Silybin, oridonin, sennoside A, dioscin, oxymatrine, phillyrin, and schizandrin B were purchased from Topscience. Doxorubicin was purchased from HARVEYBIO. In in vitro experiments, DMSO was used to dissolve silybin, oridonin, sennoside A, dioscin, oxymatrine, phillyrin, and schizandrin B to prepare mother solutions, which were diluted for use. In in-vivo experiments, oridonin was dissolved in 0.5% sodium carboxymethyl cellulose.

Both Sch9 yeast mutants and wild-type yeasts were derived from the library, Krogen lab. Standard SD medium (1% yeast extract, 1% bacterial protein, and 2% glucose) for all yeast strains cultured on a rotating shaker at 30°C and 250 rpm. Lifespan determination when yeast is in the logarithmic growth phase. As described in the previous reports (Xie et al., 2012), we observed the mother cells using a repeated microscope for 2 days. The survival curve is based on data collected from multiple experiments and corresponding controls. Survival analysis and cell cycle analysis were conducted using MATLAB.

WI-38 and 2BS cells were provided by the National Institutes for Food and Drug Control. The cells were counted and inoculated for each passage at the same density, and incubated in MEM (Gibco) supplemented with 10% FBS (Gibco) in an incubator at 37°C and 5% CO₂. WI-38 and 2BS cells before PD 30 were considered young. For the induction of premature senescence, WI-38 and 2BS cells were induced with 1 μM doxorubicin for 12 h. Oridonin treats cells for 24–48 h (Dang et al., 2020).

The cell viability was evaluated by the CCK-8 detection kit (Dojindo). 2BS or WI-38 cells were seeded into 96-well plates (5000/well), and then treated with drugs such as oridonin for 24 h. 1:10 diluted CCK-8 solution was added to the culture medium and incubated for 1 h at 37°C. Then a microplate reader (Biotek, MQX200) was used to measure the absorbance at 450 nm. Cell   survival   rate ( % ) = ( OD   treatment - OD blank ) / ( OD control - OD blank ) × 100 %

The cells were washed twice with PBS, then fixed with 4% formaldehyde at room temperature for 15 min, and washed three times. Then the cells were stained with a staining buffer {[1 mg/ml 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-gal) at 37°C, 40 mM citric acid/Sodium phosphate, pH 6.0] stain overnight, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM NaCl, 2 mM MgCl₂}, and then analyzed according to the instructions provided (CST, 9860S). Use ImageJ software to count SA-β-gal positive cells.

WI-38 cells were first induced with 1 μM doxorubicin for 12 h, and then treated with oridonin or DMSO-containing medium for 48 h. WI-38 cells homogenized in RIPA lysis buffer (Thermal Scientific) with protease inhibitors (Roche) and 1% phosphatase inhibitors (Applygen). The protein extracts were homogenized using a Dounce homogenizer and spun at 12,000 g for 20 min at 4°C. BCA analysis (Pierce) was used for protein quantification, and then SDS-PAGE was used to separate the proteins and then transfer to a PVDF membrane (Amersham Biosciences). The botting membranes were incubated with primary antibodies against ACTIN, p-FOXO1, FOXO1, p21, p53, IL-1α or IL-1β (ProteinTech), IL-6 (ABclonal), and IL-8 (Immunoway) overnight at 4°C. Secondary goat anti-rabbit or anti-mouse IgG (Scicrest) was added for blotting. ECL Plus kit (Amersham Biosciences) was used for visualization. The optical density of each band was used to quantify relative protein level.

TRIzol was used to lyse cells, and then isopropanol and chloroform were used to extract RNA. Using cDNA synthesis kit, a total of 1 μg RNA was used for cDNA reverse transcription (Abmgood). Real-time qPCR was used to evaluate the target gene’s expression and the expression level was normalized to GAPDH. Supplementary Table S1 lists all primers.

