Semen strychni (SS) was purchased from TongRenTang Pharmaceutical Group (Beijing, China) and validated by Prof. Baojian Wu (Guangzhou University of Chinese Medicine, Guangzhou, China). A voucher specimen (No. 202009-5) for semen strychni was deposited at College of Pharmacy, Jinan University. Tripterygium glycoside tablet (TGT, Z42021212) was purchased from Huangshi Feiyun Pharmaceutical (Hubei, China). Strychnine (purity: 98.52%, MUST-18120411) and brucine (purity: 99.43%, MUST-19090508) were purchased from Must Biotechnology (Chengdu, China). Triptolide (purity: ≥98%, 38748-32-2) and celastrol (purity: ≥98%, 34157-83-0) were purchased from Aladdin Chemicals (Shanghai, China). Biochemical assay kits for creatinine, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were purchased from Jiancheng Bioengineering Institute (Nanjing, China). CK-BB ELISA kit was obtained from Mlbio Bio-technology (Shanghai, China). RNAiso Plus reagent was purchased from Takara (Shiga, Japan). ChamQ Universal SYBR qPCR Master Mix was purchased from Vazyme (Nanjing, China). ELISA kits for TNF-α and IL-6 were obtained from Neobioscience Technology (Shenzhen, China). Freund’s complete adjuvant (CFA) and Freund’s incomplete adjuvant (IFA) were purchased form Yuanye Bio-technology (Shanghai, China).

Raw SS (10 g, the dry ripe seed of Strychnos nux-vomica L.) was ground into fine powder (particle size: 100–200 mesh), and suspended in 50% ethanol (100 ml). This was followed by ultrasonication for three times at room temperature, 1 h each time. After each ultrasound, the filtrate was filtered by filter paper (pore size: 30–50 μm). The combined filtrate (three times, 300 ml) was dried in vacuum at 50°C (Gao et al., 2021). Concentrated extract (15 ml) was stored at 4°C prior to use. TGT was ground to fine powder (particle size: 100–150 mesh) and suspended in 0.5% CMC-Na solution. This was followed by ultrasonication for about 30 min until a uniform dispersing solution was formed. The uniform dispersing solution was stored at 4°C prior to use (Zhao et al., 2020b).

C57BL/6 male mice were obtained from HFK Bio-technology (Beijing, China) and maintained under a 12 h light/12 h dark cycle [light on at 7 AM (= ZT0), light off at 7 PM (= ZT12)] with free access to food and water. Mice aged 8–12 weeks were used for experiments. All experiments were performed using protocols approved by the Institutional Animal Care and Use Committees of Guangzhou University of Chinese Medicine (Guangzhou, China).

Arthritis was induced in mice by immunization with type II collagen dissolved in 0.1 M acetic acid (2 mg/ml) and emulsified with an equal volume of Freund’s adjuvant as previously described (Inglis et al., 2008). On day 0, each mouse was intradermally immunized with Freund’s complete adjuvant (CFA) emulsified collagen II at the tail end and back. On day 21, mice were boosted with Freund’s incomplete adjuvant (IFA) emulsified collagen II. After the second immunization, mice showing macroscopic signs of arthritis (with a success rate of ∼65%) were selected for further study.

FLS cells were isolated from the foot joints of CIA mice as previously described (Armaka et al., 2009). Cells were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum at 37°C in a humidified 5% CO₂ atmosphere. Two sets of assays were performed with FLS. In the first set (cytotoxicity) of assays, cell viability was determined by using cell counting kit-8 (CCK-8) according to the manufacturer’s protocol (Cat. No. GK10001, GLPBIO). In brief, cells were seeded into a 96-well plate. 24 h later, cells were treated with brucine (0, 1.56, 3.13, 6.25, 12.5, 25, 50, 100, 200, and 400 μg/ml) or strychnine (0, 1.56, 3.13, 6.25, 12.5, 25, 50, 100, 200, and 400 μg/ml) or triptolide (0, 1.56, 3.13, 6.25, 12.5, 25, 50, 100, 200, and 400 ng/ml) or vehicle. After another 24 h, cells were collected for viability determination (He et al., 2019; Tang et al., 2019). In the second set of assays, cells were seeded into a 6-well plate. 24 h later, cells were treated with brucine (50 μg/ml) or strychnine (50 μg/ml) or triptolide (6.25 ng/ml) or vehicle. After 24 h, cells were collected for qPCR and determination of inflammatory factors.

