The experimental schemes were authorized by the Animal Care and Use Committee of Chinese PLA General Hospital (Beijing, China). Wild-type C57BL/6 mice and C57BL/6 mice with Ripk3^−/− (male, 20–22 g, 10–12 weeks old) were obtained from GemPharmatech (21). These mice had abundant food and water supply. Basic environment was settled as controlled temperature (21 ± 2°C) and regular circadian rhythm (12/12 h light/dark cycles). Mice (n = 6 per group) were intraperitoneally injected with LPS (6 mg/kg for 48 h) to establish septic cardiomyopathy according to a method described in a previous study (8). After treatment, blood samples were collected to measure the myocardial injury markers, such as creatine kinase-MB (CK-MB), troponin T, and lactate dehydrogenase (LDH) (8).

The heart performance was measured by Echocardiography in all mice 48 h after LPS injection (14.0 MHz transducer, Sequoia C512; Acuson, Siemens, Erlangen, Germany). The cardiomyocytes mechanical properties were assessed with SoftEdge Myocam system (IonOptix, Milton, MA, USA). The changes of cardiomyocyte length were monitored using SoftEdge software (14) and evaluated through the indices as follows: peak shortening (PS), time-to-PS (TPS), time-to-90% re-lengthening (TR90), maximal velocity of shortening (+dL/dt), and re-lengthening (-dL/dt).

Cardiomyocytes were separated from Ripk3^−/− and wild-type mice via the enzyme dissociation (26) and then cultured in high-glucose Dulbecco's Modified Eagle Medium (DMEM; Gibco/Thermo-Fisher Scientific, Waltham, MA, USA). Twenty percentage fetal bovine serum (HyClone Laboratories, Logan UT, USA) was supplied and external condition was settled as 37°C with an atmosphere of 5% CO2 and 95% air. Cardiomyocyte damage was induced by treating cardiomyocytes with LPS (5 μg/ml) in DMEM for 48 h (8).

The concentrations of inflammatory cytokines, including tumor-necrosis factor-alpha (TNF-α), interleukin (IL)-6, IL-10, monocyte chemoattractant protein-1 (MCP1), and cardiac injury markers, including creatine kinase, LDH, and troponin T, were determined with ELISA kits in accordance with manufacturer's protocols.

Electrophoresis with a 10% sodium-dodecyl polyacrylamide gel was used to isolate equal amounts (20–35 μg) of total protein, which were subsequently transferred onto an Immobilon-P PVDF membrane. Given the better protein separation efficiency of a gradient gel, the intention was to cut the membrane into small pieces to blot 2–3 proteins. Briefly, membranes were pre-treated with reagents from a protein stain kit for PVDF membranes (Pierce Biotechnologies, Waltham, MA, USA) and were split into several pieces according to the estimated molecular sizes with Spectra Multicolor Broad Range Protein Ladder (cat. no., 26623, Thermo-Fisher Scientific, USA) as the reference. The PVDF membranes were combined with 3% bovine serum albumin (BSA; prepared in 1 × TBST, pH 7.4, same as below) for 60 min at 25°C, and were incubated with the first antibodies in 3% BSA/TBST at 4°C overnight. Next day, the membranes were rinsed 3 times with 1 × TBST for 5 min and then incubated with species-relevant horseradish peroxidase-linked secondary antibodies (diluted 1:500–1:1,000) for 1 h at 20–25°C. The membranes were rinsed with TBST prior to incubation in a commercially available Western blotting detection reagent (ECL Prime Western Blotting Detection Reagent, Amersham cat. No., RPN2232, GE Healthcare, Madison, WI, USA) or Clarity Max Western ECL Substrate (cat. no., 1705062, Bio-Rad Laboratories, Hercules, CA, USA). Target bands were clarified using an ImageQuant LAS 400 system (GE Healthcare, USA) and band luminosity was quantified using the inbuilt ImageQuant TL software version 7.0 (GE Healthcare, USA) and standardized to GADPH.

