ZYP were obtained from the Pharmacy Department of Shenzhen TCM Hospital [Approval number: Guangdong Pharmaceutical Preparation Z20070083]. Its specific formula composition and preparation method refer to our previous reports (Chen et al., 2019). Formic acid (FA), methanol (HPLC grade), and acetonitrile (HPLC grade) were purchased from Thermo Fisher Scientific (Waltham, MA, United States). Ultrapure deionized water was obtained from a Unique-R20 water-purification system (Xiamen, China). MC3T3-E1 cells were purchased from Procell Life Science and Technology Company (Cat# CL-0378, Wuhan, China). Ultrapure deionized water was obtained from a Unique-R20 water-purification system (Xiamen, China).

β-glycerophosphate sodium (#G9422), ascorbic acid (#A4403) and dexamethasone (#D4902) were purchased from Sigma-Aldrich (St. Louis, MO, United States). α-MEM media, fetal bovine serum (FBS) and phosphate-buffered saline (PBS) were purchased from Gibco (Grand Island, NY, United States). The Cell Counting Kit-8 (CCK-8, #C0037), alkaline phosphatase (ALP, #P0321) color development kit, Alizarin red S staining kit (ARS, #c0148), RNAeasy™ Animal RNA Isolation Kit (Beyotime, #R0027) and horseradish peroxidase (HRP)- conjugated goat anti-rabbit or goat anti-mouse antibody (#A0208 or #A0216), BeyoECL plus kit (#P0018) were obtained from Beyotime Biotechnology Co., (Jiangsu, China). Runx2 (#ab192256) and Osterix (#ab209484) primary antibody was purchased from Abcam (Cambridge, United Kingdom). Osteoprotegerin (OPG, #41289) primary antibodies was purchased from Signalway Antibody (CA, United States). Osteopontin (OPN, #22952) primary antibody was purchased from Proteintech company (Rosemont, IL, United States). β-actin (#3700) antibody was obtained from Cell Signaling Technology (Danvers, MA, United States). The PrimerScript RT reagent kit and SYBR qPCR Master Mix Kit were purchased from Vazyme (Biotech Co., Ltd., Nanjing, China).

Phytochemical compounds were analyzed using Waters Acquity HSS T3 column (2.1 × 150 mm, 1.8 μm, Waters corporation, MA, United States) system. The mobile phase was a mixture of 0.1% formic acid in water (A) and acetonitrile (B), with an optimized gradient elution as follows: 0–5 min, 3%–8% B; 5–11 min, 8%–30% B; 11–20 min, 30%–80% B; 20–21 min, 80%–95% B; 21–27 min, 95%–95%. The injection volume was 2 μL, the column temperature was 35°C and flow rate was set at 0.30 mL/min. The MS analysis was performed using an X500B Q-TOF mass spectrometer (AB Sciex, Foster City, CA, United States). The MS data parameters were set as follows: the instrument was operated in positive (5.5 kV) or negative (−4.5 kV) ion mode, the full MS scan range was m/z 100–1,250. Then, some chemical components were identified by comparing with standard reference and an internal database (LibraryView). Others were identified by literature and online databases, such as MetaScope, ChemSpider and Progenesis QI software.

Specific pathogen-free male Sprague Dawley rats (SD, weighing 190–210 g) were provided by Beijing Vital River Laboratory Animal Technology Co., Ltd. The rats were housed in an animal facility with a standard 12 h light/dark cycle under constant temperature conditions and with free access to food and water. The experimental procedures were conducted in accordance with the Ethics Committees of the Chinese University of Hong Kong (Shenzhen, license number: CUHKSZ-AE2023014).

After 1 week of adaptive feeding, all rats were randomly divided into three groups: a control group for collecting blank bio-samples, ZYP -treated group (1.62 g/kg) for collecting plasma, urine, feces, and tissues, and ZYP -treated group (1.62 g/kg) for collecting bile. Rats in the control group were given normal saline by intragastric administration.

In different groups, blood samples were collected at 5, 15, 30, 60, 120, 240, 360, 480 and 600 min after intragastric administration and placed into heparin anticoagulant tubes. The tubes were then centrifuged at 4,500 rpm for 5 min to obtain plasma samples, which were combined for each time point and stored at −80°C until further analysis. Feces, urine and bile acid samples were collected from 0–600 min after administration, and collected every 120 min. After taking the above samples, rats were killed and the tissues were collected.

