RIP3 mediates TCN-induced necroptosis through activating mitochondrial metabolism and ROS production in apoptosis-resistant cancer cells

Xu Zhao Cancer research institute Jing Quan Cancer Research institute Yue Tan Hengyang medical college Ying Liu Hunan Traditional Chinese Medical college Chaoliang Liao cancer research institute Zhenzhen Li cancer research institute Weihua Liao Xiangya hospital Jikai Liu south-central university for nationalities Ya Cao Cancer Research institute Xiangjian Luo (  luocsu@hotmail.com ) Center for Cancer Research

Receptor-Interacting Protein Kinase-3 (RIP3) belongs to the RIP family and is a RIP1 binding protein with homology to RIP1 and RIP2. It plays an essential role in tumor necrosis factor (TNF) or Toll-like receptors (TLRs)-mediated cell necrosis signaling pathway [19]. The N-terminus of RIP3 contains an active kinase domain (KD), which is conserved among other RIP kinases. Unlike the N-terminus, each RIP family member encodes a unique C-terminus that determines different signaling molecules would be recruited. RIP3 contains a C-terminal motif called RIP homotype interaction motif (RHIM) [20,21], which also exists in the bridging intermediate domain (ID) of RIP1 and mediates the interaction of RIP1 and RIP3 to induce necroptosis [22]. The RHIM-RHIM interaction between RIPK1 and RIPK3 results in the formation of necrosome, which is an oligomeric cytoplasmic complex. RIPK3 is subsequently activated and further recruit and phosphorylate mixed lineage kinase domain-like (MLKL) [23][24][25][26]. Consequently, MLKL oligomers move to the membrane and destroy its integrity through non-speci c pore formation or interaction with TRPM7 calcium channel protein to promote calcium and sodium in ux, resulting in necroptotic cell death [27,28].
Mitochondria play a critical role in regulating cell energy metabolism, signal transduction and cell death [29]. Among these, the most important function of mitochondria is to generate ATP through oxidative phosphorylation (OXPHOS) to power cells [30,31]. During the process of ATP production, the electrons may leave the electron transfer chain and hit surrounding molecules to generate reactive oxygen species (ROS) eventually [32,33], which is mainly in the form of superoxide radical O 2 •− and is subsequently converted to hydrogen peroxide [34][35][36]. With the enhancement of aerobic respiration, mitochondrial ROS level increases concomitantly [37]. Apart from its critical role in necroptotic process, RIP3 is also regarded as an energy metabolism regulator. Several key enzymes of metabolic pathways, such as glycogen phosphorylase (PYGL), glutamine ligase (GLUL), glutamate dehydrogenase 1 (GLUD1) and pyruvate dehydrogenase (PDC) are known targets of RIP3 [37,38]. PYGL catalyzes the degradation of glycogen by releasing glucose-1-phosphate for utilizing glycogen as an energy source [39]. GLUL is responsible for the condensation of glutamate (Glu) and ammonia to generate glutamine (Gln) and GLUD1 converts Glu to a-ketoglutarate. Both of them are essential for the use of amino acid Glu or Gln as substrates for energy production by means of OXPHOS [40,41]. PDC decarboxylates pyruvate into acetyl-CoA, which enters into the TCA cycle as a supply for OXPHOS to promote aerobic respiration [38]. RIP3 activates these downstream targets to increase aerobic respiration and its by-product ROS, which can drive necroptotic cell death [37,38].
Natural products provide a unique source for the discovery of innovative leading compounds and drugs, which exhibits promising pharmacological effects and anticancer activities through inducing cell death and enhancing chemotherapeutic sensitivity [4,[42][43][44]. Trichothecin (TCN) is a natural secondary metabolite extracted from the endophytic fungi of Maytenus conferti orus and it belongs to the family of trichothecene mycotoxins [45]. Previous studies of our and other groups have showed that TCN exerts a broad-spectrum anti-tumor activity through cell cycle arrest, proliferation inhibition, apoptotic induction and invasion impairment with less toxic side effects [46,47]. In this study, we found for the rst time that TCN induces necroptosis in apoptosis-resistant cancer cells. Moreover, we have demonstrated that RIP3 mediates TCN-induced necroptosis through activating mitochondria energy metabolism and ROS production.

