The role of necroptosis in cancer biology and therapy

The downregulation of the expression of numerous key molecules in necroptotic signaling pathways has been found in different types of cancer cells, suggesting that cancer cells may evade necroptosis to survive (Table 2).

RIPK3 expression is absent or decreased in numerous cancer cell lines [65, 69]; specifically, the loss of RIPK3 protein expression was found in two-thirds of the 60+ cancer cell lines tested [65]. A decreased RIPK3 expression has also been reported in samples from human patients with cancer, such as breast cancer [7, 65], colorectal cancer [12, 66], acute myeloid leukemia (AML) [11, 68] and melanoma [69]. Moreover, Hockendorf et al. [11] reported that leukemogenesis was markedly accelerated following the knockout of RIPK3 in mice transplanted with bone marrow cells bearing a mutated AML driver gene and that the survival of the RIPK3 knockout mice was poorer than that of the wild-type mice. In addition, the tumor-suppressing effects of RIPK3 have been documented in colorectal cancer. In a cohort study involving more than one hundred patients, low RIPK3 expression was found to independently prognosticate a reduced DFS (disease-free survival) and OS (overall survival) [12]. Furthermore, RIPK3 knockout mice were reportedly at a higher risk of developing colitis-associated colorectal cancer and producing a higher number of pro-inflammatory or tumor-promoting factors [77]. Similarly, low RIPK3 expression indicates a worse prognosis in patients with breast cancer [65]. These studies suggest that RIPK3 might play an anti-inflammatory and antitumor role in cancer. Furthermore, reports indicate that genomic methylation and/or hypoxia may play a pivotal role in silencing RIPK3 expression in many cancer cell lines [26, 65, 66].

Consistent with these observations, McCormick et al. [9] reported that RIPK1 expression is downregulated in head and neck squamous cell carcinoma as well; this downregulation was proven to be correlated with disease progression. The authors suggested that the downregulation of RIPK1 expression promoted by epigenetic changes during tumor progression enables tumor cells to evade anoikis, which may stimulate tumorigenesis by enhancing the metastatic abilities of the tumor cells [9]. In addition, the expression of CYLD, which is a deubiquitinating enzyme that is a key mediator in the necroptotic pathway, was found to be decreased in chronic lymphocytic leukemia (CLL) [71] and malignant melanoma [70]. In CLL, low CYLD expression identifies a subgroup of patients with worse OS [71]. In melanoma, the repression of CYLD by the transcription factor Snail1 contributes to cell proliferation and cancer invasiveness in vitro and tumor progression and metastasis in vivo [78]. However, because CYLD is also involved in the NF-kB pathway, which mediates inflammation and tumor growth [79], the effects of CYLD downregulation may not only be associated with necroptosis.

Collectively, these studies implicate the antitumor role of the necroptosis pathway in cancer. However, the downregulation of necroptotic factors does not appear to occur in all cancers; the expression of necroptotic factors has been found to be upregulated in some cancers. For example, in glioblastoma, RIPK1 is commonly overexpressed, and the upregulation of RIPK1 expression is correlated with a poorer prognosis [10]. Similarly, RIPK1 expression is markedly elevated in both human lung cancer samples and mouse lung tumor models, and RIPK1 has been suggested to play an oncogenic role [72]. Notably, the expression of RIPK1, RIPK3, FADD and MLKL is elevated in pancreatic ductal adenocarcinoma (PDA) [64], which is accompanied by accelerated oncogenesis.

Interestingly and counterintuitively, according to Colbert et al. [73] the decreased expression of MLKL was correlated with a decreased OS in patients with early- stage resected pancreatic adenocarcinoma. Moreover, the reduced level of MLKL was markedly correlated with the reduced OS in gastric cancer [74], ovarian carcinoma [75], cervical squamous cell carcinoma [76] and colon cancers [67], probably because MLKL can affect the modulation of local tumor microenvironment immunosurveillance. These findings suggest that MLKL is a candidate prognostic biomarker in those cancers.