After mixing 60 μL plasma (600 μL in total) at each time point, 1.8 mL ice acetonitrile was added. The sample was swirled for 2 min and centrifuged (at 13,000 rpm, 10 min, 4°C). Next, 400 μL of the supernatant was removed, and dry with nitrogen. The sample was redissolved with 200 μL 50% acetonitrile, centrifuged for 10 min at 13,000 rpm. 2 μL of the supernatant was analyzed by LC-MS.

A 0.3 g aliquot of tissue or feces sample were mixed with 1 mL methanol and homogenized six times for 30 s each time. The homogenate was centrifuged at 13,000 rpm for 10 min at 4°C. Next, 200 μL tissue supernatant dry with nitrogen. Redissolved with 200 μL 50% acetonitrile, centrifuged for 10 min at 13,000 rpm, which were then absorbed 2 μL supernatant to LC-MS.

Appropriate amount of urine/bile was centrifuged, centrifuged at 4,000 rpm for 10 min. Then, 1 mL of the supernatant was added to the activated C18 solid phase extraction column, washed with 1 mL of pure water, and then eluted with 1 mL of methanol. Finally, the eluent is collected. The MS/MS spectrum, Human Metabolome Database, METLIN database and literature data were used to identify metabolites.

The prescribed dosage for ZYP was set at 18 g/d. Based on the dose equivalency conversion formula between humans and rat presented in the “Pharmacological Experimental Methodology,” a dosage of 1.62 g/kg for ZYP was determined. Ten SD rats were administered with ZYP (0.81 g/kg) via intragastric administration twice a day to gain ZYP-containing serum, while another ten rats were treated with normal saline (10 mL/kg) to gain control serum. After treatment for 3 days, blood samples were collected and then subjected to centrifugation at 3,000 rpm for 10 min at 4°C. Finally, the collected supernatant was inactivated at 56°C for 30 min, sterilized through a 0.22 µm filter, and stored at −80°C for further use.

MC3T3-E1 cells were cultured in α-MEM supplemented with 10% FBS and 1% PS (100 U/mL penicillin and 100 µg/mL streptomycin) at 37°C, 5% CO₂. Subculturing of the cells was performed when they reached 80% confluence, and the medium was refreshed every 2 days. For induction of MC3T3-E1 osteogenic differentiation, the medium was added with 10 mM β-glycerophosphate sodium, 50 μg/L ascorbic acid and 100 μM dexamethasone. ALP staining was performed after 7 days of osteogenic differentiation. The medium was changed every 2 days.

CCK-8 solution was used for the assessment of cell vitality following the instructions. 2 × 10³ cells/well was maintained for the seeding of MC3T3-E1 cells in a 96-well plate for 24 h. Then, the cells were treated with different concentrations of drug-containing serum for 48 h. Cells were divided into the following groups: (1) Control group, (2) ZYP-treated group with 2%, 4% and 8% drug-containing serum, (3) Blank group with 8% control serum. After that, the working solution containing 10% CCK-8 reagent was added to each well and incubated for 1 h. The absorbance was measured at 450 nm with a multifunctional microplate reader.

After inducing osteogenic differentiation for 7 days or 21 days, the ALP and ARS staining in different treatment groups were performed according to the manufacturer’s instructions of the corresponding assay kit. Briefly, the cells were fixed with 4% paraformaldehyde for 30 min. Then, cells were incubated with staining solution at room temperature for 18 h. After incubation, the cells were washed with PBS, and the stained cells were observed under a microscope immediately.

Total RNA was isolated from MC3T3-E1 cells using RNAeasy™ Animal RNA Isolation Kit. Then, RNA was utilized to synthesize cDNA following the manufacturer’s guide. The result of qRT-PCR was analyzed by CFX 96 system (Version 3.1, Bio-Rad Corporation, CA, United States) by using SYBR qPCR Master Mix Kit. The gene expression level was calculated based on the 2^−△△Ct method and β-actin was used as a housekeeping gene. The primer sequences were listed in Table 1.