Reagents and antibodies
z-VAD-fmk, a cell-permeable and irreversible pan-caspase inhibitor and cisplatin were from MedChemExpress (Monmouth Junction, NJ, USA). TCN was obtained from Kunming Institute of Botany, the Chinese Academy of Sciences (purity >99%, HPLC analysis). Dimethyl sulphoxide (DMSO, Sigma) was used to dissolve TCN. The antibody against β-actin was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-RIP1 was from BD Biosciences (San Jose, CA, USA). Anti-RIP3 and antiphosphorylated MLKL were from Abcam (Cambridge, MA, USA). The antibody against cleaved-Caspase 3 and Ki67 was obtained from Cell Signaling Technologies (Danvers, MA, USA).

Cell viability assay
The cell proliferation rate was assessed by plating cells into 96-well plates (2 × 103 cells per well) and absorption measurement was performed at 490 nm by a BioTek microplate reader (Beckman, Brea, CA, USA) with CCK-8 kit(Beyotime, China).

Cell permeability assay
Cell permeability assay was performed using the cell-impermeable dye Sytox Green (Invitrogen, USA). Brie y, cells were incubated with 30nM Sytox Green dead cell stain for 10 min in the dark at room temperature and then imaged with a DMI3000 uorescence microscope (Leica, Germany).

Transmission electron microscopy (TEM)
TEM was performed to identify the cells undergoing necroptosis. Cells were xed with ice-cold 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) at 4°C for 4 h, then were post-xed in 1% osmium tetroxide, dehydrated in graded alcohol and embedded in epoxy resin. The sections were counterstained with uranyl acetate and lead citrate and detected with a HT7700 TEM (Hitachi, Japan) at an accelerating voltage of 100 kV.

Immuno uorescence analysis
The cells were grown on glass coverslips overnight and treated with DMSO or TCN for 24 h, followed by xation with 4% paraformaldehyde. After permeabilization with 0.3% Triton X-100 in PBS, the cells were blocked with 5% BSA for 1 h and then incubated with the primary antibody overnight. The coverslips were washed with PBS and then incubated with appropriate uorescent secondary antibody for 1 h, followed by washing and mounting using DAPI. Images were obtained by confocal microscopy (TCS SP8, Leica) and quanti ed using the NIH Image J software.
2.7 Measurement of the cellular oxygen consumption rate using the XF extracellular ux analyzer Cells were plated in XF96 cell culture plates (Seahorse Bioscience, North Billerica, MA, USA) and incubated overnight. Cells were equilibrated with Seahorse XF base medium supplemented with glucose, glutamine, pyruvate and incubated into a 37 °C non-CO2 incubator for 45 min to 1 h prior to the XF assay. Oligomycin was prepared for nal concentrations of 1.5 μM, Carbonyl cyanide-4-(tri uoromethoxy) phenylhydrazone (FCCP), rotenone and antimycin A were prepared for a nal concentration of 0.5 μM, and they were sequentially injected from the reagent ports automatically to the wells. Real-time measurements of oxygen consumption rate (OCR) in pmole per min for cells in culture medium were plotted over time. The OCR measurements were normalized to protein levels in each well.

Measurement of mitochondrial ROS accumulation
Cells were plated into 6-well plates and stained with 5 μM MitoSOX Red (Molecular Probes, Eugene, OR, USA) for 30 min at 37 °C. Mitochondrial ROS content was detected by ow cytometry (Becton Dickinson, USA) and analyzed by owjo ver.7.6 software. The uorescence intensity re ected the amount of mitochondrial ROS and an increase in uorescence intensity represents an enhanced generation of mitochondrial ROS.

