Rubiarbonol B induces RIPK1-dependent necroptosis via NOX1-derived ROS production

The activation of receptor-interacting protein kinase 1 (RIPK1) by death-inducing signaling complex (DISC) formation is essential for triggering the necroptotic mode of cell death under apoptosis-deficient conditions. Thus, targeting the induction of necroptosis by modulating RIPK1 activity could be an effective strategy to bypass apoptosis resistance in certain types of cancer. In this study, we screened a series of arborinane triterpenoids purified from Rubia philippinesis and identified rubiarbonol B (Ru–B) as a potent caspase-8 activator that induces DISC-mediated apoptosis in multiple types of cancer cells. However, in RIPK3-expressing human colorectal cancer (CRC) cells, the pharmacological or genetic inhibition of caspase-8 shifted the mode of cell death by Ru–B from apoptosis to necroptosis though upregulation of RIPK1 phosphorylation. Conversely, Ru–B-induced cell death was almost completely abrogated by RIPK1 deficiency. The enhanced RIPK1 phosphorylation and necroptosis triggered by Ru–B treatment occurred independently of tumor necrosis factor receptor signaling and was mediated by the production of reactive oxygen species via NADPH oxidase 1 in CRC cells. Thus, we propose Ru–B as a novel anticancer agent that activates RIPK1-dependent cell death via ROS production, and suggest its potential as a novel necroptosis-targeting compound in apoptosis-resistant CRC.


Introduction
The death domain kinase, receptor-interacting protein kinase 1 (RIPK1), is an essential mediator of the activation of programmed cell death (PCD) via apoptosis and necroptosis and exerts its effects by integrating signaling complexes following the ligation of cell surface receptors, such as tumor necrosis factor receptor 1 (TNFR1) and toll-like receptor 3 (Humphries et al. 2015;Meylan et al. 2004;Silke 2011;Witt and Vucic 2017).
Upon ligation with TNFR1, RIPK1 can transduce the signal to either cell survival or PCD, depending on the engagement of activated adapter proteins and/or the cellular context. Recently, it has been proposed that post-translational modifications of RIPK1 in spatially distinct TNFR1 complexes (complex-I and -II) play an important role in determining cell fate (Kang et al. 2019;Ting and Bertrand 2016). The conjugation of non-degradable poly-ubiquitin chains to RIPK1 bound within complex-I maintains the survival function of RIPK1 by acting as a scaffold for the recruitment of pro-survival kinases, such as IκB kinase, which are essential for the activation of nuclear factor-κB (NF-κB) (Dynek et al. 2010;Fritsch et al. 2014;Wertz 2014). Conversely, de-ubiquitination of RIPK1 by loss of ubiquitin ligase such as cellular inhibitor of apoptosis 1 and 2 (cIAP1/2) and linear ubiquitin chain assembly complex (LUBAC) dissociates from complex-I and recruits caspase-8 and its adapter protein FADD to form a cytosolic death-inducing signaling complex (DISC, also termed as complex-II) that elicits characteristic RIPK1-dependent apoptosis (RDA). (Brenner et al. 2015;Micheau and Tschopp 2003;Dickens et al. 2012;Annibaldi and Meier 2018). More recent studies have reported that the ubiquitination-dependent phosphorylation of RIPK1 by IKKα/β, transforming growth factor β activated protein kinase (TAK1) and TANKbinding kinase 1 (TBK1) in complex-I protects cells from RDA by counteracting the assembly of complex-II (Dondelinger et al. 2015;Lafont et al. 2018;Geng et al. 2017;Xu et al. 2018). During RDA, RIPK1 is rapidly cleaved by the activated caspase-8, which in turn suppress the further activation of RIPK1 (Newton et al. 2019). Consequently, genetic or pharmacological inhibition of caspase-8 activity greatly increases the cytotoxic potential of RIPK1 via Ser166-autophosphorylation, which promotes the switch to RIPK1-dependent necroptosis by inducing the recruitment of RIPK3 and mixed lineage kinase-domain-like (MLKL) (Kaiser et al. 2011;Li et al. 2012;Newton 2015).
In various human cancers, genetic alterations occur that play an important role in the evasion of apoptosis, a hallmark of cancer that represents a major mechanism of cellular resistance to current cancer treatments including radiation and chemotherapeutic drugs (Croce and Reed 2016;Hanahan and Weinberg 2000). Caspase-8 is often inactivated by somatic mutations or epigenetic methylation in multiple types of human cancer, including colorectal cancer (CRC) (Hopkins-Donaldson et al. 2003;Kim et al. 2003;Teitz et al. 2000). In caspase-8-deficient CRC, the use of Smac mimetics can reduce cellular inhibition of apoptosis protein(cIAP)-mediated RIPK1 ubiquitination, which overcomes apoptosis resistance by inducing RIPK1-dependent necroptosis (He et al. 2017). Furthermore, DNA damaging compounds such as etoposide and doxorubicin induce apoptosis or necroptosis (depending on the cellular context) without the involvement of TNFR1 ligation via the Ripoptosome, a cytosolic complex containing RIPK1, FADD and caspase-8 (Bertrand and Vandenabeele 2011;Koo et al. 2015;Tenev et al. 2011). Thus, the discovery of a substance capable of inducing the necroptotic mode of cell death via the Ripoptosome could lead to an effective chemotherapeutic strategy for eradicating apoptosisresistant cancer cells.