In brief, cells were lysed in RIPA with 1% PMSF for 25 min, and then centrifuged at 12,000 rpm for 15 min at 4°C. Following that, equal protein (20 µg) was separated using 10% SDS-PAGE and blotted to a PVDF membrane. After transfer, the membrane was blocked with 5% skimmed milk for 1 h and then the following diluted primary antibodies were added at 4°C overnight: anti-Runx2 (1:1,000), anti-OPN (1:2,000), anti-OPG (1:1,000), anti-Osterix (1:1,000), anti-β-actin (1:1,000). Thereafter, the membranes were then incubated with HRP-conjugated goat anti-rabbit or anti-mouse antibody (1:5,000) at room temperature for 50 min. Finally, the chemiluminescence of target protein bands were performed with the enhanced chemiluminescence (ECL) detection system (ProteinSimple, San Jose, CA, United States).

After induction of osteoblast differentiation for 2 days, total RNA from MC3T3-E1 cells was extracted by utilizing Trizol reagent. The RNA-sequencing analysis was conducted and analyzed. Briefly, the extracted RNA was evaluated for quality and integrity, and library preparation was conducted. Differential expression analysis between two groups was performed by employing the DESeq2 R package. Genes with FDR <0.05 and │log2 (foldchange)│≥1 found by DESeq2 were regarded as differentially expressed genes (DEGs). KOBAS software was employed to test the statistical enrichment of DEGs in KEGG pathways.

All values are expressed as mean ± SD (standard deviation). Statistical difference was evaluated using one-way analysis of variance (ANOVA) followed by Tukey’s test (GraphPad Pro Prism 8.0, San Diego, CA). P < 0.05 was considered statistically significant.

As shown in Figures 1A, B, there was the representative base peak chromatogram (BPC) of ZYP in the positive and negative-ion scan. Based on the structural analysis and identification strategy of UPLC-QTOF/MS, the unknown components were classified according to the fragmenting rules and diagnostic ions of different structural types of components. A total of 152 compounds were firstly tentatively identified from identified in ZYP, including 50 flavonoids, 35 triterpenoid saponins, 17 phenylpropanoids, 11 lignans, and 14 others (Figure 1C). The herbs and the number of their components are further summarized in Figure 1D (AR 30, ABR 9, CR 14, RR 15, DR 12, SCF 12, EF 21, PR 16, NRR 19, AR/ABR 1, DR/SCF/CR 1, AR/EF 1, and AR/PR 1). The details of these compounds were listed in Supplementary Table S1.

Subsequently, fifteen chemical components were accurately identified in CR, including amino acids, nucleosides, alkaloids, and glycosides, such as codonopsis. RR also contains fifteen chemical components, with characteristic components including iridoid, phenylpropanoid, and polysaccharides. DR was found to be rich in thirteen compounds, including flavonoids and phenylpropanoids, with naringin as its representative component. Seventeen compounds were identified in PR, with main components being flavonoids such as puerarin and luteolin. Results showed that AR was rich in thirty-three compounds, mainly flavonoids like mullein isoflavones and triterpenoid saponins such as astragalus saponins. Ten compounds were identified from ABR, mainly steroids and saponins. A total of nineteen compounds were identified in NRR, characterized by triterpenoid saponins, especially ginsenosides. Thirteen compounds were identified from SCF, with lignans as the characteristic components. A total of twenty-two flavonoids, such as icariin, were identified as the characteristic components of EF. The representative structures of each herb are depicted in Figure 2.

According to the typical characteristic compounds, the secondary fragmentation and cracking behavior presented by secondary mass spectra were analyzed.

Based on the chemical characterization of ZYP, the MS² fragments and retention time were used to analyze its components in plasma, urine, feces, and bile. Icariin (85) was selected as an example. As shown in Figure 4A, the extracted ion chromatogram (XIC) and secondary spectrum of icariin (RT 14.29 min, [M+FA−H]⁻ m/z 721.2343 [M+H]⁺ m/z 677.2435) in the extraction solution revealed characteristic fragments at m/z 531, 369, and 313. Figure 4B shows the superimposed XIC of Icariin in bile, feces, plasma, and urine from both administered and blank groups. The results indicated that Icariin responded only in the administered plasma at the corresponding retention time (with no response in blank samples), and the primary ppm met the requirements (<10 ppm). The secondary spectra were highly similar to those of Icariin in the extraction solution, with consistent secondary fragments (m/z 531, 369, 313), indicating that Icariin can be absorbed into the blood and remain in the body to exert its effects.

