The data that support the findings of this study are available from the corresponding author upon reasonable request.

RAW264.7 macrophages were cultured in a Rosewell Park Memorial Institute (RPMI) 1640 medium supplemented with 20% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin at 37°C in a humidified atmosphere with 5% carbon dioxide (CO₂) and maintained in a logarithmic growth phase for all experiments. Then, the RAW264.7 macrophages were exposed to cadmium chloride (CdCl₂) (GHTECH 1.09633.010) at different concentrations for 24 h. The first set was 0.0, 2.0, 4.0, 6.0, 8.0, 10.0, 20.0, 30.0, 40.0 and 50.0 μmol/L which was used for the cell counting kit-8 (CCK-8) to evaluate the cell viability of RAW264 macrophages exposed to CdCl₂. The results from the CCK-8 showed that the cell viability of the RAW264.7 macrophages was more than 90% when exposed to CdCl₂ with a concentration of ≤ 10 μmol/L. Furthermore, to know the optimal dose of CdCl₂ within 10 μmol/L, the RAW264.7 macrophages were exposed to different concentrations of CdCl₂ (0.0, 1.0, 5.0, and 10.0 μmol/L), and the results showed that 1 μmol/L CdCl₂ had the most obvious pro-inflammatory effect. Therefore, CdCl₂ (1 μmol/L) was used in the following experiment. The CdCl₂ was ≥99.0% analytically pure and ≥98.0% chemically pure. To study the effect of mitochondrial homeostasis on the macrophages, we pretreated the RAW264.7 macrophages with 5 mM of N-Acetyl-L-cysteine (NAC) (Beyotime S0077) for 1 h to reduce the redox reactions or 50 μM of mitochondrial division inhibitor-1 (Mdivi-1) (ab144589) for 3 h to alleviate the mitochondria division. Additionally, bone marrow-derived macrophages (BMDMs) were isolated from the ApoE^−/−/RIPK3^−/− mice to research the effects of RIPK3 on Cd-exposed macrophages.

The animal experiments were carried out according to the National Institutes of Health Guidelines on the Use of Laboratory Animals and were approved by the Animal Ethics Committee of the Southern Medical University. With the exposure to Cd-contaminated food or water, the blood levels of Cd at any point in time are typically in the range of 10–100 nmol/L (17). As a result, 7-week-old male ApoE^−/− mice, fed with high-fat diets or chow diets (provided by the Guangdong Medical Laboratory Animal Center) for 3 months, were exposed to Cd in different concentrations (100 mg/L). In that way, the concentration of the blood Cd in the mice was close to or a little more than 10–100 nmol/L. To study the effect of mitochondrial homeostasis on the formation of AS plaque, we administered NAC orally (20 mM) to the drinking water of the mice exposed to a high-fat with Cd diet. Besides, Mdivi-1 was dissolved in dimethylsulfoxide (DMSO) and was i.p. injected (50 mg/kg body weight) per day. The RIPK3-knockout mice with a C57BL/6 background were obtained from GemPharmatech, Nanjing, China, then bred with ApoE^−/−mice to establish ApoE^−/−/RIPK3^−/− mice. Mice from each group were killed at 19 weeks for tissue digestion (n = 5–7 per group).

After being fed with a high-fat or chow diet for to 3 months, the mice were killed with deep anesthesia for blood collection. Plasma was obtained from the blood samples of the mice by centrifugation (Thermo Scientific™ Medifuge™, Massachusetts, United States) at 1,000 × g for 15 min 4°C and then stored at −80°C. The mice were subsequently fixed by perfusion through the cardiac apex with a phosphate buffer saline (PBS), and the specimens of the aortic root were obtained without the peripheral adipose tissue.

The RAW264.7 cells in the logarithmic growth phase were inoculated in a 96-well plate according to 10,000 cells/well. The next day, the macrophages were observed to grow up to 70%, then treated with different concentration gradients of CdCl₂ for 24 h. The cell viability was detected by a CCK-8 kit purchased from Dojindo (CK04), Kumamoto, Japan, in a Plate Reader (Bio Tek Instruments Epoch™, Vermont, United States).