Triterpenoids comprise the largest group of plant natural products and possess a diverse range of pharmacological activities (Gill et al. 2016). Pentacyclic triterpenoids in particular exhibit promising antitumor activity, regulating multiple cellular pathways related to apoptosis, the cell cycle and angiogenesis (Markov et al. 2017;Patlolla and Rao 2012). Arborinane-type triterpenoids constitute a rare group of pentacyclic triterpenoids. Recently, arborinanetype triterpenoids such as rubiarbonol G and myrotheols A have attracted attention from chemists and pharmacologists due to their potential to induce apoptosis and cell cycle arrest in various cancer cell types (Basnet et al. 2019;Zeng et al. 2018). However, the activity of arborinane-type triterpenoids toward necroptotic inducers and Ripoptosome formation is largely unknown. The genus Rubia is a rich source of arborinane-type triterpenoids. In a previous phytochemical study, we isolated a series of arborinanetype triterpenoids from Rubia philippinesis (R. philippinesis) (Quan et al. 2016). In the present study, we show that a novel arborinane triterpenoid isolated from R. philippinesis, rubiarbonol B (Ru-B), elicited apoptotic and necroptotic cell death via Ripoptosome formation in RIPK3-expressing CRC cells. When apoptosis was blocked, Ru-B triggered a shift from apoptotic to RIPK1-dependent necroptotic cell death. The RIPK1-dependent cell death was mediated by NADPH oxidase 1 (NOX1)-derived reactive oxygen species (ROS) generation, which led to TNFR1independent RIPK1 phosphorylation. Our findings provide insight into the interplay between necroptotic cell death and ROS-mediated RIPK1 phosphorylation that underlies the cytotoxic potential of Ru-B and offer a potential therapeutic strategy for the treatment of refractory CRC, which is resistant to proapoptotic stimuli.

Extraction of rubiarbonol B
Arborinane-type triterpenoids, rubiarbonol B (Ru-B) was isolated from our previous chemical investigation on R. phillippinensis (Quan et al. 2016

CRISPR/Cas-9 mediated KO cells generation
For the depletion of RIPK1, RIPK3, caspase-8, and FADD in HT-29 cells, oligos were synthesized and inserted into the px330-puro vector through a standard protocol to generate gRNA with hCas9 protein. gRNA sequences were designed using the open-access software provided at http:// chopc hop. cbu. uib. no/. gRNA sequences were as follows: RIPK1-CTC GGG CGC CAT GTA GTA GA; RIPK3-CGG GCG CAA CAT AGG AAG TG; caspase-8-CAC CGA ACG AGA TAT ATC CCG GAT G; FADD-ACA CGC TCT GTC AGG TTG CG. The targeting plasmid was transfected into HT-29 cells using Lipofectamine 2000 reagent according to the manufacturer's instructions (Invitrogen Life Technologies, Franklin, MA, USA). After 24 h, cells were exposed with 3 μg/ml puromycin for two days, and clones propagated from single cells were picked out. The depletion of target genes was confirmed by both immunoblotting and genomic DNA sequencing.
Caspase-8 activation assay HCT116 cells were plated in 96-well plates and treated with a series of constituents (10 μM) derived from R. phillippinensis for 24 h. Caspase-8 activity was measured using a Caspase-Glo 8 assay kit (Promega, USA) that utilizes luminogenic caspase-8 substrates, following the manufacturer's instructions. The luminescence intensity of each sample was measured in a plate-reading luminometer (Infinite 200pro, Tecan, Switzerland).

Determination of cell death
After treatment as described in the figure legends, a cell viability assay was conducted utilizing Cell Titer-glo Luminescent Cell Viability Assay kit (Promega, USA), which measures cell viability based on ATP levels present in live cells. Luminescent measurements were taken on a microplate leader (Infinite 200pro, Tecan, Switzerland). Representative images were also taken by an inverted microscope (EVOS M5000, Thermo Fisher Scientific, USA). For the measurement of early/late apoptotic or necrotic cell death, cells were stained with 10 μM fluorescein isothiocyanate (FITC)labeled annexin V and propidium iodide (PI), in a Ca 2+ -enriched binding buffer (10 mM HEPES, pH 7.4, 140 mM NaCl, and 2.5 mM CaCl 2 ), and analyzed by two-color flow cytometry. The fluorescence of cells was analyzed by NovoCyte Flow Cytometer (ACEA Biosciences, USA).

Immunoblot analysis and immunoprecipitation
After treatment as described in the figure legends, cells were collected and lysed in M2 buffer (20 mM Tris, pH 7.6, 0.5% NP-40, 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 2 mM dithiothreitol, 0.5 mM PMSF, 20 mM β-glycerol phosphate, 1 mM sodium vanadate, and 1 µg/mL leupeptin). Cell lysates were fractionated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized by enhanced chemiluminescence (Thermo Fisher, USA). For immunoprecipitation assay, the lysates were precipitated with the relevant antibodies and protein A-or G-sepharose beads overnight at 4 °C. The beads were washed three times with M2 buffer, and the bound proteins were resolved in 10% SDS-PAGE for immunoblot analysis.

