The study was approved by the Research Ethics Committee of the Institute of Basic Theory of Chinese Medicine, China Academy of Chinese Medical Sciences, Beijing, China (SYXK 2016-0021, Beijing, China). All animal studies were treated in accordance with the guidelines and regulations for the use and care of animals of the Center for Laboratory Animal Care, China Academy of Chinese Medical Sciences.

Crude drugs of gypsum (60 g), Anemarrhena asphodeloides Bge. (15 g), Cinnamomum cassia Presl (10 g), Oryza sativa L. (30 g), and Glycyrrhiza uralensis Fisch (5 g) were purchased from Beijing Tongrentang Co., Ltd. (Beijing, China). BHGZD was prepared according to the original composition of the formula recorded in the Chinese Pharmacopoeia 2020 edition and obtained based on our previous study (Supplementary Material Section 1). The filtrate was combined and concentrated under reduced pressure and dried out in an oven at 70°C overnight to obtain the BHGZD powder.

Male Lewis rats (n = 51, 6∼8-week-old, 200 ± 20 g in weight) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (production license no. SCXK 2016-0006, Beijing, China). All animals were maintained under specific pathogen-free conditions in a temperature-controlled room at a constant temperature of 24 ± 1°C with a 12°h light/dark cycle and had free access to the standard rodent diet and water ad libitum.

A total of 51 rats were randomly divided into four groups: 1) normal control group (n = 15); 2) AIA-M (n = 15); 3) AIA-M-BHGZD treatment (n = 15); and 4) AIA-M-MTX treatment (positive control group, n = 6).

Induction of the AIA-M was performed as previously reported (Li W. et al., 2020; Mao et al., 2020), and AIA-M rats were established to simulate the pathological changes and characteristics of active RA after AIA-M induction. In brief, male Lewis rats were injected intradermally through the base of the tail with 10 mg/mL M tuberculosis H37 Ra (Difco, BD company, New Jersey, United States) suspended in liquid paraffin (Freund’s complete adjuvant (CFA)). From the day of primary immunization, the male Lewis rats were kept in an artificial climate box (production license no. RXZ-380A-LED) for 2 h daily with certain wind velocity (6 m/s), temperature (37°C), and humidity (90%) for a period of 15 days.

In the AIA-M-BHGZD treatment group, the dosage selection for BHGZD was 21.4 g/kg, nearly equivalent to two times of the daily dosage of RA patients in clinics, which has been proved to exert the most prominent therapeutic efficacy as in our previous study (Li W. et al., 2020). The dosage for MTX was 0.2 mg/kg. All treatments were performed for 30 days via oral administration from the day of primary immunization.

The severity of arthritis was evaluated by arthritis accidence, arthritis score, diameter of the limb, and body weight as per our previous studies (Wang et al., 2015; Zhang Y. et al., 2016; Guo et al., 2016; Li W. et al., 2020; ). Detailed information has been given in Supplementary Material Section 2. The temperature of the articular surface was detected as described in Supplementary Material Section 3.

The pain threshold was evaluated by mechanically, acetone-, and thermally induced hyperalgesia as described previously (Wang et al., 2015; Guo et al., 2016; Li W. et al., 2020; Zhang Y. et al., 2016). Detailed information on the protocol is provided in Supplementary Material Sections 4–6.

The weight ratios of the thymus, spleen, liver, and kidney relative to the total brain weight were calculated, as described in Supplementary Material Section 7.

Mouse macrophage cell lines (Raw264.7) and immortalized cell lines of synovial fibroblasts from the articular cavity in RA patients (MH7A cells) were used for experiment validation in vitro. Raw264.7 macrophage cells were maintained in DMEM (Hyclone, Logan, UT, United States), supplemented with 10% fetal bovine serum (FBS, Gibco, Carlsbad, CA, United States), 100 U/mL penicillin, 100 μg/ml streptomycin, and 2 mML glutamine (TBD, Tianjin, China) in a humidified 5% CO₂ incubator at the temperature of 37°C. The Raw264.7 macrophage cells of low passage number were used in the current experiment validations.