The total RNA was extracted from the macrophages using a Trizol reagent (Invitrogen, Massachusetts, United States, 15596026). The RNA was quantified and reverse-transcribed into complementary DNAs using a PrimeScriptTM RT Master Mix kit (TaKaRa, Kusatsu, Shiga, Japan, RR036A). Finally, a quantitative real-time PCR analysis (qRT-PCR) was performed using a TB Green Premix Ex TaqTMII (TaKaRa RR820A) kit on a Light Cycler96 PCR instrument (Roche, Basel, Switzerland). The Vazyme cycling conditions were: 95°C for 30 s followed by 39 cycles at 95°C for 10 s and 60°C for 30 s. Then, a melting curve analysis was performed by increasing the temperature from 65 to 95°C for 15 min. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a loading control. The PCR primers used in this study were synthesized using TSINGKE (Beijing, China) and the sequences were: GAPDH [CATCACTGCCACCCAGAAGACTG (F), ATGCCAGTGAGCTTCCCGTTCAG (R)], IL-1β [TGGACCTTCCAGGATGAGGACA (F), GTTCATCTCGGAGCCTGTAGTG (R)], IL-6 [TACCACTTCACAAGTCGGAGGC (F), CTGCAAGTGCATCATCGTTGTTC (R)], IL-10 [CGGGAAGACAATAACTGCACCC (F), CGGTTAGCAGTATGTTGTCCAGC (R)], TGF-β [CGAAGCGGACTACTATGCTAAA (F), TCCCGAATGTCTGACGTATTG (R)], CD86 [ACGTATTGGAAGGAGATTACAGCT (F), TCTGTCAGCGTTACTATCCCGC (R)], CD206 [GTTCACCTGGAGTGATGGTTCTC (F), AGGACATGCCAGGGTCACCTTT (R)].

The proteins were extracted from the cells, tissues, or mitochondria, while the mitochondria were isolated from the RAW264.7 macrophages or aortic root using a cell mitochondria isolation kit (Beyotime C3601, Beyotime Biotechnology, Beijing, China) or tissue mitochondria isolation kit (Beyotime C3606). After the proteins were measured using BCA (Beyotime Biotechnology) to the same concentration in each group, we boiled the proteins for 10 min, and separated them using 10% sodium dodecyl sulfate-(SDS)-polyacrylamide gels then transferred onto a polyvinylidene difluoride (PVDF) membrane. After blocking the non-specific binding sites with 5% non-fat milk in Tris-buffered saline-Tween 20 (TBS-T), we incubated the membranes overnight with antibodies. The following primary antibodies were used: rabbit polyclonal antibody against GAPDH (ab9485), rabbit polyclonal antibody against NLRP3 (ab214185), rabbit monoclonal antibody against IL-1 beta[EPR16805-15] (ab234437), rabbit monoclonal antibody against TNF alpha [EPR19147] (ab183218), rabbit monoclonal antibody against Opa1[EPR11057(B)] (ab157457), rabbit monoclonal antibody against MLKL (phospho S345) [EPR9515(2)] (ab196436), and rabbit monoclonal antibody against RIP3 (phospho S232) [EPR9516(N)-25] (ab195117), all from Abcam (Cambridge, United Kingdom). The rabbit monoclonal antibody against Phospho-NF-kappaB p65 (Ser536) (93H1) (Cell Signaling, #3033S) and rabbit monoclonal antibody against COX IV (D6I4K) (Rodent Specific) (Cell Signaling, #38563) were from Cell Signaling Technology (Dancers, Massachusetts, United States). The rabbit polyclonal antibody against LC3I/II (WL01506) and rabbit polyclonal antibody against Drp1 (WL03028) were from Wanlei Biology, Shenyang, China. After having been incubated with the corresponding secondary antibodies (Boster BA1054) for 1.5 h at room temperature, the immunoblot bands were detected using enhanced chemiluminescence (Engreen 29100, Engreen Inc., San Jose, California, United States). Prestained molecular weight marker proteins (Thermo#26616) were used to calculate the molecular weights of proteins. The membranes were exposed to autoradiography film following the incubation with an enhanced chemiluminescence imaging system (Engreen 29100). The signal intensities were checked using the Gel-Pro Analyzer 4.0 software (Media Cybernetics, Silver Spring, Maryland, United States).