Determination of ROS production
Production of intracellular ROS was measured using a fluorescent dye dihydroethidium (DHE) in HT-29 cells. After treatment, as described in the figure legends, cells were incubated with 10 μM DHE in phosphate-buffered saline (PBS) solution containing 10% FBS for 30 min. The stained cells were analyzed with flow cytometry (NovoCyte Flow Cytometer, USA), and the mean fluorescence intensity (MFI) was calculated after correction for autofluorescence is presented. For the quantification of the mitochondrialderived superoxide (O 2 − ) production, cells were incubated with a mitochondria-target probe, Mito-SOX Red (5 μM) with a MitoTracker green (200 nM) for 10 min, and the images were captured using a fluorescent microscope (EVOS M5000, Thermo Fisher Scientific, USA).

Statistical analysis
Data are expressed as the mean ± SE from at least three separate experiments performed triplicate. Statistical analysis was carried out using one-way analysis of variance, followed by the Bonferroni t test for multi-group comparison tests. The difference between two groups was analyzed using the Student′s t test. P < 0.05 is considered statistically significant.

Ru-B induces caspase-8-mediated apoptosis
To identify novel small molecules with cytotoxic potential that act on caspase-8, we first screened a set of arborinane-type triterpenoids purified from R. phillippinensis by conducting a protease activity assay using a luminogenic substrate specific for caspase-8. In initial an investigation, Ru-B was found to be a potent caspase-8 activator, exhibiting modest cytotoxicity in HCT116 human CRC cells (Table 1, Fig. 1A).
To determine whether the cytotoxic potential of Ru-B is due to caspase-8 activation, we compared the effects of the irreversible caspase-8 and caspase-3 inhibitors z-IETD-fluoromethyl ketone (z-IETD) and z-DEVD-fluoromethyl ketone (z-DEVD) against Ru-B-induced cell death, respectively. Pretreatment Table 1 Screening for caspase-8 activation and cytotoxicity in a series of arborinane-type triterpenoids isolated from R. philippinesis with z-IETD effectively abrogated cell death in response to Ru-B treatment in multiple types of human cancer cells, including HCT116, HeLa and MCF7 cells, while z-DEVD had a marginally preventive effect against cell death (Fig. 1B, 1C). To examine the mode of cell death triggered by Ru-B, HCT116 cells treated with Ru-B were subjected to annexin V and propidium iodide (PI) staining, followed by flow cytometry. Ru-B treatment resulted in significant increases in both the early and late phases of apoptosis (52.8% and 9.14%, respectively), while very few cells were stained exclusively with PI (2.51%) (Fig. 1D). Consistently, such an increased population of cell stained with Annexin V following Ru-B treatment was significantly reduced by pretreatment with z-IETD, but not with z-DEVD. To further investigate the signaling pathway underlying Ru-Binduced apoptosis, we analyzed the sequence of activation processes in the caspase cascades. In a kinetic analysis, treatment of Ru-B resulted in sequential activation of caspase-8 and the resultant cleavage of RIPK1, Bid, caspase-3 and PARP, which was completely inhibited by pretreatment with z-IETD ( Fig. 1E). Caspase-8 is a key initiator of death receptor (DR)-mediated apoptosis upon DR ligation by associating with RIPK1 and FADD to form the DISC. However, catalytic activation of caspase-8 triggers the cleavage of DISC-associated RIPK1, resulting in destabilization of the DISC (Tummers and Green 2017). Accordingly, an immunoprecipitation assay using an anti-caspase-8 antibody showed that no evident DISC formation was observed in HCT116 cells following treatment with Ru-B only (Fig. 1F, first to third rows). However, treatment of cells with Ru-B in the presence of z-IETD led to drastic recruitment of RIPK1 and FADD to the isolated caspase-8 (Fig. 1F, fourth and fifth rows). Taken together, these results indicate that caspase-8 activation is a major element Table 1 (continued) † RLU, relative luminescence unit ‡ HCT116 cells were treated with a series of arboriane-type triterpenoids (10 μM) for 24 h, and cell death was quantified using the cell viability assay kit (Promega) and the results were expressed as mean ± SE for Ru-B-induced apoptosis via DISC formation, even though it also serves in the process of destabilization of the DISC by the cleavage of RIPK1.
Ru-B triggers a shift from apoptosis to necroptosis in RIPK3-expressing cancer cells Given that caspase-8 inhibits RIPK3-MLKL-mediated necroptosis (Shalini et al. 2015;Newton et al. 2019), it is hypothesized that RIPK3 expression in cancer cells may be play a role in determining the cell fate (apoptosis or necroptosis) in response to Ru-B treatment. To investigate this hypothesis, we compared the effects of z-IETD against Ru-B-induced cell death in pairs of CRC cells either lacking or harboring RIPK3 expression. In line with previous findings, z-IETD pretreatment drastically suppressed Ru-B-induced cell death in HCT116, DLD1 and Caco2 cells, all of which lack RIPK3 expression ( Fig. 2A). Conversely, in RIPK3-expressing cells (HT-29 and SW620 cells), Ru-B-induced cell death was significantly enhanced in the presence of z-IETD ( Fig. 2A). Important to note, pretreatment of HT-29 cells with a RIPK1-specific inhibitor necrostatin-1 (Nec-1) almost completely protected against Ru-B/z-IETD-induced cell death, which was accompanied by impaired phosphorylation of RIPK1, RIPK3 and MLKL (Fig. 2B, Fig. 2C). These and z-DEVD-fmk (10 μM) for 30 min, and then treated with Ru-B (10 μM) for the indicated times. Cell death was quantified by using Cell Titer-glo Luminescent cell viability assay as described in Materials and Methods. The data represent as mean ± S.E. of three experiments carried out in triplicate. *P < 0.05, compared with Ru-B treated group. (C-E) HCT116 cells were pretreated with z-IETD-fmk and z-DEVD-fmk for 30 min, and then treated with Ru-B for the indicated times.
