance is shown in Fig. 4C and E.

At the phylum level, Firmicutes and Bacteroidota account for more than 80% of species abundance. Compared with the control group, the relative abundance of Firmicutes in the model group decreased by 21.4%, the relative abundance of Bacteroidota decreased by 7.8%, unidentified_Bacteria increased by 1.4 times, Actinobacteria increased by 1.85 times, Spirochaetota increased by 19 times, Proteobacteria increased by 1.48 times, and Campilobacterota increased by 1.18 Times. Compared with the model group, the relative abundance of Firmicutes in the THSWD group increased by 27.5%, unidentified_Bacteria decreased by 59.5%, Actinobacteria decreased by 29.8%, Spirochaetota decreased by 84.0%, Proteobacteria decreased by 65.2%, and Campilobacterota decreased by 72.1%. At the genus level, compared with the control group, the relative abundance of Lactobacillus in the model group decreased by 65.9%, Prevotella decreased by 89.4%, Prevotellaceae_UCG-003 increased by 33.6 times, and Christensenel decreased by 29.2%; compared with the model group, Lactobacillus increased in THSWD group 32.2%, Prevotella increased by 2.13 times, Prevotellaceae_UCG-003 decreased by 17.9%, and Christensenellaceae_R-7_group increased by 1.3 times.

MetaStat method was used to analyze the abundance data of species between groups. Species with significant differences were screened based on q-values. The relative abundance of each sample group differed significantly at the level of phylum, family, and genus (Table 5). The relativeabundances of Actinobacteriota and Deferribacteres in the model group were significantly higher than the control group at the phylum level. At the family level, the relative abundance of Rikenellaceae, Rs-E47_termite_group, and Prevotellaceae was significantly higher than the model group. At the genus level, the relative abundances of Prevotellaceae_UCG-004, Rikenellaceae_RC9_gut_group, Bilophila, and Odoribacter were significantly higher than the model group, while the relative abundance of [Eubacterium]_siraeum_group, [Bacteroides]_pectinophilus_group, Marvinbryantia were significantly lower.

Table 5 Species relative abundance (percentage) of each group (n = 8) The potential relationship between plasma metabolites and gut microbiota

We analyzed the association between gut bacteria and host metabolism by calculating the Spearman correlation coefficient and found that some flora has a strong correlation with metabolites (p > 0.3 or p < -0.3) (Fig. 5A). The phylum level includes Actinobacteriota, Acidobacteriota, Proteobacteria, Actinobacteria, Desulfobacterota, and the genus level includes Prevotellaceae_UCG-004, Rikenellaceae_RC9_gut_group, [Eubacterium]_siraeum_group, [Bacteroides]_pectinophilus_group, Marvinbryantia. These microflora are closely related to the metabolic pathways and may participate in interference with blood deficiency and blood stasis syndrome. Regarding potential biomarkers, Eugenol, 12-Keto-leukotriene B4, Tetracosanoic acid, 6-Methylmercaptopurine, Cyclic AMP, (R)-3-Hydroxybutyric acid, l-Arogenate, d-Alanyl-d-alanine, P-Aminobenzoic acid, Gentisic acid, 3-Indoleacetonitrile, beta-Alanine, Methyl (indol-3-yl) acetate, Deethylatrazine, Deoxyuridine, Docosapentaenoic acid (22n-3), (S)-Methylmalonic acid semialdehyde, 5-Aminopentanoate, Tetrahydrodipicolinate are closely related to the gut microbiota. The results are shown in Fig. 5A. At the phylum level, l-Arogenate, p-Aminobenzoic acid, and Actinobacteriota were significantly positively correlated, while Cyclic AMP (R)-3-Hydroxybutyric acid, d-Alanyl-d-alanine and Actinobacteriota were significantly negatively correlated. At the genus level, l-Arogenate and p-Aminobenzoic acid were significantly positively correlated with Prevotellaceae_UCG-004 and Rikenellaceae_RC9_gut_group, but were significantly negatively correlated with [Eubacterium]_siraeum_group, [Bacteroides]_pectinophilus_group, Marvinbryantia. It can be seen from the results that the metabolites were closely related to the gut microbiota. THSWD affected the host blood system through the mutual adjustment of these two factors and relieved the blood deficiency and blood stasis syndrome in rats.

Fig. 5

Correlation diagram between the relative abundance of gut microbiota and potential biomarkers, the correlation between the phylum level difference flora and 29 different metabolites; the difference and correlation between genus level flora and 29 different metabolites (A); the color of the box represents the changing trend of the model group and the control group, in which yellow represents promotion, and blue represents inhibition. The arrow on the right of the metabolite in the box indicates the trend of THSWD and the model group; the up arrow represents promotion, and the down arrow represents inhibition (B)