Immunohistochemical analysis
The tumor tissue sections were depara nized in environmentally friendly dewaxing agent (solarbio, china) and rehydrated with an ethanol-aqueous solutions of decreasing concentrations. For antigen retrieval, tissue sections were incubated in 10 mM sodium citrate buffer (pH = 6.0) for 20 min in a microwave oven. The endogenous peroxidase activity was removed by incubating with 3% hydrogen peroxide for 10 min and was blocked in normal donkey serum for 30 min. The primary antibodies (anti-RIP3, anti-Ki67) were applied at 4 °C overnight. Chromogen was developed using DAB (Zsgbbio, China) and counterstained with hematoxylin staining kit. Immunohistochemical staining of these sections was evaluated based on all of the available tumor cells or epithelial cells meeting the typical morphological criteria by 3 pathologists using the qualitative scale that is described in the literature.

Tumor xenograft studies
Xenograft experiment using HT29 cells was performed by subcutaneous injection of 3 × 10 6

cells into 5-
week-old BALB/c nu/nu female mice. After the tumors grew to a volume of about 80-100 mm 3 , mice were randomly divided into four groups (n = 4 for each): untreated (vehicle), treatment with TCN (1 mg/kg) or DDP (1 mg/kg), treatment with TCN (0.5mg/kg) combined with DDP (1mg/kg) by intraperitoneal injection every day for 13 days, respectively. Tumor volume was calculated according to the formula (V = length x width 2 /2). At the end of experiments, the mice were euthanized by CO2 inhalation and the tumors were stripped and weighed. Animal care experimental procedures were conducted in accordance with the approval of Xiangya hospital of Central South University (Changsha, China).

Statistical analysis
Statistical analysis was performed using ANOVA and two-tailed Student's t test and a p value <0.05 was considered statistically signi cant. Statistical calculations were performed using SPSS ver.16.0 software.

TCN induces necroptosis in apoptosis-resistant cancer cells
In the presence of different concentrations of zVAD.fmk, a pan-caspase inhibitor, the protein levels of cleaved form of caspase-3 and PARP were substantially decreased in a dose-dependent manner both in HT29 and Raji cells, which indicate an effective inhibition of apoptotic pathway (Supplementary gure 1). Hence, the dosage of 20 μM zVAD.fmk was adopted to pre-treat the cancer cells to mimic a prototypic model of apoptosis-resistance due to caspase pathway inhibition. To gain insight into the potential role of Trichothecin (TCN) in apoptosis-resistant cancer cells, rstly, CCK-8 assay was performed to investigate its effect on cell proliferation. Cells were pretreated with zVAD.fmk (20 μM) for 1 h, followed by TCN treatment for 24 h. The results showed that TCN treatment dose-dependently inhibited the cell proliferation in apoptosis-resistant HT29 and Raji cells ( Figure 1A-1B). Sytox Green Nucleic Acid Stain is a high-a nity nucleic acid uorescent probe that can easily penetrate cells with compromised plasma membrane, while be unable to across the membrane of live cells. We used sytox green to stain apoptosisresistant HT29 and Raji cells upon TCN treatment and observed that green uorescence was substantially intensi ed in a dose-dependent manner ( Figure 1C). This suggests that TCN treatment signi cantly disrupts the cell membrane integrity. Accordingly, we further used TME to observe the morphology of TCN-induced cell death in apoptotic-resistant HT29 and Raji cells. Compared with the control group, TEM analysis showed necroptotic-like morphological changes in the TCN group such as mitochondrial swelling and plasma membrane rupture ( Figure 1D). Taken together, these experiments indicate that TCN treatment might induce necroptotic cell death in apoptosis-resistant cancer cells.