After treatment, the samples were washed in phosphate-buffered saline (PBS) for 3 times and fixed using 4% paraformaldehyde at 25°C for 30 min. Then, the samples were incubated using Ripk3 (1:1,000 dilution, cat. no., 95702, Cell Signaling Technology, Danvers, MA, USA) and cardiac troponin T (1:1000, ab50576, Abcam, Cambridge, United Kingdom) antibodies at 4°C for 12 h. Subsequently, the samples were rinsed with PBS 3 times, followed by staining with fluorescent secondary antibody for 30 min at 37°C. DAPI was applied for nuclear staining. Fluorescence microscope (Olympus Corporation, Tokyo, Japan) was used for image taking.

qRT-PCR was performed according to a standard protocol. The primers below were used for PCR: LAMP1, forward, 5′-ACTGGTAACAACGGAACCTG−3′, reverse, 5′-ACACATTGGGGTTAGGAACA−3′; LAMP2, forward, 5′-CTAGGAGCCGTTCAGTCCAA−3′, reverse: 5′-CTTGCAGGTGAATACCCCAA−3′; GADPH, forward: 5′-TGGAGTCTACTGGCGTCTT−3′, reverse: 5′-TGTCATATTTCTCGTGGTTCA−3′. Concentrations of mRNA were measured using qRT-PCR for each RNA in triplicate, with GADPH as the endogenous control.

TFEB and YAP were knocked down in RIPK3^−/− cells using small interfering RNAs (siRNA). We purchased siRNAs and their non-targeting sequences (negative controls) from GenePharma Co., Ltd., Shanghai, China.

Changes in mitochondrial membrane potential (ΔΨm) were evaluated with the JC-1 Kit (Beyotime, China). mPTP opening was evaluated according to the rapid dissipation of tetramethylrhodamine ethyl ester (TMRE) fluorescence, in accordance with a method described in a previous study by the authors (27). The ATP generation was assessed with a commercially available ATP bioluminescence assay kit (Beyotime, China) as described previously (28).

ROS were measured using dihydroethidium (Invitrogen, San Diego, CA, USA) staining and assessed with a microscope according to our previous study (21). Cellular viability was evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (29). A commercially available caspase 3 colorimetric assay kit (cat. no., APT165, Millipore/Sigma, Burlington, MA, USA) was also used for assessment.

All variables are presented as mean ± standard error of mean (SEM). One- or two-way ANOVA and Student-Newman-Keuls post-hoc tests or t-tests were used for statistical analysis. P-value < 0.05 was considered statistically significant.

First, alterations in RIPK3 in the heart was measured via immunofluorescence after LPS treatment. As shown in Figures 1A,B, LPS treatment significantly enhanced RIPK3 expression compared to the control group. RIPK3^−/− mice were used to further investigate the role of RIPK3 in septic cardiomyopathy subsequently. LPS treatment notably enhanced the serum levels of myocardial injury markers comparing to the control group (Figures 1C–E). Interestingly, RIPK3 deletion notably decreased the concentrations of myocardial injury markers. An enhanced inflammatory response and oxidative stress have been shown to be involved in the pathophysiology of sepsis-induced cardiomyopathy. As shown in Figures 1F–K, LPS-mediated sepsis significantly increased inflammatory cytokines than the control group, including TNF-α, IL-6, IL-10, MCP1, as well as ROS production, in the mouse heart. However, compared to the LPS group, knockout of RIPK3 significantly reduced the levels of inflammatory factors and ROS production. Cardiomyocyte death is a well-established marker of myocardial injury caused by sepsis; thus, the TUNEL assay and caspase-3 activity were used to detect cardiomyocyte death. As shown in Figures 1L–N, the LPS-treated group had higher percentage of TUNEL-positive cells and caspase-3 activity in heart tissue than that in the control group. However, the loss of RIPK3 reduces LPS-induced cardiomyocyte death. Based on the above findings, sepsis-induced myocardial injury was associated with upregulated expression of RIPK3.

Cardiac function was also measured using echocardiography. Comparing with the control group, LPS treatment significantly reduced LVEF, LV fractional shortening (LVFS), and LV end-systolic pressure (LVESP), and increased LV end-diastolic pressure (LVEDP) (Figures 2A–D). Interestingly, RIPK3 deletion largely reversed LPS-induced alterations in cardiac function parameters. Subsequently, changes in cardiomyocyte contraction properties were characterized by acute isolation of cardiomyocytes from wild-type and RIPK3^−/− mice according to our previous work. As shown in Figure 2E, LPS treatment and genetic ablation of RIPK3 had no effect on cardiomyocyte length. However, compared to the control group, LPS impaired the contractile and diastolic functions of cardiomyocytes, accordance with the remarkably reduced PS, ± dL/dt, TR90, and TPS (Figures 2F–J). In contrast, LPS-induced myocardial dysfunction was shown to be decreased significantly or reversed by the loss of RIPK3. Collectively, these results confirmed that knockout of RIPK3 preserves cardiac function in patients with septic cardiomyopathy.