The immunosurveillance of cancer refers to the process by which the immune system identifies and eliminates cancerous and/or precancerous cells based on tumor-specific antigens (TSAs) or tumor-associated antigens (TAAs) [80] before these cells constitute a threat to our health [81]. This process is mediated by innate and adaptive immune cells and effector molecules, including dendritic cells (DC), cytotoxic T cells, M1 macrophages, natural killer (NK) cells, natural killer T (NKT) cells and their corresponding cytokines [81, 82]. RIPK3 has been found to be required in the regulation of cytokine expression in DCs, which are crucial sentinels that regulate immune homeostasis by expressing modulatory cytokines and interlinking the innate and adaptive immune systems [83]. Additionally, despite evidence indicating that RIPK3 signaling may not play a role in the regulation of the activation of T lymphocytes, B lymphocytes and macrophages [84], RIPK3 has been suggested to regulate NKT cell function and promote the NKT cell- mediated anti-tumor immune response by activating mitochondrial phosphatase phosphoglycerate mutase 5 (PGAM5) through a process that is independent of the necroptosis pathway [85]. Additionally, while the function of apoptosis in the maintenance of central tolerance has been well defined, several reports have indicated that necroptosis plays a regulatory role in the antigen-induced proliferation of T cells mainly via the elimination of excessive T cells, which is essential for maintaining homeostasis in peripheral T cells and the survival of T cells when they are activated by stimuli, and necroptosis-dependent process is negatively regulated by caspase-8 [86]. Necroptosis reportedly occurs during the late stage of T-cell proliferation and necroptotic signaling is markedly intensified in T cells absent in FADD, suggesting that FADD may negatively regulate necroptosis mediated by T cell receptors [87].

In addition to the direct interactions with immune cells, necroptosis initiates adaptive immune responses by releasing DAMPs into the tissue microenvironment [37], and after necroptotic cells are phagocytosed, phagocytic cells, such as DCs and macrophages, can release pro- inflammatory cytokines that increase stimulating molecules and amplify cross-presentation, incurring strong immune responses [36, 88]. Necroptotic cells can provide both antigens and inflammatory cytokines to DCs for antigen cross-priming which activates cytotoxic CD8+ T lymphocytes [89‐91]. Yatim et al. demonstrated that RIPK1 expression and NF-κB activation during programmed cell death are essential for initiating CD8+ T cell adaptive immunity and that CD8+ T cells activated by immune responses that necroptotic cells incurred released various effector cytokines, demonstrating in vivo cytolytic effects and defended mice against tumorigenesis [89]. Werthmöller et al. reported that the use of a combination of the pan-caspase inhibitor zVAD-fmk, which has been demonstrated to induce necroptosis, and other therapeutics, including radiotherapy, chemotherapy and hyperthermia, for the treatment of melanoma remarkably reduced tumor growth by reducing the tumor infiltration of regulatory T cells (Tregs) and increasing DC and CD8+ T-cell infiltration in the tumor microenvironment [24]. Schmidt et al. [92] showed that necroptotic cervical cancer cells induced by PolyI:C, which is a viral dsRNA analog that triggers necroptosis in cervical cancer cells, produced interleukin-1α (IL-1α), which is essential for the activation of DCs to release IL-12, which is a cytokine pivotal for antitumor effects, and that the expression level of RIPK3 in cervical carcinoma cells may predict the efficacy of PolyI:C-induced immunotherapy; therefore, immunotherapeutic treatment should be customized according to the RIPK3 level.

Despite the role of necroptosis in the induction and amplification of cancer immunity, multiple lines of evidence indicate that the immune inflammatory cells recruited by necrosis/necroptosis can promote tumor development by fostering angiogenesis, promoting cancer cell proliferation, and accelerating cancer metastasis [1, 93]. Additionally, necrotic/necroptotic cells can release regulatory cytokines, such as IL-1α, which can directly stimulate the proliferation of neighboring cells and potentially facilitate neoplastic progression [1, 93]. Activated inflammatory cells may also release reactive nitrogen intermediates (RNI) and ROS that can damage DNA and lead to genomic instability, thereby facilitating tumorigenesis [93] (Fig. 2).

Through the release of DAMPs into the tissue microenvironment, necroptotic tumor cells may provide both antigens and inflammatory cytokines to DCs for antigen cross-priming which activates cytotoxic CD8+ T lymphocytes, resulting in tumor cell elimination. However, DAMPs released by necroptotic cells may also recruit immune inflammatory cells and incur inflammation, which can promote tumor development by fostering angiogenesis, promoting cancer cell proliferation, and accelerating cancer metastasis. Activated inflammatory cells may also release reactive nitrogen intermediates (RNI) and ROS that can damage DNA and lead to genomic instability, thereby facilitating tumorigenesis. Moreover, RIPK1 expression and NF-κB activation during programmed cell death are essential for initiating CD8+ T cell adaptive immunity, and RIPK3 has been suggested to regulate NKT cell function and promote the NKT cell-mediated anti-tumor immune response by activating mitochondrial phosphatase phosphoglycerate mutase 5 (PGAM5).