(C) After 24 h, cells were visualized using an inverted phasecontrast microscope. (D) Cells were stained with FITC-labeled annexin V and PI and analyzed by flow cytometry as described in Materials and Methods. (E) Whole cell extracts were subjected to immunoblotting with the indicated antibodies. (F) HCT116 cells were pretreated with z-IETD-fmk for 30 min, and then treated with Ru-B for the indicated times. Cell extracts from each sample were subjected to immunoprecipitation (IP) with anti-caspase-8 antibody. Immunoprecipitates were analyzed by immunoblotting with the indicated antibodies. A total of 1% of the cell extract volume from each sample was used as input control results suggest that Ru-B facilitates RIPK1-dependent necroptosis in RIPK3-expressing cells under the caspase-8-blocked condition. Consistent with this notion, cell death induced by Ru-B/z-IETD was drastically abolished by RIPK3-and MLKLspecific inhibitors GSK872 and necrosulfonamide (NSA) pretreatment, respectively (Fig. 2D). Since caspase-8 activation decreases the stability of DISC by RIPK1 cleavage, Nec-1 has been shown to partially block RDA, while it completely blocks RIPK1-dependent necroptosis (Degterev et al. 2019;Xu et al. 2018;Kang et al. 2020). Consistently, in the absence of z-IETD, the extent of cell death by Ru-B was significantly but not completely inhibited by Nec-1 pretreatment (Fig. 2B), despite Nec-1 efficiently preventing Ru-B-induced caspase-8 cleavage (Fig. 2C), and thus it is believed that RIPK1 kinase activation at least partially contribute to Ru-B-induced apoptosis. Consistent with previous reports (Degterev et al. 2019;Xu et al. 2018;  Cells were treated with Ru-B (10 μM) in the absence or presence of z-IETD-fmk (20 μM) for the indicated times, and cell death was quantified as in Fig. 1B (right). The data represent as mean ± S.E. *P < 0.05, compared with Ru-B treated group. (B, C) HT-29 cells were untreated or pretreated with Nec-1 (50 μM) for 30 min and then treated with Ru-B or in combination with z-IETD-fmk for the indicated times and 24 h, respectively. (B) Cell death was quantified as in A. The data represent as mean ± S.E. *P < 0.05, compared with Ru-B treated group. # P < 0.05, compared with Ru-B/z-IETD-fmk treated group. (C) Whole cell lysates from each sample were subjected to immunoblotting with the indicated antibodies. (D) HT-29 cells were treated with Ru-B or in combination with z-IETDfmk for 24 h in the absence or presence of necroptosis inhibitors Nec-1 (50 μM), GSK-872 (3 μM), and NSA (2 μM). Cell death was quantified as in A. The data represent as mean ± S.E. *P < 0.05, compared with Ru-B treated group. # P < 0.05, compared with Ru-B/z-IETD-fmk treated group. (E) WT, CASP8and FADD-KO HT-29 cells were treated with Ru-B for the indicated times. Whole cell lysates were subjected to immunoblotting with the indicated antibodies. (F, G) WT, CASP8-and FADD-KO HT-29 cells were treated with Ru-B or in combination with z-IETD-fmk for the indicated times. Cell extracts from each sample were subjected to IP with anti-caspase-8 (F) and anti-RIPK3 (G) antibodies, respectively. Immunoprecipitates were analyzed by immunoblotting with the indicated antibodies. A total of 1% of the cell extract volume from each sample was used as input control et al. 2020), although Nec-1 completely abolished RIPK1-dependent necroptosis by TNF/SM164/z-IETD (TSZ), RDA by TNF/SM164 (TS) could only be partially protected ( Supplementary Fig. S1). To further assess the functional relationship between apoptosis and necroptosis following Ru-B treatment, we used CRISPR-Cas9 to knock out caspase-8 or FADD in HT-29 cells and examined the modes of cell death. As expected, caspase-8 signaling cascades were activated without triggering necroptosisrelated events after Ru-B treatment in wild-type (WT) HT-29 cells (Fig. 2E, left panel). By contrast, RIPK1, RIPK3 and MLKL were markedly phosphorylated in both caspase-8-and FADD-deficient HT-29 cells upon Ru-B treatment (Fig. 2E, middle and right panels). To gain further insight into the molecular mechanisms underlying Ru-B-induced RIPK1/3 and MLKL phosphorylation, we examined whether Ru-B induces the necrosome formation under the condition of pharmacological or genetic blockade of DISC-mediated apoptosis. As expected, RIPK1, RIPK3, and MLKL were associated with caspase-8 in HT-29 cells after treatment with Ru-B in the presence of z-IETD (Fig. 2F). In parallel, we observed the Ru-B-induced association of necrosome components including RIPK1, RIPK3, and MLKL in caspase-8 deficient HT-29 cells (Fig. 2G). Our results suggest that under physiological conditions where both apoptosis and necroptosis are preserved, cells preferentially undergo apoptosis in response to Ru-B treatment, but can be converted to necroptosis through necrosome formation under apoptosis-limiting condition. RIPK1 phosphorylation plays an essential role in Ru-B-induced necroptosis RIPK1 and its phosphorylation play an essential role in inducing necroptosis by forming an RIPK3-containing necrosome under apoptosis-deficient conditions (Newton 2015). To directly determine