Based on the phase I and phase II metabolism patterns and the similarity of the secondary mass spectrum profiles, metabolite libraries distinct from the prototype components can be quickly screened from the matrix. These metabolites are then automatically matched with the prototype components to assist in their identification and annotation. For example, the mass deviation between Icariin and M69 (m/z 369.1331, RT 20.93 min) was Δm = 308.1107, consistent with the molecular weight loss of Rhamnose and Glucose. This biological transformation pathway conforms to “Rhamnose hydrolysis and Glucose hydrolysis.” The XIC superposition of M69 in each sample, as shown in Figure 4C, indicated a weak response in urine, but its chromatogram m/z was 369.1513 with ppm >10. The secondary profile of M69 in administered feces was similar to Icariin, with consistent fragments, including m/z 369 and m/z 313, indicating that M69 is a metabolite of Icariin. Additionally, M69’s response in stool suggests it may be metabolized by intestinal flora. Figure 4D shows the structural formula of Icariin-M69. Following this principle, a total of six metabolites were matched to Icariin, and their structural association diagrams are presented in Figure 5.

A total of 110, 65, 63, 59, 27 compounds were extracted from feces, plasma, urine, intestine and colon, respectively. And 70 compounds were detected in other tissues (Figure 6A). In detail, 70 prototype compounds were extracted from plasma, urine, feces, or bile (Figure 6B). Among them, 41 were detected in plasma, 25 in urine, 48 in feces, and 1 in bile. Further analysis of these prototype compounds revealed a total of 99 matched metabolites, with 24 detectable in plasma, 38 in urine, 62 in feces, and 3 in bile (Figure 6C). The tissue distribution of these compounds indicated that both prototypes and metabolites could be detected in various tissues, particularly in the intestines, the primary organ for digestion and absorption. Specifically, 18 prototypes and 9 metabolites were detected in the colon, while 36 prototypes and 23 metabolites were detected in the small intestine. These findings suggest that ZYP’s impact on overall host metabolism, through regulation of intestinal flora, may be a significant mechanism of action. The detailed distribution of prototypes and metabolites is presented in Supplementary Tables S2, S3, respectively.

Figure 7 illustrates the association network between the prototypes and their related metabolites. Metabolites detected in feces may be processed by intestinal flora, while those metabolized by the liver are detectable in bile, plasma, and urine. Detailed biotransformation and annotation of prototype and metabolite components are provided in Supplementary Table S4.

Figure 8A is a simple cell study protocols. As shown in Figure 8B, ZYP-containing serum (2%, 4%, and 8%) and 8% blank serum exerted no obvious effect on the viability of MC3T3-E1 cells. In addition, we performed ALP staining to assess the osteogenic differentiation ability of MC3T3-E1 cells. Although the positive rate of ALP and ARS staining in 8% blank serum treated samples was similar to that of the control samples, our results showed that ZYP-containing serum treatment resulted in a higher rate of positive ALP and ARS staining in MC3T3-E1 cells compared to the control group (Figures 8C, D).

The results revealed that the addition of 8% blank serum to MC3T3-E1 cells did not yield any noticeable differences in the mRNA and protein levels of Runx2, Opn, Opg and Osterix when compared with the control group. Conversely, the mRNA and protein levels of Runx2, Opn, Opg and Osterix were elevated by ZYP-containing serum at concentrations of 2%, 4%, and 8% in MC3T3-E1 cells when compared with the blank group (as depicted in Figure 9). Taken together, these findings suggest the role of ZYP-containing serum promotes the levels of osteogenic differentiation-related proteins in MC3T3-E1 cells.

To elucidate the potential mechanism underlying the impact of ZYP-containing serum in promoting the osteogenic differentiation, RNA-seq analysis was further conducted. A total of 5,249 DEGs with 2,700 upregulated and 2,549 downregulated genes were confirmed between MC3T3-E1 treated with or without ZYP-containing serum (Figure 10A). Moreover, the heat map suggested that a set of genes (Cox5a, Cox7a2, Cox7c, Ndufa2, Ndufs8, Atp5g3, Atp5o, Atp5d, Ndufb5, Ndufa9, Ndufa12, Uqcrh, Ndufb7, Ndufb8, Ppa1, Ndufb10 and Ndufab1) of MC3T3-E1 were upregulated by ZYP-containing serum (Figure 10B).

The top 20 enriched KEGG pathways and GO terms (biological process (BP), cellular component (CC), molecular function (MF)) were shown in Figure 11. The results suggested that ZYP-containing serum significantly promoted osteogenic differentiation may be closely related to the upregulation of mitochondrial oxidative phosphorylation.