The levels of the inflammation markers in the cell supernatant or mouse plasma, including interleukin 1β (IL-1β) and interleukin 6 (IL-6), were examined using ELISA kits (DAKEWE #1210122 #1210602, Shenzhen, China).

The cell surfaces were stained with a phycoerythrin (PE)-conjugated anti-mouse F4/80 antibody (Biolegend 123110, California, United States), APC-conjugated anti-mouse CD206 (MMR) antibody (Biolegend 141708, California, United States), and FITC-conjugated anti-mouse CD86 antibody (Biolegend 105006, California, United States). The expression of F4/80 was determined using flow cytometry to identify the macrophages. The expression of CD86 was used to delineate the M1 macrophages, while the cells stained by the APC-conjugated anti-mouse CD206 (MMR) antibody could be identified as M2 macrophages. The cell suspensions were stained for 30 min on ice with specific antibodies and washed two times with 3 ml of PBS buffer supplemented with 0.5% bovine serum albumin (BSA). Finally, we analyzed it using flow cytometry (CytoFlex A00-1-1102, California, United States).

The cultured Raw264.7 macrophages or aortic root tissues were fixed with 4% paraformaldehyde, incubated in 1% Triton for 10 min and then in 2% BSA for 1 h, successively incubated with primary antibodies overnight at 4°C, followed by incubation with secondary antibodies. The following primary antibodies were used: F4/80 (BM8.1) rat mAb (#71299S) from Cell Signaling Technology, CD86 Polyclonal rabbit antibody (13395-1-AP), CD206 Polyclonal rabbit antibody (18704-1-AP), Tom20 Polyclonal rabbit antibody (11802-1-AP), and LC3B-Specific Polyclonal rabbit antibody (18725-1-AP) from Proteintech, Chicago, United States. Then, the samples were stained with 6-diamidino-2-phenylindole (DAPI) to visualize the nucleus and observed by a laser confocal microscope (LEICA SP8). All the images and staining intensities were acquired and measured using analysis software (Image J).

The mitochondrial superoxide of the Raw264.7 macrophages was detected using a MitoSOX™ Red mitochondrial superoxide indicator (Invitrogen™ M36008), and the mitochondrial membrane potential was determined via a mitochondrial membrane potential assay kit with JC-1 (Beyotime C2006) according to the instructions of the manufacturer.

The macrophages growing to the logarithmic phase were scraped and collected to be fixed with glutaraldehyde for 4 h. Then, sections (60 nm) were cut and stained with lead citrate and uranyl acetate at room temperature for 4 h. Next, samples were embedded in resin at room temperature for another 2 h. A Hitachi H-7500 Transmission Electron Microscope (Hitachi, Ltd., Tokyo Japan) was used to observe the samples.

After the serial 10 μm sections were cut, the optimal cutting temperature (OCT) compound-embedded tissues were stained with Oil Red O (Solarbio 08010) to evaluate the lipid content, and Re-dyed with hematoxylin (Beyotime), differentiated by hydrochloric alcohol (Beyotime), washed by double-distilled water, and finally observed under a microscope (OLYMPUS, Shinjuku City, Tokyo, Japan).

We added 500 μl 1% HNO3 to 250 μl of the whole blood of mice and added 0.01% TritonX-100 (Amresco 0694) to 5 ml. Then we measured the concentration of the blood Cd using an ICP-Mass Spectrometer (PerkinElmer NexIONTM 350X).

Data were generated through at least three independent experiments and were presented as the means ± SEM, then analyzed with the method of t-test between two groups or one-way ANOVA followed by a Bonferroni comparison test among three groups. Statistical analyses were carried out using Prism 8 (GraphPad, San Diego, California, United States). A two-tailed P < 0.05 was considered to be statistically significant.