whether Ru-B-induced PCD is achieved by targeting RIPK1 or RIPK3, we examined the cytotoxic efficacy of Ru-B and Ru-B/z-IETD treatment in WT, RIPK1and RIPK3-knockout (KO) HT-29 cells. Consistent with both RIPK1 and RIPK3 not being involved in the TNF/cycloheximide (CHX)-induced cell death pathway (Kang et al. 2020;Lin et al. 2004), the cytotoxic effects of TNF/CHX in RIPK1-KO and RIPK3-KO cells was comparable to that of WT HT-29 cells ( Fig. 3A-3C). Of note, Ru-B-induced apoptotic cell death was almost completely abolished in RIPK1-KO, but not in RIPK3-KO HT-29 cells (Fig. 3B, 3C), as evident by caspase-8 cascade activation (Fig. 3D), suggesting that RIPK1 plays an essential role in Ru-B-induced apoptosis. In addition, necroptotic cell death accompanied by RIPK1, RIPK3, and MLKL phosphorylation was completely abolished in RIPK1-KO and RIPK3-KO HT-29 cells treated with Ru-B or TS in the presence of z-IETD (Fig. 3B-3D). Importantly, the phosphorylation of RIPK3 and MLKL was completely abrogated in RIPK1-KO HT-29 cells by Ru-B/z-IETD treatment (Fig. 3D, fourth to sixth rows), suggesting that RIPK1 functions as an upstream kinase responsible for RIPK3 activation to induce Ru-B/z-IETD-induced necroptosis. Similar effects were found in RIPK1-deficient Jurkat T cells (Fig. 3E), confirming that RIPK1 indeed plays a critical role in Ru-B-induced apoptotic and necroptotic cell death. Previously, RIPK3 has been reported to contribute to RDA in mouse embryonic fibroblasts through yet unknown mechanisms (Dondelinger et al. 2013). However, Ru-B-and TS-induced apoptosis observed in WT HT-29 cells was not affected by either RIPK3 deficiency or NSA pretreatment (Fig. 3B, Fig. 3C, Supplementary Fig. S1), thus excluding the possible involvement of RIPK3 and MLKL in RDA process. Consistently, we found that pretreatment of Nec-1 significantly suppressed Ru-Band TS-induced cell death in RIPK3-deficient cells including HCT116, HeLa and MCF7 cells, confirming that RIPK3 unlikely involves in Ru-B-induced RDA ( Supplementary Fig. S2). Because the phosphorylation of RIPK1 at serine residue 166 (Ser 166 ) triggers RIPK1 kinase activity to trigger the downstream cell death signaling (Kang et al. 2019), we next investigated whether Ru-B could induce RIPK1 phosphorylation. We found that in response to Ru-B, RIPK1 was phosphorylated in WT HT-29 cells, peaking at 1 h after Ru-B treatment (Fig. 4A, left panel). Important to note, Ru-Binduced RIPK1 phosphorylation was markedly prolonged and enhanced by z-IETD pretreatment, which was subsequently accompanied with the enhanced phosphorylation of RIPK3 and MLKL (Fig. 4A, right  panel). These results suggest that persistent RIPK1 phosphorylation promotes RIPK3/MLKL-mediated necroptosis when the apoptotic pathway is blocked.
Consistent with this notion, Ru-B-induced RIPK1 phosphorylation was persistent in caspase-8-KO and FADD-KO HT-29 cell, but transient in WT HT-29 cells (Fig. 4B). Moreover, Ru-B-induced RIPK1 and RIPK3 phosphorylation was almost completely inhibited by Nec-1 (Fig. 4C). Subsequent immunoprecipitation assays revealed that Ru-B treatment led to the recruitment of RIPK1 and MLKL into the isolated RIPK3 in caspase-8-KO HT-29 cells, and this necrosome formation was abrogated by Nec-1 (Fig. 4D). These results indicate that the increased RIPK1 phosphorylation triggered by Ru-B treatment likely occurs upstream of RIPK3 and actively drives necroptosis via necrosome formation under apoptosis-deficient conditions.
Previously it has been reported that, in the presence of caspase inhibitor, some anti-cancer chemicals including 5-fluorouracil and Smac mimetics induces RIPK1-dependent necroptosis via autocrine TNF-α production (Oliver Metzig et al. 2016; Gerges et al. with Ru-B for the indicated times, and whole cell extracts from each sample were subjected to IP with anti-RIPK3 antibody. Immunoprecipitates were analyzed by immunoblotting with the indicated antibodies. A total of 1% of the cell extract volume from each sample was used as input control 2016). To explore the possibility that enhanced necroptosis triggered by Ru-B/z-IETD is caused by autocrine TNF-α production, we analyzed the mRNA expression of TNF-α in HT-29 cells. As expected, treatment of SM-164 and z-IETD led to a marked increase of TNF-α expression whereas SM-164 alone had only a marginal effect. By contrast, no detectable transcriptional induction of TNF-α was observed in HT-29 cells upon Ru-B alone or Ru-B/z-IETD treatment (Supplementary Fig. S3A). Furthermore, enhanced RIPK1 phosphorylation by Ru-B/z-IETD was not affected by the cycloheximide pretreatment ( Supplementary  Fig. S3B). These results suggest that the enhanced RIPK1 phosphorylation and necroptosis by Ru-B/z-IETD does not require de novo TNF-α synthesis. Consistent with these results, we found that the degree of cell death by Ru-B/z-IETD occurred at a similar level in TNFR1-knockdown HT-29 cells when compared to control cells ( Supplementary Fig. S3C). Hence, these data confirm that Ru-B-induced RIPK1-dependent necroptosis under caspase-8 inhibited conditions is independent of TNFR1 signaling.