Necroptosis has also been shown to generate an immunosuppressive tumor microenvironment in vivo in PDA and, thus, promote the oncogenesis of pancreatic cancer [64]. In RIPK3 knockout p48^Cre;Kras^G12D pancreases, the percentages of B cells and T cells were elevated, the percentages of peritumoral myeloid-derived suppressor cells (MDSCs) and tumor-associated macrophages (TAMs), both of which can not only inhibit antitumor immune reactions but also stimulate tumor growth and metastasis [94, 95], were reduced and the expression of programmed death-ligand 1(PD-L1), a ligand which negatively regulates T cell antigen receptor signaling through interacting with its receptor PD-1 [96], in macrophages was decreased. Furthermore, the blockade of necroptosis may expand and activate T cells, which is a promising avenue for ameliorating the unsatisfactory efficacy of checkpoint-based immunotherapeutics in pancreatic cancer [64].

As mentioned above, necroptosis has been shown to perform antitumor functions in cancer; however, mounting evidence suggests that as a pathway triggering inflammatory responses, necroptosis may also play a tumor- promoting role, suggesting that the necroptosis pathway is a double-edged sword in cancer.

For instance, a study conducted by Liu et al. [17] demonstrated that in several breast cancer cell lines, the knockout of the RIPK1, RIPK3, or MLKL genes in cancer cells markedly reduced their tumorigenicity and appeared to sensitize breast cancer cells to radiotherapy. Moreover, in a xenograft model, the necroptosis inhibitor NSA (necrosulfonamide) greatly delayed tumor growth [17]. Additionally, the authors also reported that the higher phosphorylation levels of MLKL were correlated with a poorer prognosis and shorter survival in human patients with colon and esophageal cancer, indicating that necroptotic genes play a critical role in tumor promotion [17]. Furthermore, Seifert et al. [64] reported that the in vivo deletion of RIPK3 or RIPK1 attenuated tumor progression and immunosuppression in mice and explained these results by suggesting that necroptosis promoted pancreatic oncogenesis because CXCL1 (chemokine (C-X-C motif) ligand 1), which is a chemokine attractant, and Mincle signaling induced by necroptosis promotes the induction of adaptive immunosuppression by myeloid cells. Collectively, these studies indicate that the necroptosis pathway may increase the risk of tumor progression. The mechanisms underlying this seemingly paradoxical phenomenon may be related to the inflammatory response triggered by necroptosis, which may provide a tumor- promoting inflammatory microenvironment or elevated reactive oxygen species (ROS) production, which is correlated with genomic instability [97], ultimately accelerating malignant transformation and cancer progression [98, 99].

Metastasis is the primary cause of resultant mortality in cancer patients and involves the dissemination of cancer cells from the primary site to distant organs through the circulatory system.

The role of necroptosis in metastasis also exhibits duality. Fu et al. reported that in an in vivo osteosarcoma model, not only primary tumors but also lung metastases were markedly reduced by shikonin, which is a component used in Chinese herbal medicine, probably by inducing RIPK1- and RIPK3-dependent necroptosis [100]. One possible mechanism underlying the antimetastatic role of necroptosis may be its function in the regulation of ROS production, which involves necroptosis in extracellular matrix (ECM) detachment and metabolism and, ultimately, in cancer metastasis [101]. Indeed, RIPK3 has been demonstrated to regulate the production of downstream ROS [101, 102] and could activate multiple metabolic enzymes to modulate TNF-induced ROS production [51] during the process of necroptosis, which jointly induces a considerable production of ROS, thus enabling necroptosis to kill metastatic cancer cells by incurring ROS bursts. Accordingly, necroptosis may be a critical pathway inhibiting tumor metastasis.