TCN up-regulates RIP3 upon caspase inhibition
In order to further clarify the underlying mechanism of TCN-induced necroptosis in apoptotic-resistant cancer cells, we examined both the mRNA and protein levels of RIP1 and RIP3 in HT29 and Raji cells. The results demonstrated that TCN treatment augmented the expression of RIP3 rather than RIP1 in a dosedependent manner (Figure 2A-2B). Moreover, the uorescence intensity of RIP3 was enhanced upon TCN treatment ( Figure 2C). Although the mRNA level was up-regulated, we did not observe coincident change on protein expression of RIP1 with TCN treatment (Supplementary gure 2A-2B). Similar results were con rmed by immuno uorescent assay (Supplementary gure 2C). We further tested whether TCN treatment facilitated the interaction between RIPK1 and RIPK3. Computerized quanti cation of uorescent co-localization between RIPK1 and RIPK3 showed no signi cant difference in the presence or absent of TCN (Supplementary gure 2D), which suggests that TCN treatment might have no effect on the interaction between RIP1 and RIP3 as well as the RIP1/3 necrosome formation. These ndings indicate that TCN might induce necroptosis through up-regulating RIP3, not RIP1 or the formation of RIP1/3 necrosomes. MLKL can be recruited and further be phosphorylated by RIP3, which is essential for executing necroptosis. We tested p-MLKL expression levels upon TCN treatment and observed that it was substantially augmented in a TCN dose-dependent manner both in HT29 and Raji cells ( Figure 2B). Altogether, these data indicate that TCN treatment leads to up-regulation of RIP3, consequently recruits and phosphorylates downstream MLKL to induce necroptosis in apoptosis-resistant cancer cells, while RIP1 might not be involved in this event.

Silencing of RIP3 rescues cancer cells from TCN-induced necroptosis
RIP3 has emerged as a critical regulator of necroptosis. To further investigate the role of RIPK3 in TCNinduced necroptosis, four non-overlapping short hairpin RNAs targeting RIP3 (shRIP3 3#-6#) were transfected into HT29 cells, respectively. Western blot analysis con rmed that shRIP3 3# and 6# exhibited the best knock down e ciency of RIP3 (Supplementary gure 3). CCK-8 assay was performed to examine the cell proliferation of each designated group ( Figure 3A). The results showed that TCN treatment signi cantly inhibited cell proliferation in apoptotic-resistant HT29 cells. However, knocking down of RIP3 restored cell viability suppressed by TCN treatment ( Figure 3A). It suggests that silencing of RIP3 reduces the sensitivity of HT29 cells upon TCN treatment and protects cells from TCN-induced cell death. Similar results were also observed in Raji cells ( Figure 3B). Sytox Green staining assay further con rmed that TCN-induced necroptosis was rescued by RIP3 knocking down ( Figure 3C). Moreover, TEM analysis illustrated that TCN treatment induced cell death characterized by disrupt plasma membrane, mitochondrial swelling and vacuolation. Whereas, depletion of RIP3 substantially attenuated TCNinduced morphological changes both in apoptotic-resistant HT29 and Raji cells ( Figure 3D). These ndings support the notion that RIP3 mediates the TCN-induced necroptosis in apoptotic-resistant cancer cells.