NOX1-derived ROS production induced by
Ru-B is required for RIPK1 phosphorylation and RIPK1-dependent cell death Previous in vitro and in vivo experimental studies reported that ROS derived from superoxide (O 2 − ) are involved in RIPK1-dependent necroptosis (Goossens et al. 1995;Kim et al. 2007;Roca and Ramakrishnan 2013). Furthermore, ROS function as a positive feedback loop to enhance necrosome formation via RIPK1 autophosphorylation at Ser 161 (Schenk and Fulda 2015;Zhang et al. 2017). Therefore, we investigated whether intracellular O 2 − levels was increased following Ru-B treatment using dihydroethidium, an oxidative fluorescent dye. Ru-B and Ru-B/z-IETD treatment caused a dramatic increase in O 2 − levels within 30 min in HT-29 cells, which peaked at 1 h after treatment (Fig. 5A). This increase was attenuated when the cells were pretreated with either antioxidants such as butylated hydroxyanisole (BHA) and apocynin or a NOX inhibitor diphenyleneiodonium (DPI); however, intracellular O 2 − levels were not reduced by pretreatment with the mitochondriatargeting antioxidant Mito-TEMPO (Fig. 5B). These results suggest that Ru-B induces non-mitochondrial ROS production, potentially via NOX. To exclude the possibility that the ROS production triggered by Ru-B treatment was mitochondrial, we used a mitochondriatargeting hydroethidine analog, MitoSOX Red, to monitor mitochondrial O 2 − production. Treatment of HT-29 cells with carbonyl cyanide chlorophenylhydrazone, a mitochondrial uncoupler, dramatically enhanced the MitoSOX Red oxidation signal, being consistent with the well-established mitochondrial uncoupling effect (Fig. 5C, bottom panel). By contrast, Ru-B treatment did not induce MitoSOX Red oxidation (Fig. 5C, middle panel), indicating that Ru-B-induced ROS production occurs independently of the mitochondria. To determine whether increased ROS production plays a role in Ru-B or Ru-B/z-IETD-induced cell death, we pretreated HT-29 cells with various antioxidants. Pretreatment with BHA or the NOX inhibitors, but not with Mito-TEMPO, significantly prevented cell death in response to Ru-B and Ru-B/z-IETD treatment; this was correlated with their ROS quenching efficiencies (Fig. 5D). Moreover, the sequential phosphorylation of RIPK1, RIPK3 and MLKL upon Ru-B/z-IETD treatment was markedly attenuated in the presence of apocynin (Fig. 5E). These results indicate that ROS generated by NOX enzymes play an important role in RIPK1 phosphorylation, which subsequently leads to RIPK1-dependent apoptosis and necroptosis in response to Ru-B and Ru-B/z-IETD, respectively.
Of the known NOX enzymes, NOX1 is expressed in several types of non-phagocytic cells, while NOX2/ gp91 is mainly found in phagocytic cells (Geiszt et al. 2003;Suh et al. 1999). Next, we investigated whether the expression of various NOX isoforms was responsible for Ru-B-induced ROS production. As shown in Fig. 5F, NOX1 and NOX2, were constitutively expressed in various types of CRCs whereas the expression of NOX4 and NOX5 was variable depending on the cell types. Notably, the mRNA levels of NOX1 were high compared to those of the other four NOX isoforms (Fig. 5F). Knockdown of NOX1 led to a significant decrease the ROS production following Ru-B treatment (Fig. 5G), suggesting that NOX1 is the major NOX responsible for Ru-B-induced ROS production in CRCs. Furthermore, knockdown of NOX1, but not NOX2, caused a marked attenuation of cell death against to both apoptotic (Ru-B) and necroptotic (Ru-B/z-IETD) triggers, respectively (Fig. 5H), which was accompanied by decreased cleavage of caspase-8 cascades and reduced phosphorylation of RIPK1 and RIPK3 (Fig. 5I). These data suggest that NOX1derived ROS production plays an essential role in RIPK1mediated cell death in CRC cells.