However, contrasting evidence indicates that under certain circumstances, necroptosis may promote cancer cell metastasis. Extravasation, which is the process of tumor cell exit from the blood vessels and entry into a secondary site, is a crucial step in metastasis. Strilic et al. reported that tumor cells can induce necroptotic endothelial cell death to promote tumor cell extravasation and cancer metastasis via the activation of DR6 (death receptor 6) [63]. That study demonstrated that when cocultured with tumor cells, endothelial cells undergo necroptotic cell death. Similarly, after treatment with metastatic tumor cells, murine lung epithelial cells demonstrated necroptotic features. Moreover, the binding of DR6 to its ligand APP (amyloid precursor protein) promoted endothelial cell death and cancer cell extravasation. Strilic et al. explained that endothelial cells subjected to necroptotic death provide a tunnel through which tumor cells can pass and start to extravasate and/or the DAMP (damage-associated molecular pattern) molecules generated by necroptotic cells exert effects on tumor cells and adjacent endothelial cells, thus promoting the extravasation and metastasis of cancer cells. Thus, the authors suggested that therapies targeting DR6-mediated endothelial cell necroptosis may represent a new approach for preventing cancer metastasis.

In summary, the net effect of necroptosis on oncogenesis and cancer metastasis remains undefined because the specific role of necroptosis may not be universal and may vary according to the different biological traits or tumor microenvironments of each cancer type. Whether necroptosis facilitates or suppresses tumor growth and metastasis cannot be conclusively determined.

Recently, Seehawer et al. coincidently found that of the two major subtypes of liver cancer, i.e., hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC), the type that animal models develop is determined by the gene-delivery technique, i.e., HCC via hydrodynamic tail-vein injection (HDTV) and ICC via in vivo electroporation, and that the cancer-promoting genes that they transferred were the same [6].

These confounding results were deciphered by the following findings: HDTV triggered apoptosis in the microenvironment, whereas electroporation triggered necroptosis, and the electroporated livers showed higher levels of phosphorylated MLKL and elevated mRNA expression of RIPK3 [6], which are biomarkers of necroptosis. The authors explained that necroptotic cells may release DAMPs that can shape the microenvironment via cytokines released by immune cells with pattern recognition receptors (PRRs) and that the necroptotic microenvironment may direct the lineage commitment of liver cancer, causing the switch from HCC to ICC development; this process is independent of the oncogenic drivers but may be involved in the epigenetic regulation of the genes Tbx3 and Prdm5 [6]. Additionally, the pharmacological or genetic inhibition of necroptosis reverts the necroptosis-dominated microenvironment and converts ICC to HCC [6], which further substantiates the role of necroptosis in determining liver cancer subtypes. The study provides a revolutionary insight into how tumor microenvironment may be shaped by a specific cell death modality and may eventually regulate lineage commitment in liver tumorigenesis and thus determine cancer subtypes, and further investigations are warranted to explore the role of necroptosis on other types of cancer and the fundamental mechanisms behind.

Inducing and/or manipulating necroptosis in anti-cancer therapies represent a promising therapeutic approach for bypassing acquired or intrinsic apoptosis-resistance, serving as an alternative way to eliminate apoptosis- resistant cancer cells. A growing arsenal of compounds and multiple chemotherapeutic agents have been reported to trigger necroptosis in cancer cells (Table 3).

Oncolytic viruses (OVs) are novel anticancer agents whose antitumor activities are attributed to oncolysis and induced antitumor immunity [147]. OVs induce mostly immunogenic cancer cell death (ICD), which includes necroptosis, through exposure to calreticulin and the release of ATP, high-mobility group box 1 (HMGB1), DAMP, and PAMP, which may activate dendritic cells and incur adaptive antitumor immunity [124, 147]. These viruses also encode certain genes to regulate ICDs, including necroptosis, which according to the authors, can be genetically modified to elicit a certain desired modality of ICD in the cancer cells infected by the viruses [124]. For instance, in ovarian cancer cells, vaccinia virus can cause necroptosis [125]. In glioma, ICD induced by newcastle disease virus occurred in a caspase-independent fashion and was blocked by nec-1, implicating the contribution of necroptosis [126].

Hemagglutinating virus of Japan-envelope (HVJ-E) also triggers necroptosis in xenograft model of human neuroblastoma cells, which is triggered by elevated level of cytoplasmic calcium activating calcium-calmodulin kinase II (CaMK II) [127].