To evaluate the cytotoxicity caused by Cd, CCK8 measurement was performed after exposing the RAW264.7 macrophages to CdCl₂. The cell viability of the RAW264.7 macrophages was more than 90% when exposed to CdCl₂ with a concentration of ≤ 10 μmol/L (Figure 1A). Subsequently, the macrophages were treated with 0.0, 1.0, 5.0, and 10.0 μmol/L CdCl₂ for the qRT-PCR detection of related inflammatory factors, and the results showed that a lower concentration of CdCl₂ (1 μmol/L) had the most obvious pro-inflammatory effect (Figure 1B). To further verify the effect of Cd on macrophage polarization in AS, 1 μmol/L Cd exposure was used for subsequent assays, including flow cytometry, western blot (WB), and ELISA after the RAW264.7 was foamed with ox-LDL (50 μg/ml) in vitro. Upon being exposed to Cd, the surface-specific molecule of the M1-type macrophages, CD86, was markedly upregulated while the surface-specific molecule of the M2-type macrophages, CD206, was significantly downregulated (Figures 1C–E). Then, the results of the WB showed that the expressions of NLRP3 and pro-IL-1β (Figures 1F,G), and pNF-KB and TNF-α (Supplementary Figures 1A,B), were increased after Cd exposure. Moreover, the inflammatory factor, IL-6, in the supernatant of macrophages was increased after the Cd exposure, measured using an ELISA kit (Supplementary Figure 1C). The proteins mentioned above were closely connected with the M1-type polarization of macrophages (18). Finally, the Oil red O staining of the aortic root revealed that the plaque area of the high-fat feeding group with Cd exposure was significantly increased compared with those without Cd exposure (Figures 1H,I). Taken together, Cd promoted macrophage polarization toward the pro-inflammatory phenotype and subsequent AS.

To further research the mechanism of Cd-mediated AS, ApoE^−/− mice fed up with a high-fat diet were exposed to Cd in vivo. The results showed that the blood Cd concentration was proportional to the Cd exposure level. The blood Cd of the group exposed to the 100 mg/L concentration of Cd is closer to the blood Cd of a human body exposed to Cd pollution (10–100 nmol/L) (17) (Figure 2A). Meanwhile, we found that the expression of RIPK3 and its downstream protein, p-MLKL, were significantly increased under Cd exposure both in vivo (Figures 2B,C) and in vitro (Figures 2D,E).

To explore the effect of RIPK3 on Cd-induced AS, we knocked out this protein to breed ApoE^−/−RIPK3^−/− mice (Figure 3A). The mice were fed with a high-fat diet and exposed to Cd. The Oil red O staining of the aortic root showed that the area of the AS plaques were significantly decreased in the ApoE^−/−/RIPK3^−/− mice compared with the ApoE^−/− mice (Figures 3B,C). The immunofluorescence staining of the aortic roots showed that the level of M1-type polarization (CD86⁺/F4/80⁺) decreased while the level of M2-type polarization (CD206⁺/F4/80⁺) increased in the ApoE^−/−/RIPK3^−/− mice compared with the ApoE^−/− mice (Figures 3D–F). The expressions of NLRP3 and pNF-KB were downregulated in the aortic root (Figures 3G,H). Additionally, the deletion of RIPK3 inhibited the polarity shift toward the M1-type BMDMs extracted from the ApoE^−/−/RIPK3^−/−, as shown by flow cytometry (Figures 3I,J). The expressions of NLRP3 and pNF-KB were inhibited in the BMDMs of the ApoE^−/−/RIPK3^−/− mice (Supplementary Figures 2A,B). The plasma level of the inflammatory cytokines, IL-1β and IL-6, secreted by the M1-type macrophages were significantly decreased in the ApoE^−/−/RIPK3^−/− mice (Supplementary Figure 1C). These results indicated that the deletion of RIPK3 attenuated Cd-induced inflammatory responses and subsequent AS.