MH7A cells were cultured in a sterile synoviocyte growth medium (Cell Applications, Inc., San Diego, CA) added with 10% synoviocyte growth supplement (Cell Applications, Inc., San Diego, CA), 100 U/mL penicillin, 100 μg/ml streptomycin, and 2 mML glutamine (TBD, Tianjin, China) in a humidified 5% CO₂ incubator at the temperature of 37°C. MH7A cells of passage numbers four to eight were used in the current experiment validations.

To induce a conventional NLRP3 inflammasome activation, Raw264.7 macrophage or MH7A cells were induced with 0.2 μg/ml or 1 μg/ml lipopolysaccharide [LPS, Escherichia coli (O111:B4), Sigma-Aldrich, St Louis, MO, United States] for 4 or 6 h, respectively, followed by adding a 3 mM ATP (adenosine triphosphate) (ATP disodium salt hydrate, Sigma-Aldrich, St Louis, MO, United States) for 1 h and incubation at 37°C/5% CO₂ (Li et al., 2018; Zhao L. R. et al., 2018), in the presence or absence of the NLRP3 inflammasome inhibitor.

A selective NLRP3 inhibitor MCC950 was purchased from Selleck Chemicals (CP-456773 Sodium, Selleck, Houston, TX, United States). 100nM MCC950 sodium was added for 30 min before NLRP3 stimulation (Huang et al., 2017).

On the 31st day, all rats were anesthetized by intraperitoneal injection of pentobarbital sodium (50 mg/kg). The rats were secured in the supine position, and the blood was taken from the abdominal aorta using one-time anticoagulant negative pressure blood collection tubes. The blood was placed at room temperature for 20 min and centrifuged at 12,000 rpm at low temperature (4°C) for 15 min. The supernatant was collected, vortexed, and centrifuged again at high speed for 15 min. The cells were centrifuged, and supernatants were collected.

The expression levels of TLR4, IL-1β, and IL-18 proteins and activity of caspase-1 in the sera of AIA-M rats, as well as IL-1β and IL-18 in culture supernatants, were estimated using the ELISA kit (ML Bio, Shanghai, China) according to the manufacturer’s instructions. The absorbance was measured using a Multiskan™ GO microplate spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States). Detailed information on the ELISA kit used is listed in Supplementary Material Table S1.

Pyroptosis was assessed by LDH release. LDH release in the sera and culture supernatant were assayed using the LDH ELISA kit (ML Bio, Shanghai, China), according to manufacturer’s instructions. The absorbance was measured using a Multiskan™ GO microplate spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States). Detailed information on the ELISA kit used is listed in Supplementary Material Table S1.

To evaluate the regulatory activities of BHGZD on the candidate therapeutic targets in the arthritic tissue sample and Raw264.7 macrophage and MH7A cells treated with drugs, western blotting analysis was performed following the protocol of our previous studies (Li W. et al., 2020). The following antibodies were used: TLR4 (rabbit anti-TLR4 antibody, dilution 1:1,000, Bioss Antibodies, Beijing, China), NLRP3 (NLRP3 rabbit antibody, dilution 1:1,000, ABclonal Technology, Wuhan, China), ASC (ASC antibody, dilution 1:1,000, Santa Cruz Biotechnology, Santa Cruz, CA, United States), caspase-1 (caspase-1 (D7F10) rabbit antibody, dilution 1:1,000, Cell Signaling Technology, Danvers, Massachusetts, United States), GSDMD (recombinant anti-DFNA5/GSDME antibody-N-terminal, dilution 1:1,000, Abcam, Cambridge, United Kingdom), and IL-1β (IL-1beta rabbit antibody, dilution 1:1,000, ABclonal Technology, Wuhan, China). The mean normalized protein expression±S.D. was calculated from independent experiments. GAPDH (GAPDH polyclonal antibody, dilution 1:10,000, Proteintech, Chicago, United States) and β-actin (anti-beta-actin/β-actin antibody, dilution 1:2000, Boster Biological Technology, California, United States) were used as loading controls of arthritic tissue samples and cultured cells, respectively.