Therapeutic effects of SS and TGT were assessed using CIA mice. For dose-response experiments, CIA mice were randomly divided into seven groups (n = 5 mice in each group). These groups of mice were gavaged with SS extract (20, 40, and 80 mg/kg) or TGT suspension (15, 30, and 45 mg/kg) or vehicle at ZT2 once daily for 7 days. On day 8, mice were sacrificed at ZT2 to collect plasma and hind foot samples (Chen et al., 2011). For chronoefficacy experiments, CIA mice were gavaged with SS extract (40 mg/kg) or TGT suspension (30 mg/kg) or vehicle (control group) at each circadian time point (ZT2, ZT6, ZT10, ZT14, ZT18 or ZT22) once daily for 7 days. On day 8, mice were sacrificed at ZT2 to collect plasma and hind foot samples.

Joint samples were fixed with paraformaldehyde, decalcified with sodium citrate, embedded with paraffin and sectioned at 6 mm, followed by hematoxylin and eosin (H&E) staining. Synovial hyperplasia, inflammatory cell infiltration, and cartilage damage were assessed as previously described (Feuerherm et al., 2019). Disease severity was graded from 0 to 4 using a scoring criterion (0, normal; 1, mild: 2, moderate; 3, marked; 4, severe). In the histological images, black arrow indicated articular periosteal hyperplasia, with a large number of infiltrating inflammatory cells (lymphocytes mainly); red arrows indicated synovial connective tissue hyperplasia, with a large number of infiltrating inflammatory cells (lymphocytes mainly); yellow arrow indicated local erosion of articular cartilage; green arrow indicated trabecular bone surrounded with a large number of osteoblasts.

Pharmacokinetic experiments were performed with CIA mice. Mice were randomly divided into two groups (SS and TGT groups). Mice in the SS group were gavaged with SS extract (40 mg/kg) at ZT6 or ZT18. Plasma samples were then collected at predetermined time points (i.e., 5, 7.5, 15, 30, 60, 360, and 720 min) by means of orbital plasma collection. Mice in the TGT group were gavaged with TGT suspension (30 mg/kg) at ZT2 or ZT14. Plasma samples were collected through retro-orbital bleeding at predetermined time points (i.e., 5, 15, 30, 60, 120, 360, and 720 min). All plasma samples were processed for UPLC-QTOF/MS analysis as previously described (Zhao et al., 2020a; Gao et al., 2021). AUC (area under the curve) and other pharmacokinetic parameters were obtained by using WinNonlin software (Pharsight, Cary, NC, United States) with the non-compartmental method.

TNF-α and IL-6 in biological samples were measured by using ELISA kits (EMC102a and EMC004.96) according to the manufacturer’s protocol (Neobioscience, Shenzhen, China). CK-BB in biological samples was measured by using ELISA kit (ml026271) according to the manufacturer’s protocol (Mlbio, Shanghai, China). Creatinine, ALT and AST were determined by using for their assay kits (C011-2-1, C009-2-1, and C010-2-1) according to the manufacturer’s protocol (Jiancheng Bioengineering, Nanjing, China).

RNA was extracted from biological samples using RNAiso Plus reagent and reverse-transcribed into cDNA by PrimeScript RT Master Mix (Vazyme, Nanjing, China). SYBR Green PCR Master Mix (Vazyme, Nanjing, China) was then used for amplification reactions (an initial denaturation at 95°C for 5 min, 40 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s). Hmbs was used as an internal control. The 2^−ΔΔCT method was used to calculate relative gene expression (Lin et al., 2019a; Yu et al., 2019). The primers are listed in Table 1.