All naturally aged mice were purchased from SPF Biotechnology Co., Ltd., Beijing, China. The mice were kept in an environment with minimum pressure and standard conditions, namely constant temperature and a light/dark cycle of 12:12 h. The animals were randomly assigned to the control group and the treatment group.

Life expectancy test: 16 months old, male C57BL/6J mice were intraperitoneally injected with oridonin for 1 week every month; the dosage was 2 mg/kg for 6 months, and the control group was given the same volume of 0.5% sodium carboxymethyl cellulose. The death time of mice was recorded.

Grip strength test: the mice were placed on top of the grip strength meter, so they grasped the grid with all four paws. The meter was set to “peak tension” (T-PK) mode, the grip of 5 trials was recorded, and the final data were the maximum grip (g).

Learning and memory test: on the first day, the mouse was placed in a frame 40 cm long, 40 cm wide, and 40 cm high to adapt to the environment for 5 min. On the second day, two identical objects are placed in the frame. The mouse explores autonomously for 5 min on the third day. One of the objects was replaced and the time was recorded when the mouse autonomously analyzed new and old objects. Mouse score calculated as DI = (NEWtime-OLDtime)/(NEWtime + OLDtime).

Working memory ability: Y maze test was used. The mouse was placed at the end of one arm. The order was recorded in which the mouse entered each component within 5 min. Alternation is defined as entering three components continuously, such as (1, 2, 3 or 1, 3, 2), the maximum alternation is the sum of the number of component advances -2, and then the percentage is calculated = alternation/maximum alteration×100%.

All mice were purchased from SPF Biotechnology Co., Ltd., Beijing, China. The mice were kept in an environment with minimum pressure and standard conditions, namely constant temperature and a light/dark cycle of 12:12 h. The animals were randomly assigned to the control group and the treatment groups. 2-month-old male C57BL/6J mice were injected intraperitoneally with doxorubicin (10 mg/kg) every 5 days for 3 times. Oridonin (2 mg/kg) was injected intraperitoneally every 3 days. The control group was given 0.5% sodium carboxymethyl cellulose, and the death of the mice was recorded.

Masson’s trichrome staining: to test heart tissue fibrosis, Masson’s trichrome staining kit (Solarbio, G1340) was used. The sections fixed with formalin and embedded in paraffin were dealkylated and rehydrated with 100% alcohol, 95% alcohol and tap water. Then the sections were stained according to the manufacturer’s instructions. Images were collected using a strong light microscope (×10 magnification objective lens) (OLYMPUS, BX43), and analyzed by ImageJ software (NIH) to measure the area of fibrosis and the total area. The percentage of fibrosis was calculated by dividing the fibrosis-positive area by the entire area. Finally, the statistical results are normalized.

H&E staining: The kidney was fixed with 4% formaldehyde to test the tissue morphology, followed by paraffin embedding and sectioning. The sections were stained with H&E for morphological analysis. The renal tubule damage was scored. When the brush border of the renal tubule was lost, the lumen was dilated, and the epithelial cells were flattened and atrophied, the damaged tubules were counted. Scores were calculated according to the following criteria: 0 = normal; 1=< 20%; 2 = 20–40%; 3 = 40–60%; 4 = 60–80%; and 5≥ 80%.

WI-38 cells were first induced with 1 μM doxorubicin for 12 h, and then treated with oridonin or DMSO-containing medium for 48 h. Total RNA was extracted from the tissue using TRIzol® Reagent according to the manufacturer’s instructions (Invitrogen) and genomic DNA was removed using DNase I (TaKara). Then RNA quality was determined by 2100 Bioanalyser (Agilent) and quantified using the ND-2000 (NanoDrop Technologies). Only a high-quality RNA sample (OD260/280 = 1.8–2.2, OD260/230 ≥ 2.0, RIN≥ 6.5, 28S:18S ≥ 1.0, >2 μg) was used to construct the sequencing library. After quantified by TBS380, paired-end RNA-seq sequencing library was sequenced with the Illumina HiSeq X ten/NovaSeq 6000 sequencer (2 × 150 bp read length). RNA sequencing and raw data quality control were performed by Shanghai Majorbio Bio-pharm Technology Co., Ltd. Two threshold criteria for gene set selection: fold up or down-regulation fold change> 1.20 and an associated p-value of less than 0.05.