For NLRP3/ASC double immunofluorescence, rabbit anti-NLRP3 (dilution 1:100, ABclonal Technology, Wuhan, China) and mouse anti-ASC (dilution 1:50, Santa Cruz Biotechnology, Santa Cruz, CA, United States) primary antibodies were used. Sections were then labeled with FITC– and Cy3–conjugated secondary antibodies (dilution 1:200, Servicebio, Wuhan, China) for 2 h at room temperature, avoiding light, followed by counterstaining with 4′,6-diamidino-2-phenylindole (DAPI) staining solution (Servicebio, Wuhan, China). Subsequently, the anti-fluorescence quencher was used to seal the section (Servicebio, Wuhan, China). Fluorescence images were photographed with a Zeiss LSM 880 confocal microscope (Carl Zeiss, Jena, Germany).

Active caspase-1 was visualized with a FAM-FLICA caspase-1 assay kit using the FAM-YVAD-FMK inhibitor probe (ImmunoChemistry Technologies, Bloomington, MN, United States), according to manufacturer’s guidelines. Fluorescence images were photographed with a Zeiss LSM 880 confocal microscope (Carl Zeiss, Jena, Germany).

Cell apoptosis was determined using a TUNEL Andy Fluor™ 594 Apoptosis Detection kit (ABP Bioscience, Wuhan, China), in accordance with the manufacturer’s protocol. Fluorescence images were photographed with an inverted fluorescence microscope (MSHOT, Guangzhou, China).

After treatment, Raw264.7 macrophage cells were double-stained with Annexin V-fluorescein isothiocyanate and propidium iodide (PI) (FITC Apoptosis Detection Kit I; BD Biosciences, San Jose, CA, United States) according to the instructions. Quantification was then performed using a flow cytometer (NovoCyte 2040R; ACEA Bioscience, San Diego, CA, United States) and analyzed using NovoExpress 1.4.1 Software (ACEA Bioscience, San Diego, CA, United States).

Statistical analyses were performed using GraphPad Prism 8.0 Software (San Diego, CA, United States). Data are expressed as the mean ± S.D. and were analyzed by one-way ANOVA with Bonferroni’s or Dunnett’s post hoc test for comparison of multiple columns. Differences were considered statistically significant when the p value was less than 0.05.

To determine the pharmacological effects of BHGZD against AIA-M, we successfully established AIA-M rats simulating pathological changes and characteristics of RA with severe redness and swelling (Figure 1A). The final incidence (incidence on the 25th day after the first immunization), mean arthritis score, diameter of the limb, and articular temperature were approximately 100%, 43.56, 10.95 cm, and 33.65°C, respectively (Figure 1D∼F). BHGZD treatment strikingly improved the severity of arthritis, including the reduced arthritis incidence (p < 0.05, Figure 1D), arthritis scores (p < 0.001, Figure 1E), diameter of the limb (p < 0.05, Figure 1F), and articular temperature (p < 0.05, Figure 1G), which were consistent with the macroscopic evidence of arthritis (Figure 1A).

In terms of the response to inflammation, the thymus and spleen indexes in different groups were determined (Figure 1H∼I). BHGZD treatment decreased the thymus and spleen indexes of AIA-M rats significantly.

Simultaneously, the body weight (Figure 1C) and pain thresholds (mechanically, acetone-, and thermally induced hyperalgesia) in AIA-M rats were significantly elevated when treated with BHGZD (all p < 0.05, Figure 1J∼K). The pharmacological effects of BHGZD were similar to that of the positive drug MTX. No hepatic or renal toxicities were observed after BHGZD treatment (Supplementary Material Figure S1).

Consistent with our previous study (Li W. et al., 2020), the expression of TLR4 proteins was significantly higher in both sera and the knee joint of AIA-M rats, which was reduced by the treatment with BHGZD (all p < 0.001, Figure 2).