Three cysteine residues on RIPK1 play a crucial role in triggering necroptosis by regulating the ROS-mediated RIPK1 phosphorylation induced by Ru-B It is noteworthy that the three cysteine residues (C257, C268, and C586) in RIPK1 sense ROS signals and thus play a crucial role in RIPK1 autophosphorylation by forming oxidized disulfide bonds and causing RIPK1 to aggregate (Zhang et al. 2017). Thus, it is estimated that ROS induced by Ru-B may function as an upstream signaling component for inducing RIPK1-dependent necroptosis. To explore the underlying upstream regulatory mechanisms leading to RIPK1 phosphorylation and necroptosis by Ru-B/z-IETD, we reconstituted RIPK1 expression in RIPK1-KO HT-29 cells with WT or three cysteine mutants (3CS) RIPK1 expression vector.  Fig. 1B. The data represent as mean ± S.E. *p < 0.05, compared with Ru-B treated group. # p˂0.05, compared with the Ru-B/z-IETDtreated group. (E) HT-29 cells were pretreated with apocynin (20 μM) and then treated with Ru-B/z-IETD for the indicated times. Whole cell lysates were performed immunoblotting with the indicated antibodies. (F) Total RNA was prepared from the indicated cell lines, and RT-PCR was performed with the primers specific to human NOX isoforms. After PCR amplification, the products were analyzed by agarose gel electrophoresis and visualized using ethidium bromide staining. (G-I) HT-29 cells were transfected with either a nonspecific control siRNA or siRNA specific for NOX1 and NOX2 for 48 h. (G) Cells were treated Ru-B for 1 h, and the superoxide production was analyzed in A. (H) Cells were treated Ru-B or Ru-B/z-IETD for 24 h; cell death was quantified as in D. (I) Cells were treated Ru-B or Ru-B/z-IETD for the indicated times. Whole-cell lysates were performed immunoblotting with the indicated antibodies Consistent with a previous report (Zhang et al. 2017), no significant differences in the recruitment of ubiquitinated-RIPK1 and TRADD into TNFR1 were detected between RIPK1-KO HT-29 cells reconstituted with either WT or 3CS RIPK1 (Fig. 6A). We also observed that treating cells with TNF showed no obvious difference in NF-κB activation between these cells, as evidenced by phosphorylation of IKK and p65 or the degradation of IκBα (Fig. 6B), confirming that residues of these cysteine on RIPK1 are unlikely to be involved in upstream NF-κB activation. By contrast, following Ru-B/z-IETD treatment, RIPK1 phosphorylation (Fig. 6C), as well as the association of the RIPK1, RIPK3 and MLKL with caspase-8 (Fig. 6D), was dramatically reduced in RIPK1-KO HT-29 cells expressing 3CS-RIPK1 compared to those expressing WT-RIPK1. This suggests that the modification of the RIPK1 cysteine residues serves a critical role in Ru-B-mediated RIPK1 phosphorylation and necrosome formation. To further investigate whether cysteine residues on RIPK1 are specifically involved in the necroptosis process, the cell death induced by various stimuli was compared in RIPK1-KO cells reconstituted with WT-RIPK1 and 3CS-RIPK1. As expected, no differences were observed between RIPK1-KO HT-29 cells expressing WT or 3CS-RIPK1 following TNF-related apoptosisinducing ligand (TRAIL) treatment. (Fig. 6E). By (E-G) RIPK1-KO HT-29 cells reconstituted with the indicated RIPK1 constructs were treated with TRAIL (100 ng/ml) or the indicated combination of compounds (10 μM Ru-B; 20 μM z-IETD-fmk, 15 ng/ml TNF; 100 nM SM-164) for 24 h. (E) Cell death was quantified as in Fig. 1B. The data represent as mean ± S.E. *p < 0.05, compared with RIPK1 KO HT-29 cells expressing WT-RIPK1. (F) Cells were visualized using an inverted phase-contrast microscope. (G) Whole cell lysates from each sample were subjected to immunoblotting with the indicated antibodies contrast, necroptotic cell death triggered by either Ru-B/z-IETD or TNF/SM/z-IETD was significantly lower in RIPK1-KO cells reconstituted with 3CS-RIPK1 compared with those expressing WT-RIPK1, as evidenced by cell viability, cell morphology and the phosphorylation of RIPK1, RIPK3 and MLKL (Fig. 6E-6G). Taken together, these results suggest that the three cysteine residues on RIPK1 are essential for Ru-B to induce ROS-mediated necroptosis via amplifying RIPK1 phosphorylation.

Discussion
Given the pivotal role of RIPK1 in triggering necroptosis, small molecules capable of activating RIPK1 kinase activity and RIPK1-dependent PCD in human cancer cells present an alternative means of eradicating cancer cells, by inducing the necroptotic mode of cell death in apoptosis-resistant cancer cells. As part of our search for PCD-inducing bioactive compounds at the DISC level, the novel arborinane triterpenoid Ru-B, which was isolated from R. philippinesis, was identified as a potent inducer of dual RIPK1-dependent modes of apoptosis and necroptosis in RIPK3expressing CRC cells. In this study, we found that Ru-B markedly enhances necroptosis through upregulation of RIPK1 phosphorylation by NOX1-derived ROS production under apoptosis-limiting conditions. Thus, we propose that Ru-B is a novel RIPK1 activator that can provide an efficient strategy for inducing necroptosis to overcome CRC cells resistant apoptosis.
Caspase-8 activation is triggered by RIPK1-associated DISC formation following DR ligation, and functions as an initiator caspase that induces the extrinsic apoptotic signaling pathway (Tummers and Green 2017). The anti-cancer properties of several bioactive compounds, including pentacyclic triterpenoids, are known to be closely related to DISC-independent activation of executor caspases (e.g. caspase-3) by inducing mitochondrial dysfunction (Fulda 2010;Fulda and Kroemer 2009;Markov et al. 2017). However, in this study, we found that Ru-B induces caspase-8 activation and DISC formation without affecting the mitochondrial pathway in multiple types of CRC cells (Fig. 1). In addition, pretreatment with the caspase-8 inhibitor significantly protected Ru-B-induced apoptosis in CRC cells lacking RIPK3 expression (Fig. 1).