Smac mimetics are small molecule mimics of second mitochondrial activator of caspases (Smac), which is an endogenous protein that promotes apoptosis by inhibiting cIAPs [19]. Smac mimetics have been recognized as emerging anticancer agents [150]. Similar to shikonin, a Smac mimetic was reported to permit cells to bypass apoptosis resistance via the necroptotic pathway [132]. Laukens et al. demonstrated that in leukemia cells, a Smac mimetic augmented TNFα-induced cell death via necroptosis in the absence of FADD or caspase-8 in apoptosis-resistant cells or via apoptosis in apoptosis-proficient cells, suggesting that Smac mimetics may be developed as novel chemotherapeutics to promote necroptosis as an alternative programmed cell death process to overcome apoptosis resistance [132]. Furthermore, ROS were found to be required for the regulation of Smac mimetic/TNFα-induced necroptotic signaling, specifically in the enhancement of RIPK1/RIPK3 necrosome stabilization when stimulated by a Smac mimetic/TNFα [133]. Recently, Hannes et al. [134] reported that the Smac mimetic BV6 alone or combined with TNFα could induce necroptosis in pancreatic cancer cells in which apoptosis was blocked by inducing the formation of the RIPK1/RIPK3 necrosome.

The ubiquitin-proteasome system is the major mechanism by which cells selectively degrade unneeded or damaged cellular proteins by proteolysis and thus, has an essential role in many cellular processes including cell cycle and cell demise [151]. Proteasome inhibitors have therefore been deemed as a promising anti-cancer agent clinically. For instance, a classic proteasome inhibitor, bortezomib, has been a success in treating multiple myeloma [152]. Recently, Moriwaki et al. [135] reported that proteasome inhibitors MG132 and bortezomib can trigger the RIPK3/MLKL dependent necroptosis in both fibroblasts in mouse models and human leukemia cells. Notably, proteasome- inhibitor- induced necroptotic pathway is independent of caspase inhibition, yet still requires intact RIP homotypic interaction motif (RHIM) [135]. The study has indicated that the ubiquitin-proteasome system may be a potential regulatory pathway for RIPK3-dependent necroptosis and that proteasome inhibitors can be developed as an anti-cancer agent targeting necroptotic pathway.

Obatoclax (GX15–070), which is a small-molecule Bcl-2 inhibitor of antiapoptotic Bcl-2 proteins, was recently reported to trigger necroptosis by assembling the necrosome onto autophagosomes; this process links the induction of autophagy to cell death via necroptosis [136].

Polyinosinic:polycytidylic acid (PolyI:C), which is a viral dsRNA analog, was reported to trigger necroptosis in cervical cancer, which strictly depended on the expression of RIPK3. Necroptotic cervical cancer cells produce IL-1α, which was essential for activating DCs to release IL-12, which is a cytokine crucial for antitumor activities [92]. In colon carcinoma cell lines, in addition to inducing immune or macrophage activation, PolyI:C can trigger necroptosis alone or combined with the pan-caspase inhibitor zVAD, supporting tumor retardation mediated by immune effectors in vivo [57].

NUPRI1 is a member of intrinsically disordered proteins (IDPs), which engages in various cancer-related processes including cell-cycle regulation, apoptosis [153], cancer metastasis [154], DNA repair response [155] and etc. It has recently drawn remarkable attention because it has been shown to promote progression and development of pancreatic cancer [137], making it a promising anti-cancer therapeutic target. A newly synthesized NUPR1 inhibitor, named ZZW-115 was reported to inhibit the growth pancreatic xenografted tumors in vivo dose-dependently via the induction of necroptosis by inducing mitochondrial metabolism rupture, and there was no evidence of off-target effects [156]. The study suggested that ZZW-115 is a promising therapeutic agent for treating various cancers due to its effectiveness in targeting NUPR1.

Since necroptosis may also occur under various physiological conditions, one major concern about pronecroptotic therapy is whether necroptosis can be selectively induced in cancer cells yet cause no harm to normal cells. In fact, multiple agents and chemotherapeutic drugs that have been granted for marketing or in clinical trials have been recognized as selective necroptosis inducer in cancer cells in specific cancer types, including shikonin and its analogs [41, 157], TRAIL [16], obatoclax [158], metal nanoparticles [148] et cetera. The safety of those compounds and drugs in vivo have been verified, suggesting triggering necroptosis in cancer cells is not necessarily injurious to normal cells [159]. However, to further improve the selectivity of the agents and drugs targeting necroptosis, future therapeutics can combine necroptosis inducers to tumor-guiding agents or tumor-targeting antibodies.