3.4 RIP3 mediates TCN-induced necroptosis through activating mitochondria energy metabolism and ROS production RIP3 acts as an energy metabolism regulator associated with switching cell death from apoptosis to necroptosis. In order to explore the downstream effector of RIP3, we screened the mRNA levels of RIP3 targets involved in energy metabolism, including PYGL, GLUL, GLUD1 and pyruvate dehydrogenase complex (PDC-E1α,E1β,E2,E3 subunits) upon TCN treatment in apotosis-resistant cancer cells. All of them are responsible for the use of energy substrates such as glucose, glutamate and glutamine, to boost aerobic respiration. Among them, PYGL and PDC-E1α mRNAs were substantially up-regulated induced by TCN both in HT29 ( Figure 4A) and Raji cells ( Figure 4B). In addition, the mRNA levels of PDC-E3 in HT29 and GLUL in Raji cells were also augmented with TCN treatment (Figure 4A-4B). Moreover, knockdown of RIP3 led to substantial reduction of the mRNA levels of both PYGL and PDC-E1α ( Figure 4C-4D). These results indicate that RIP3 mediates the up-regulation of PYGL and PDC-E1α transcriptions induced by TCN in apotosis-resistant cancer cells.
Since mitochondria play an important role in regulating cell energy metabolism and the energy substrates are metabolized to acetyl-CoA and enter into the Krebs cycle to boost OXPHOS, we further tested the change of oxygen consumption rate (OCR) in the absence or presence of TCN in apoptotic-resistant cancer cells. As expected, compared to the DMSO control, the basal respiration and maximal respiration were both enhanced upon TCN treatment ( Figure 5A). However, knockdown of RIP3 (shRIP3 3# 6#) reversed the up-regulation of OCRs induced by TCN treatment ( Figure 5A). Considering that the mitochondrial respiratory chain consists of the main intracellular source of reactive oxygen species (ROS) in most tissues, the boosted mitochondrial energy metabolism should be accompanied by enhanced ROS production. Accordingly, we further tested mitochondrial ROS levels in the presence or absence of TCN in apoptotis-resistant cancer cells. The results illustrated that mitochondrial ROS was markedly augmented upon TCN treatment. Knock down of RIP3 substantially attenuated ROS production in the mitochondria ( Figure 5B-5C). Moreover, treatment with antioxidant, N-acetylcysteine (NAC) markedly rescued cell death induced by TCN ( Figure 5D-5E). Collectively, these ndings support the conclusion that up-regulation of RIP3 to activate mitochondria energy metabolism and its by-product ROS is responsible for TCN to trigger necroptosis, especially when apoptosis is restricted in cancer cells.