Wilcoxon rank-sum test was used to compare life span differences. Evaluate the results through a two-tailed Student’s t-test or analysis of variance to compare two and more samples, respectively. The data is the mean ± SEM. p < 0.05 is the significance threshold. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. n.s., not significant. Unless otherwise stated, the results are based on a minimum of three independent experiments. Unless otherwise stated, use GraphPad Prism for statistics.

Considering that muscle is the most abundant tissue in the human body, we analyzed the transcriptional changes of muscle tissue of the elderly compared with the muscle tissue of the young, which was carried out by Todd R Golub et al. (Lamb et al., 2006). Gene set enrichment analysis calculated the anti-aging score (Figures 1A,B). The results revealed 7 top-ranked and commercially available compounds from the subset, including silybin, oridonin, sennoside A, dioscin, oxymatrine, phillyrin, and schizandrin B of natural compounds. To check the antiaging effects of these compounds, we firstly examined the cell viability in human embryonic lung diploid cells 2BS treated with different concentrations of these compounds, and found that oxymatrine, schizandrin B and oridonin significantly enhanced the viability of senescent cells (Figure 1C and Supplementary Figures S1A,B), while silybin, phillyrin, sennoside A and dioscin did not improve the viability (Supplementary Figures S1C–F). To determine whether this observation was limited to a specific cell type, the effect of schizandrin B and oridonin on cell viability was further verified in another senescent cell line embryonic lung fibroblast WI-38. The results demonstrated that oridonin significantly improved the cell viability of senescent WI-38 cells in a dose-dependent manner (Figure 1D), while schizandrin B did not improve the cell viability (Supplementary Figure S1G). Based on these assessments, oridonin, the top-ranked compound from the CMAP analysis (red dots as shown in Figures 1A,B), was chosen as a potent natural anti-aging compound for further study.

To validate the effect of oridonin, we examined the senescence-associated β-galactosidase (SA-β-gal) activity in WI-38 and 2BS cells, which is a widely used marker in different types of senescent cells for characterizing aging in vivo and in vitro. Young WI-38 cells (PD30) had a flat, scattered appearance and were evenly spaced apart, while old cells (PD45) showed senescence characteristics (Figure 1E). Compared with PD45 cells in the elderly control group, only 11.3% of SA-β-gal positive cells were observed in PD30 cells, and the percentage of SA-β-gal positive cells in oridonin-treated WI-38 cells (PD45) was 48.5% (51.5% reduction, Figures 1E,F). Then we used 2BS cells to verify further that oridonin inhibited cellular senescence. Similarly, young 2BS cells (PD30) had a flat, spread appearance and even spacing, while cells in senescent cells (PD45) showed the characteristics of senescence, granular cytoplasm and accumulation of inclusions (Figure 1G). Compared with the PD45 cells, only 9.7% of SA-β-gal positive cells were observed in PD30 cells, and the proportion of SA-β-gal positive cells in senescent cells treated with oridonin was reduced to 76.1% (Figures 1G,H). Collectively, these results demonstrate that oridonin identified by CMAP analysis effectively maintains the viability of senescent cells and delays the replicative senescence progression.

Developing anti-aging drugs is a tedious and time-consuming process, and only long-term life tests can obtain meaningful results. The yeast strain Saccharomyces cerevisiae (Xie et al., 2015) is one of the commonly used model organisms for aging research. The aging mechanism in yeast and other model organisms is also highly conserved. Budding yeast shows progressive senescence similar to that of mammals (Gershon and Gershon, 2000). In our attempt to determine the replicative lifespan of yeast, an automated device based on a microfluidic chip described previously was employed (Jo et al., 2015), which can measure the lifespan of yeast within 3 days.