Strychnine, brucine, triptolide and their metabolites were quantified using a UPLC-QTOF/MS system (Waters, Milford, MA, United States) consisting of a Waters ACQUITY UPLC and a Xevo G2 QTOF/MS as described in our previous publications (Zhao et al., 2020b; Gao et al., 2021). Chromatographic separation was performed on an ACE UltraCore SuperC18 column (Phenomenex, Torrance, CA, United States) with a flow rate of 0.3 ml/min. The mobile phases were 0.1% formic acid (mobile phase A) and 0.1% formic acid in acetonitrile (mobile phase B). The gradient elution program was 5% B at 0–0.5 min; 5–90% B at 0.5–4 min; 90% B at 4–4.5 min; and 90–5% B at 4.5–5 min. Mass spectrometer was operated in positive ESI mode. Strychnine, brucine, pseudostrychnine, pseudobrucine, dihydroxystrychnine, triptolide, and celastrol were respectively quantified by using extracted ion chromatograms of m/z 335.17, 395.19, 350.16, 410.18, 367.16, 361.16, and 451.28 Da, with a mass window of ±0.05 Da (Supplementary Table S1). The representative extracted ion chromatograms are provided in Supplementary Figures S1, S2.

Data are presented as mean ± SD (standard deviation). Student’s t test was used to test for mean differences between two groups. One-way or two-way ANOVA followed by Bonferroni post hoc test was used for multiple group comparisons. The level of significance was set at p < 0.05 (*).

Collagen-induced arthritis (CIA) was established in mice as previously described (Figure 1A) (Inglis et al., 2008). As expected, CIA mice showed much higher production of TNF-α and IL-6 as well as higher mRNA expression levels of TNF-α, IL-6, COX-2, and iNOS as compared to normal mice (Figures 1B,C, 2A,B; Supplementary Table S2). In addition, H&E staining showed extensive infiltration of inflammatory cells in the synovial membranes in the joints of CIA mice and severe cartilage destruction (Figures 1E, 2D). Interestingly, SS dose-dependently reduced the production of TNF-α and IL-6 in CIA mice (Figure 1B). Furthermore, SS down-regulated the inflammatory factors TNF-α, IL-6, COX-2 and iNOS in a dose-dependent manner (Figure 1C; Supplementary Table S2). SS also significantly ameliorated paw swelling in CIA mice (Figure 1D). In addition, SS-treated mice showed reduced infiltration of inflammatory cells and attenuated cartilage destruction according to histopathological examinations (Figure 1E). Likewise, TGT dose-dependently reduced the production of TNF-α and IL-6, and decreased the mRNA levels of TNF-α, IL-6, COX-2 and iNOS in CIA mice (Figures 2A,B; Supplementary Table S2). In the meantime, TGT-treated mice showed reduced paw swelling, decreased inflammatory cell infiltration, and attenuated cartilage destruction (Figures 2C,D). These data clearly indicated anti-arthritis effects of both SS and TGT in mice. It was noteworthy that we did not observe changes in CK-BB and creatinine (respective biomarkers for neurotoxicity and nephrotoxicity) in SS (40 mg/kg)-treated mice or changes in plasma ALT and AST (two indexes of hepatotoxicity) in TGT (30 mg/kg)-treated mice (Figures 1F, 2E) (Zhao et al., 2020a; Gao et al., 2021).

According to the dose-response experiments (Figures 1, 2), a dose of 40 mg/kg for SS and a dose of 30 mg/kg for TGT were selected to assess dosing time-dependent efficacy against CIA in mice. CIA mice were treated with SS or TGT at each of six circadian time points (ZT2, ZT6, ZT10, ZT14, ZT18, and ZT22). The therapeutic efficacy of SS and TGT against CIA displayed significant circadian rhythms (Figures 3, 4). SS efficacy was the best when drug was dosed at ZT18, and was poorer when drug was dosed at other times (particularly at ZT6, Figure 3). To be specific, ZT18 dosing generated lower levels of inflammatory factors (TNF-α, IL-6, COX-2, and iNOS) compared to drug dosing at other times such as ZT6 (Figures 3A,B; Supplementary Table S3). Supporting this, ZT18-treated mice showed more significantly attenuated arthritis as compared to ZT6-treated mice according to histopathological examinations (Figure 3C). As for TGT, the efficacy was better when drug was dosed at ZT2, and was poorer when drug was dosed at other times, particularly at ZT14 (Figure 4). This was accompanied by lower levels of inflammatory factors (TNF-α, IL-6, COX-2, and iNOS) and a smaller histopathological score at ZT2 than at ZT14 (Figures 4A–C; Supplementary Table S3). Taken together, both SS and TGT showed dosing time-dependency of anti-arthritis effects in CIA mice.