Previous studies have reported that mitochondrial damage is a hallmark feature of sepsis-mediated myocardial injury. First, mitochondrial membrane potential was determined using an in vitro JC-1 assay (Figures 3A,B). LPS treatment significantly reduced mitochondrial membrane potential in cardiomyocytes when comparing to the control group. Knockout of RIPK3 stabilized mitochondrial membrane potential to near-normal levels in LPS-treated cells. Furthermore, cellular ATP generation was decreased following LPS treatment, which was reversed by RIPK3 loss (Figure 3C). Compared with the control group, the mPTP opening time in LPS-treated cells was found to be significantly longer (Figure 3D). However, loss of RIPK3 significantly decreased mPTP opening time following LPS treatment. A previous study by the authors demonstrated the cardioprotective role of mitophagy in LPS-mediated mitochondrial injury and septic cardiomyopathy. A focus on mitophagy was used to explore the underlying mechanism by which RIPK3 promotes mitochondrial injury. Urolithin A (UA), mostly used as an inducer of mitophagy, was applied in cardiomyocytes under LPS challenge. Mitophagy activity of cardiomyocytes was assessed using the Mt-Keima assay. Compared to the control group, mitophagy was moderately and intensely activated by LPS and UA, separately (Figures 3C,E,F). However, RIPK3 knockout largely promoted activation of mitophagy, which was consistent with the application of UA in response to LPS. Furthermore, UA also inhibited LPS-induced mitochondrial injury as indicated by enhanced mitochondrial membrane potential, increased ATP generation, and decreased mPTP opening time, which was like the loss of RIPK3. To gain more evidence, cardiomyocyte death was also evaluated using caspase 3 activity (Figure 3G) and MTT assay (Figure 3H). According to current study, LPS induction notably reduced cardiomyocyte viability and increased cardiomyocyte death, which could be reversed by knockout of RIPK3. Thus, these results confirmed that the mitochondrial injury mediated by RIPK3 during septic cardiomyopathy could be inhibited and alleviated by the activated mitophagy.

According to previous studies, the disruption of mitophagy is associated with lysosomal dysfunction (24, 30). Given the pivotal role of TFEB in myocardial lysosomal biogenesis, we detected the expression of TFEB and mRNA of key lysosomal proteins (LAMP1 and LAMP2). As shown in Figures 4A–C, deletion of RIPK3 enabled a more extensive expression of TFEB at the protein level, and expression of LAMP1 and LAMP2 at the mRNA level than those in the LPS group, confirming the negative regulatory effect of RIPK3 on TFEB and subsequent lysosomal biogenesis. Furthermore, the upregulated TFEB expression by RIPK3 knockout was consistent with activation of mitophagy as indicated by enhanced mito-LC3B and reduced p62 expression in respond to LPS treatment (Figures 4D,E). To further explore whether TFEB is required for RIPK3 induced mitophagy inhibition, siRNA against TFEB (si-TFEB) was transfected into RIPK3^−/− cells to reduce TFEB expression. As shown in Figures 4F,G, si-TFEB inhibited mitophagy and reduced mitochondrial potential in RIPK3^−/− cells. These results confirmed that upregulated RIPK3 reduced mitophagy activation by inhibiting TFEB-mediated lysosomal biogenesis.

To explore the mechanism by which RIPK3 regulates TFEB, the YAP pathway was explored. As shown in Figure 5A, LPS stimulation reduced YAP expression compared to that in the control group. However, RIPK3 knockout increased YAP expression in response to LPS. To verify whether YAP was required for RIPK3-mediated TFEB inactivation, siRNA against YAP (siYAP) was transfected into RIPK3^−/− cells under LPS treatment. As shown in Figure 5B, blockade of the YAP pathway abolished the inhibitory effects of RIPK3 deletion on TFEB. Subsequently, it was investigated whether RIPK3 modulated mitochondrial function and cardiomyocyte viability via the YAP/TFEB pathway. The ATP generation analysis revealed that LPS-mediated ATP reduction could be reversed by RIPK3, and this effect was ineffective once the YAP pathway was blocked (Figure 5C). Furthermore, cardiomyocyte viability decreased under LPS treatment but increased to normal levels on RIPK3 deletion. The inhibition of YAP by siRNA in RIPK3^−/− cells reduced cardiomyocyte viability (Figure 5D). Collectively, these results demonstrate that RIPK3 controls TFEB-mediated cardiac protection via the YAP pathway.