Yeast was incubated in SD medium for 20 h, and then collected and loaded on the chip. The lifespan was monitored when the oridonin concentration was 0, 0.25, 0.5, 1, 2, 4, and 8 μM. The experimental results showed that oridonin led to a robust lifespan extension of 14.6–48.9% in the BY4741 strain in a dose-dependent manner (Figure 2A) compared with control media. In addition, oridonin treatment significantly reduced the number of cells with longer cell cycle duration (Figures 2B–H), thus decreasing the heterogeneity of cell cycle length, and prolonged lifespan.

To further test whether oridonin can prolong the lifespan of mice, we selected 16-month-old C57BL/6J male mice to conduct lifespan and healthy lifespan tests under the intervention of oridonin. During the experiment, the oridonin treatment group extended the average life span of mice from 394 to 479 days (85 days, approximately 21.6% extension). Throughout the life process, the oridonin treatment group extended the average life span of mice from 874 to 959 days (85 days, approximately 9.7% extension) (Figure 3A). Moreover, we evaluated the physical function of old mice treated with oridonin through functional tests. Compared with the control group, oridonin treatment significantly improved grip strength (Figure 3B) and learning memory ability (Figure 3C). There was no significant difference in the average weekly weight between the control and treatment groups (Figure 3D), indicating the safety of taking 2 mg/kg of oridonin regularly. These results suggest that oridonin improves the lifespan and health of naturally aging mice.

Off-target toxicity limits the maximum tolerated dose of chemotherapeutic drugs and causes long-term health problems in cancer survivors, including accelerated aging (Henderson et al., 2014). The expression of the senescence markers p53 and p21 is upregulated, which leads to the generation of senescence-associated secretory phenotype (SASP) (Van Deursen, 2014; Childs et al., 2017; Wakita et al., 2020). Since chemotherapy can induce senescence (Ewald et al., 2010), we further explored whether oridonin could delay cellular senescence caused by chemotherapy drug doxorubicin. PD30 WI-38 and 2BS cells were treated with 1 μM doxorubicin for 12 h to detect whether oridonin could suppress the cellular senescence. In WI-38 cells, the percentage of SA-β-gal positive cells in the control group was 16.2% of the model group, and the rate of SA-β-gal positive cells in the oridonin-treated model group decreased in a dose-dependent manner (Figures 4A,B). At the same time, we also tested senescence markers, including p21, p53, and SASP. RT-qPCR results showed that compared with the control group, the mRNA and protein expression of IL-1α, IL-1β, IL-6, IL-8, p21, and p53, were upregulated in the model group. After oridonin treatment, the expression of these genes was downregulated (Figures 4C–F and Supplementary Figures S2A–E). Similarly, in 2BS cells, compared with the model group, the SA-β-gal positive cell rate in the control group was 10.0%, and the SA-β-gal positive cell rate in the oridonin-treated model group was significantly reduced (Figures 4G,H). The results indicate that oridonin delays cell senescence induced by doxorubicin.

We speculated that oridonin could extend the lifespan of an accelerated aging mouse model induced by doxorubicin (Dang et al., 2020; Wiley et al., 2021). In this experiment, 8-week-old male C57BL/6J mice have induced senescence with a dose of 10 mg/kg doxorubicin. We found that oridonin treatment extended the average life span of doxorubicin-induced aging mice from 19 to 29 days (10 days, approximately 53.8% extension) compared to the control group (Figure 5A). Notably, mice treated with oridonin had significantly increased grip strength than the control group (Figure 5B), but there was no significant change in body weight (Figure 5C). Doxorubicin has been reported to induce nephrotoxicity and cardiotoxicity (Majumder et al., 2018; Galan-Arriola et al., 2019; Guerrero et al., 2019; Dang et al., 2020; Yang et al., 2020). Masson’s trichrome staining results showed that the degree of cardiac fibrosis of the doxorubicin-treated group was significantly increased compared to control mice, and oridonin treatment significantly attenuated doxorubicin-induced myocardial fibrosis (Figures 5D,E). Hematoxylin and eosin (H&E) staining conducted a histomorphological analysis of the kidney. Compared with the control group, the model group had noticeable damage to the renal tubules, while the oridonin treatment group had significantly less damage (Figures 5F–H). These results suggest that oridonin prolongs the lifespan, improves health span, and improves tissue morphological changes in doxorubicin-induced aging mice.