Our previous studies have shown that the time-varying pharmacokinetics in normal mice contributes to chronotoxicity of SS and TGT at a toxic dose (Zhao et al., 2020b; Gao et al., 2021). Since pharmacokinetics can be a determinant of both drug efficacy and toxicity, it is of interest to test whether the pharmacokinetics of SS and TGT in CIA mice are circadian-dependent or not at a therapeutic dose. CIA mice were treated with SS (40 mg/kg) at ZT6 (corresponding to low drug efficacy) or ZT18 (corresponding to high drug efficacy), and with TGT (30 mg/kg) at ZT2 (corresponding to high drug efficacy) or ZT14 (corresponding to low drug efficacy) (Figures 5, 6). Pharmacokinetic analysis of SS focused on four constituents (i.e., brucine, strychnine, pseudostrychnine, and pseudobrucine) and one metabolite (dihydroxystrychnine from strychnine) that can be quantified as noted previously (Zhao et al., 2020a; Gao et al., 2021). Plasma levels of brucine and strychnine after SS dosing at ZT18 were significantly higher than those after drug dosing at ZT6, particularly in the first 30 min (Figures 5A,B). Accordingly, the AUC values of brucine and strychnine were higher at ZT18 than at ZT6 (Table 2). By contrast, pharmacokinetic behaviors of other two constituents (pseudostrychnine and pseudobrucine) showed no differences between two circadian time points (Figures 5C,D). Furthermore, a significant dosing-time effect was observed for strychnine metabolism (i.e., formation of dihydroxystrychnine) (Figure 5E). Plasma dihydroxystrychnine levels were higher at dosing time of ZT6 than at ZT18 (Figure 5E). Lower metabolite formation may account for higher exposure of strychnine at ZT18 (Figures 5B,E).

Triptolide, celastrol, tripdiolide, wilfordine, and wilfortrine are five main putative active constituents of TGT (Bao and Dai, 2011). It was hypothesized that the chronoefficacy of TGT might be associated with time-varying exposure of these compounds, which were then the focuses of pharmacokinetic analysis for TGT. We found that TGT dosing at ZT2 generated higher plasma concentrations of triptolide as compared to ZT14 dosing (Figure 6A). Consequently, the systemic exposure (reflected by AUC) of triptolide at ZT2 was higher than that at ZT14 (Table 3). In contrast, plasma concentrations of celastrol showed no significant differences between two dosing times (Figure 6B). It was noteworthy that we were unable to analyze pharmacokinetic behaviors of the other three constitutes tripdiolide, wilfordine and wilfortrine because their plasma levels were below the lower limit of quantification. Taken together, time-varying exposure of triptolide may account for dosing time-dependent efficacy of TGT.

To verify whether the herbal constituents, brucine, strychine and triptolide possess an anti-arthritis activity, we examined their effects on the proliferation of FLS and the expression of the inflammatory factors TNF-α, IL-6, COX-2, and iNOS. FLS cells were isolated from the joints of CIA mice, and treated with different concentrations of brucine, strychine or triptolide. We found that brucine, strychnine and triptolide significantly inhibited the proliferation of FLS cells with low IC₅₀ values (234.2, 215.8 μg/ml and 42.99 ng/ml for brucine, strychnine and triptolide, respectively) (Figures 7A, 8A). Furthermore, TNF-α and IL-6 production were significantly reduced in brucine-, strychine-, and triptolide-treated FLS (Figures 7B, 8B). In addition, brucine, strychine and triptolide significantly inhibited the expression of arthritis-related inflammatory factors (i.e., TNF-α, IL-6, COX-2, and iNOS) (Figures 7C, 8C). Taken together, brucine and strychine (two putative active constitutes of SS) as well as triptolide (a putative active constitute of TGT) can inhibit the proliferation of FLS and the expression of key inflammatory factors, suggesting that they do possess an anti-arthritis activity.