To address the regulation of BHGZD on NLRP3 inflammasome activation, the expression levels of the NLRP3 inflammasome pathway were detected by ELISA and Western blotting, respectively. The protein expression levels of NLRP3, ASC, and caspase-1 p20 (the active form of caspase-1) were markedly increased, which were decreased by the treatment with BHGZD (all p < 0.001, Figure 2). In addition, the activity of caspase-1 in sera was detected by ELISA, which demonstrated the reduced effect of BHGZD on the activity of caspase-1. Notably, we found that the protein expression levels of GSDMD and inflammatory cytokines (IL-1β and IL-18) were markedly increased, which were reduced by the treatment with BHGZD (p < 0.001, Figure 2). Moreover, the LDH activity was used to assess pore formation and release of intracellular soluble components. The enhancing activity of LDH in the sera of AIA-M rats was observed, which was decreased by BHGZD treatment (p < 0.05, Figure 2A).

To verify our in vivo findings based on AIA-M rats, an established method (LPS plus ATP) was applied to induce NLRP3 inflammasome activation in both Raw264.7 macrophage and MH7A cells. We initially evaluated the cytotoxicity of BHGZD on the growth of Raw264.7 macrophage cells using flow cytometry analysis, exhibiting no cell toxicity under BHGZD treatments of 7.14, 14.27, and 28.54 μg/ml, as shown in Supplementary Material Figure S2. Thus, the low, middle, and high doses of BHGZD treatment were chosen in the following assays.

The aggravated pyroptotic cell death morphology in cultured Raw264.7 macrophage and MH7A cells (membrane swell, Figure 3D) was observed, which was reversed by the treatment with BHGZD in different doses (7.14, 14.27, and 28.54 μg/ml). Then, flow cytometry analysis revealed that BHGZD treatment prominently reduced the PI–positive cell rate [a marker of cells that stains necrotic, dead, and membrane-compromised cells (Yang J. et al., 2016)], indicating the amelioration of cell membrane pore formation induced by LPS/ATP (Figure 3A∼B). In addition, the co-staining with FAM-FLICA caspase-1 and PI was performed to visualize pyroptosis. Importantly, activation of caspase-1 was observed in LPS/ATP–induced Raw264.7 macrophage cells exposed to radiation by FAM-FLICA caspase-1 staining, which was suppressed by the treatment with BHGZD (Figure 3C). Moreover, BHGZD apparently decreased TUNEL–positive cells in LPS/ATP–induced MH7A cells (Figure 3E).

To further explore the mechanism of BHGZD against pyroptosis, the expression levels of the proteins related to the NLRP3 inflammasome pathway were detected by immunofluorescence staining, ELISA, and Western blotting. The immunofluorescence staining result showed that the co-expression level of NLRP3 and ASC was increased, which was significantly reduced by the treatment with BHGZD in both Raw264.7 macrophage and MH7A cells, similar to the effects of MCC950 (Figure 4A and Figure 5A). As shown in Figure 4B∼D, BHGZD treatment decreased the expression levels of IL-1β proteins in both cultured cells and supernatants and LDH release in LPS/ATP–induced Raw264.7 macrophage cells (all p < 0.05 in 28.54 μg/ml BHGZD). Consistent with the data, BHGZD treatment decreased the expression levels of IL-1β proteins in both cultured cells and supernatants, as well as the level of IL-18 and LDH release in MH7A cell supernatants (all p < 0.05 in 14.27 and 28.54 μg/ml BHGZD treatment, Figure 5B∼D). As the maturation of IL-1β results from NLRP3 inflammasome activation, which was characterized by decreased protein expression levels of NLRP3, ASC, and caspase-1 when treated with BHGZD in LPS/ATP–induced Raw264.7 macrophage and MH7A cells (Figure 4C∼D and Figure 5C∼D).

In addition, LPS/ATP induction profoundly increased the expression of TLR4 proteins, while BHGZD treatment significantly reduced the TLR4 protein expression (all p < 0.05, Figure 4C∼D and Figure 5C∼D), revealing that BHGZD suppressed toll-like signaling activation. Moreover, Western blotting results showed that GSDMD occurred in the LPS/ATP–induced group and was decreased by the treatment with BHGZD (all p < 0.001, Figure 4C∼D and Figure 5C∼D).