It has been reported that oridonin is an inhibitor of AKT (Song et al., 2018). We first knocked out the homologous gene Sch9 of AKT in yeast to explore how oridonin plays an antiaging effect and possible targets. Oridonin treatment failed to extend its lifespan in the Sch9 knockout strain. At the same time, the Sch9 knockout strain had a more prolongated lifespan than the wild type. The strain’s lifespan is extended by 42.3% (Figures 6A,B). This result indicates that oridonin prolongs the lifespan of yeast through Sch9, and knocking out Sch9 can prolong the lifespan of yeast.

Western blot detected downstream signaling pathways of AKT and aging-related proteins to confirm that oridonin functions through AKT in human cells. Compared with the control group, the doxorubicin-induced senescent cell model group had upregulated FOXO1 phosphorylation level and p21 expression. In contrast, FOXO1 phosphorylation and p21 expression were downregulated in the oridonin-treated cells, thereby prolonging lifespan (Figures 6C–E). Oridonin has been reported to act directly on NLRP3 inflammasomes and block the assembly of NLRP3 inflammasomes (He et al., 2018). Therefore, we used the Western blot to detect its downstream signal pathways, such as IL-1β. As expected, compared with the control group, the expression of IL-1β in the doxorubicin-induced senescent cell model group was upregulated. In contrast, the expression of IL-1β in the oridonin-treated group was downregulated (Figures 6C–E). These results indicate that oridonin exerts anti-aging effects through AKT and NLRP3.

To further study whether oridonin also affects other signaling pathways. In WI-38 cells, we used RNA-seq to detect genetic changes in young cells, doxorubicin-induced senescent cells, and oridonin-treated doxorubicin-induced senescent cells. Doxorubicin-induced cellular senescence (CVSD) and oridonin intervention (DVSO) had 352 genes that changed in common (Figure 7A). The heatmap found that doxorubicin caused upregulation or downregulation of genes, and oridonin intervention reversed most genes’ changes (Figure 7B). Next, we performed GO enrichment analysis based on these 352 genes, and found the main modifications in cell signal transduction (Supplementary Figure S3A). Therefore, we performed a KEGG enrichment analysis, and found that the FOXO signaling pathway, mTOR signaling pathway had changed (Figures 7C,D), noting that these two signal pathways are also directly related to AKT. Additionally, we also found changes in the T cell receptor, cGMP-PKG, and Hippo signaling pathways (Figures 7C,D). Finally, we used the differentially expressed genes caused by doxorubicin-induced cell senescence (CVSD) and the differentially expressed genes caused by oridonin treatment of senescent cells (DVSO) to perform GO and KEGG enrichment analysis. The same signal transduction-related genes had major changes (Supplementary Figures S3B–D). Oridonin treatment also changed the AGE-RAGE, NF-κB, and TNF signaling pathways (Supplementary Figure S3D). The RT–qPCR results also supported these findings (Figure 7E). Surprisingly, among the differentially expressed genes caused by oridonin, we found that the top-ranked genes included histone and ubiquitin ligase-related genes, and RT–qPCR also confirmed this finding (Figure 7F). This result indicates that oridonin may delay aging by influencing epigenetics. These results indicate that oridonin reversed the genetic changes caused by doxorubicin-induced cell senescence.