To investigate the potential mechanism of RIPK3, we performed immunofluorescence staining on the macrophages with the mitochondrial membrane protein Tom20 antibody. The results showed that mitochondrial fragmentation was greatly reduced in the ApoE^−/−/RIPK3^−/− mice compared with the ApoE⁻/⁻ mice under Cd exposure (Figure 4A). In addition, the results from the MitoSOX™ Red mitochondrial superoxide indicator showed that the level of mitochondrial superoxides (mROS) was significantly decreased in the BMDMs from the ApoE^−/−/RIPK3^−/− mice compared with those from the ApoE⁻/⁻ mice (Figures 4B,C). Moreover, the RIPK3 knockout enhanced the mitochondrial membrane potential in Figures 4D,E. The WB results revealed that Cd exposure led to the increased expression of the mitochondrial division protein Drp1, decreased fusion protein Opa1, and decreased protective autophagy (the ratio of LC3II/I) both in BMDMs (Supplementary Figures 3C–E) and in vivo (Figures 4F–H). RIPK3 knockout restored the effects above. Furthermore, we performed an immunofluorescence staining with autophagy-related protein LC3II, and the results showed that the protective autophagosomes in BMDMs were significantly decreased under Cd exposure while RIPK3 knockout enhanced the protective autophagy (Supplementary Figures 3A,B). These results indicated that RIPK3 knockout significantly reversed the Cd-disrupted mitochondrial homeostasis.

To improve mitochondrial homeostasis, the antioxidant NAC was used to alleviate the mROS production, and Mdivi-1, a mitochondrial fission inhibitor, was used to inhibit the expression of Drp1. First, the expression of RIPK3 in the mitochondria was downregulated after the NAC or Mdivi-1 treatment (Figures 5A,B). The results showed that NAC and Mdivi-1 effectively reduced the mitochondrial fragments of the Cd-exposed macrophages (Figure 5C), decreased the production of mROS (Supplementary Figures 4A,B), and recovered the mitochondrial membrane potential (Supplementary Figures 4C,D), indicating the restoration of mitochondrial homeostasis. Additionally, results from the WB identified that NAC or Mdivi-1 decreased the expression of Drp1, and increased the expression of Opa1 and the ratio of LC3II/I (Supplementary Figures 4E–G). Meanwhile, increased autophagosomes were presented using immunofluorescence with an LC3II antibody after the NAC and Mdivi-1 treatment (Figures 5D,E). Similarly, the results from the TEM showed that the treatment with NAC and Mdivi-1 significantly increased the content of autophagosomes in macrophages, while the production of mitophagy was more obvious in the Mdivi-1-intervened group (Supplementary Figure 5A). The decreased expression of CD86 and the increased expression of CD206 were both demonstrated using flow cytometry with the recovery of the mitochondrial homeostasis (Figure 1E). Moreover, NAC and Mdivi-1 markedly decreased the expression of NLRP3, pNF-KB, and their downstream proteins, pro-IL-1β and TNF-α (Figures 5F,G). These results indicated that improving mitochondrial homeostasis with NAC or Mdivi-1 effectively reversed Cd-induced macrophage polarization toward the pro-inflammatory phenotype via RIPK3 inhibition.

First, the expression of RIPK3 and its downstream protein, p-MLKL, were significantly downregulated after the treatment with NAC or Mdivi-1 (Figures 6A–D). The Oil red O staining of the aortic root revealed that NAC and Mdivi-1 effectively reduced the plaque area in the mice with high-fat and Cd exposure (Figures 1H,I). Additionally, the mice treated with NAC or Mdivi-1 showed a significant increase in the expression of Opa1 and the ratio of LC3II/I but a remarked decrease in the expression of Drp1 (Figures 6E–G, Supplementary Figures 6A–C). The treatment with NAC or Mdivi-1 significantly decreased the M1-type polarization (the ratio of CD86⁺/F4/80⁺; Figures 6H,I), while increasing the M2-type polarization (Supplementary Figures 6D,E). Moreover, the expression of NLRP3 and pNF-KB decreased after the treatment with NAC or Mdivi-1 (Figures 6J,K, Supplementary Figures 6F,G). The results from ELISA further confirmed that the levels of the M1-type inflammatory markers, il-1β and il-6, were significantly decreased after the treatment with NAC or Mdivi-1 (Supplementary Figures 6H,I). Taken together, maintaining the mitochondrial homeostasis with NAC or Mdivi-1 could reverse Cd-induced macrophage polarization to the pro-inflammatory phenotype and thus counteract AS in vivo.