Moreover, we provide evidence that genetic or pharmacological inhibition of caspase-8 not only accelerates cell death by Ru-B, but also can shift the balance of cell death to necroptosis in RIPK3-expressing CRC cells (Fig. 2). Therefore, we propose that Ru-Binduced caspase-8 activation at the DISC level is the major determinant of cell death type (apoptotic or necroptotic). In this sense, the mechanism driving caspase-8 activation and DISC formation in response to Ru-B treatment is a question that remains yet largely unresolved. It has been reported that, in human cancer epithelial cells, including CRC cells, certain pentacyclic triterpenoids can activate caspase-8 by upregulating DRs such as DR5 and FAS in cell surface (Byun et al. 2018;Mou et al. 2011;Sung et al. 2014). However, we observed that the mRNA and protein expression levels of DR5 and FAS were not significantly affected by Ru-B treatment (data not shown). Thus, Ru-B-induced caspase-8 activation via DISC formation is unlikely associated with DR signaling pathway.
ROS actively participate in the execution of necroptosis, which is induced by a variety of stimuli including TNF (Vanden Berghe et al. 2010;Vanlangenakker et al. 2011;Zhang et al. 2009), FAS ligand (Chen et al. 2009) and plant-derived natural products (Sun et al. 2019;Zhao et al. 2021). However, the signaling pathways governing the crosstalk between ROS and RIPK1 activation are still under debate. For example, RIPK1 plays an essential role in TNF-induced ROS generation, which is required for the initiation of necroptosis (Kim et al. 2007;Lin et al. 2004); this suggests that the ROS production driving necroptosis occurs downstream of RIPK1. On the other hand, ROS promotes RIPK1 phosphorylation at Ser161 via its three cysteine sites, which leads to RIPK1 oligomerization and promotes RIPK1/RIPK3 interaction (Zhang et al. 2017); this suggests that ROS functions as a positive feedback loop of necrosome formation at the upstream level of RIPK1. In this study, we found that ROS accumulated after Ru-B treatment, presumably via NOX1 (Fig. 5), and both Ru-B-induced apoptosis and necroptosis were abrogated in RIPK1-deficient HT29 cells (Fig. 3). In this regard, it was interest whether ROS induced by Ru-B plays a role in controlling the cytotoxic potential of RIPK1. An important finding from this study is that Ru-B-induced RIPK1 phosphorylation was markedly prolonged and enhanced under necroptotic conditions, such as caspase-8 inhibition and FADD deficiency (Fig. 4). This indicated that the upregulation of RIPK1 phosphorylation functions as a positive effector that enables RIPK3 phosphorylation, thus facilitating necrosome formation. We also found that the ROS scavenging activity of BHA and two NOX inhibitors, diphenyleneiodonium and apocynin, correlated well with the inhibition of RIPK1 phosphorylation and Ru-B/z-IETD-induced necroptosis (Fig. 5). Thus, we propose that ROS production by NOX1 likely functions upstream of Ru-B-induced RIPK1 phosphorylation, and thus, it can switch the cell death mode from apoptosis to necroptosis in RIPK3-expressing CRC cells. In this sense, whether the NOX1-mediated RIPK1 activation by Ru-B triggers a selective cytotoxicity in cancer cells is a critical question to be further studied. It has been proposed that dysregulation of ROS production has long been implicated as a risk factor in cancer development (Liou and Storz 2010;Meitzler et al. 2014;Perillo et al. 2020). Furthermore, NOX1 is mainly expressed in the colon and has been shown to induce malignant transformation and cancer cells growth (Mitsushita et al. 2004;Suh et al. 1999). Indeed, activation of NOX1 or enhanced level of NOX1 expression has been commonly observed in several malignant cancers (Banskota et al. 2015;Juhasz et al. 2017;Laurent et al. 2008;Rudolf et al. 2018). Thus, it is possible that NOX1-mediated cytotoxic activity of Ru-B will be of clinical value for selective therapeutic approach in malignant cancer cells harboring abundant NOX1 activity, rather than non-transformed normal cells.
Nevertheless, the findings from this study also raise several questions that should be addressed. Although the results including ours showed that extramitochondrial ROS production by NOX1 is responsible for necroptosis by TNF and some anti-cancer compounds (Kim et al. 2011(Kim et al. , 2007, it has been also reported that mitochondrial involvement in this process (Schenk and Fulda 2015;Zhang et al. 2017). Although our knowledge regarding mitochondrial structure in specific cancer types is limited, it has been established that malignant transformation disturbs redox homeostasis in cancer cells (Gorrini et al. 2013). Thus, this discrepancy may depend on the cell type and/or cellular level of molecular context such as the NOX family members. Further research is needed to elucidate the dynamic interactions between Ru-B and NOX1, as well as the mechanism by which Ru-B targets NOX1 to induce ROS generation. Further in vivo studies investigating the anti-cancer efficacy of Ru-B in mice lacking caspase-8 or FADD are also necessary for the development of Ru-B as a cancer chemotherapeutic to overcome apoptosis resistance.