TCN sensitizes cisplatin (DDP) in tumor treatment in vivo
A xenograft tumor model using HT29 cells, which is a representative of the in vitro studies, was adopted to examine whether TCN would promote chemosensitivity in tumor treatment in vivo. Administration with TCN at a dose of 1 mg/kg resulted in inhibition of tumor load by 2-fold compared to the vehicle group, while DDP treatment had no effect on tumor progression (Fig. 6A-6B). In addition, DDP treatment showed a severe side effect with obvious loss of weight relative to vehicle group (Fig. 6C). The combination of DDP with TCN (0.5mg/kg each) achieved a similar tumor suppression effect as that in TCN group. Moreover, this regimen effectively ameliorated the side effect caused by administration of DDP alone. Immunohistochemistry (IHC) staining of Ki67 showed a signi cant inhibition of tumor cell proliferation both in TCN and DDP/TCN-treated group compared with vehicle or DDP group (Fig. 6D). The upregulation of RIP3 in TCN as well as DDP/TCN-treated group (Fig. 6D) further suggested that RIP3-mediated necroptotic cell death should contribute to the bioactivity of TCN in vivo. Therefore, these results indicate that TCN may augment RIP3 expression to sensitize tumor cells to DDP chemotherapy.

Discussion
Necroptosis is typically de ned as a form of programmed necrotic cell death mediated by RIP1 and RIP3 kinases. RIPK1 and RIPK3 are regarded as central regulators, which are activated and form the necrosome complex through RHIM domain interaction to initiate necroptosis. In this paper, we have demonstrated that upon TCN treatment, the expression of RIP3 as well as the phosphorylation level of its downstream effector MLKL was substantially augmented in apoptosis-resistant cancer cells. Whereas, the change of RIP1 protein expression was not so evident. Moreover, the interaction between RIPK1 and RIPK3 can't be enhanced by TCN treatment, which suggests that TCN treatment has no effect on the RIP1/3 necrosome formation. Although RIP1 plays an important regulatory role in the process of necrosis, it is not necessary for the occurrence of necroptosis. RIP3 can be activated by several stimuli to trigger necroptosis [48,49]. lipopolysaccharide (LPS) or double-stranded RNA interacts with Toll-like receptor 3/4 via TRIF (TIR-domain-containing adapter-inducing interferon-β), which is a RHIM binding protein, to activate RIP3 directly [50]. During virus infection-induced necroptosis, a DNA-dependent activator of interferon-regulatory factors (DAI) interacts with RIP3 via RHIM to mimic the RIP1/3 necrosome complex, resulting cell death [51]. Moreover, RIP3 oligomerization can be enhanced by chemical agents to induce necroptosis in a RHIM domain-independent manner [52]. Thus, we conclude that TCN treatment might up-regulate RIP3 to induce necroptosis of apoptosis-resistant cancer cells in a RIP1-independent manner.
Cancer cells frequently display fundamentally altered energy metabolism, which is featured by increased aerobic glycolysis and reduced OXPHOS [53,54]. Here, we illustrated that TCN addition reactivated OXPHOS level to promote mitochondrial metabolism in apoptotic-resistant cancer cells. As an energy regulator, RIP3 mediates the action of TCN. PYGL and PDC are all downstream targets of RIP3, both of which are responsible for the utilization of carbon sources to boost OXPHOS. Through up-regulation of PYGL and PDC-E1α, RIP3 prompts aerobic respiration.
The enhanced respiration should be accompanied by accumulated ROS production. Growing evidence supports that ROS accumulation exerts the executioner and mediator of necroptosis [55][56][57][58]. ROS could change mitochondrial permeability, eventually resulting in necroptosis. Curcumol, a natural sesquiterpenoid hemiketal compound, induces necroptosis in hepatic stellate cells through JNK1/2-ROS signaling [57]. Mitochondrial ROS generation induced by C/EBP homologous protein promoted necroptotic cell death in human lung cancer cells [56]. In this study, we have demonstrated that concomitant ROS generation induced by OXPHOS augmentation facilitates necroptotic cell death and contributes to the anti-tumor activity of TCN. Moreover, RIP3 mediates the augmentation of ROS content induced by TCN.
Necroptosis might act as an alternative pathway to initiate cell death when apoptosis is restrained. Since RIP3 serves as a critical regulator of necroptosis, silencing RIP3 causes cancer cells to develop resistance to 5-FU [59,60], cisplatin [61][62][63], camptothecin and etoposide and the activation of key regulators in necroptotic pathway can increase the sensitivity of cells to chemotherapy drugs [5]. Here, we demonstrated that administration of TCN combination with cisplatin signi cantly increased the chemotherapeutic sensitivity of cancer cells in vivo.

Conclusion
Our study supports the view that TCN up-regulates RIP3 expression rather than RIP1 or the interaction between RIP1 and RIP3 forming necrosomes to induce necroptosis. RIP3 mediates this process through reprogramming of mitochondrial energy metabolism and ROS production. Thus, activating RIP3 to reprogram mitochondrial energy metabolism and redox homeostasis might be a novel strategy for intervention against cancer, especially in the case of apoptosis being restrained. TCN might facilitate therapeutic gain via single administration or combination with chemotherapy drugs by up-regulation of RIP3 to induce necroptotic cell death and is deserved for further exploration.  were detected by western blot assay in HT29 and Raji cells. β-actin was used as a loading control. (c) HT29 cells were seeded on coverslips overnight and pretreated with zVAD.fmk (20 μM) for 1 h, followed by TCN (0,1 μM) treatment for 24 h. The uorescence of RIPK3 was detected by confocal microscopy and nuclei were stained blue. Quanti cation of the uorescent intensity of RIPK3 imaging was shown as bar graphs. Scale bar, 25 μm. Data are shown as mean values ± S.D. of independent, triplicate experiments.