Necroptosis: an alternative cell death program defending against cancer

Abstract

One of the hallmarks of cancer is resistance to programmed cell death, which maintains the survival of cells en route to oncogenic transformation and underlies therapeutic resistance. Recent studies demonstrate that programmed cell death is not confined to caspase-dependent apoptosis, but includes necroptosis, a form of necrotic death governed by Receptor-Interacting Protein 1 (RIP1), RIP3, and Mixed Lineage Kinase Domain-Like (MLKL). Necroptosis serves as a critical cell-killing mechanism in response to severe stress and blocked apoptosis, and can be induced by inflammatory cytokines or chemotherapeutic drugs. Genetic or epigenetic alterations of necroptosis regulators such as RIP3 and cylindromatosis (CYLD), are frequently found in human tumors. Unlike apoptosis, necroptosis elicits a more robust immune response that may function as a defensive mechanism by eliminating tumor-causing mutations and viruses. Furthermore, several classes of anticancer agents currently under clinical development, such as SMAC and BH3 mimetics, can promote necroptosis in addition to apoptosis. A more complete understanding of the interplay among necroptosis, apoptosis, and other cell death modalities is critical for developing new therapeutic strategies to enhance killing of tumor cells.

1. Introduction

Eliminating unwanted or damaged cells through cell death is essential for fundamental biological processes such as animal development, tissue homoeostasis, and stress response. Deregulated cell death contributes to a number of human diseases including cancer and inflammation. During oncogenic transformation, neoplastic cells become resistant to cell death, which allows them to survive and acquire additional oncogenic mutations [1]. Cell death is also one of the most critical anticancer mechanisms triggered by a variety of chemotherapeutic drugs. Thus, targeting cell death pathways is an attractive strategy for developing novel anti-cancer therapies.

Cell death can be classified into several major forms based on morphological features and biochemical characteristics, including apoptosis, necrosis, autophagic death, and mitotic catastrophe [2]. Apoptosis and necrosis represent the two most distinctive and best understood cell death modalities. Apoptosis, known as programmed cell death, is highly regulated, while necrosis was previously thought to be a random and passive process without much regulation. However, this traditional view has been changed by studies in the past two decades that led to the identification of a regulated form of necrosis. It was observed that inhibition of caspases, proteases that are essential for apoptosis, did not suppress but exacerbated cell death with signs of necrosis in certain cell types [3]. Receptor-interacting kinase 1 (RIP1; RIPK1) was shown to be a critical regulator of necrosis in caspase-inhibited cells [4]. Chemical biology studies identified a small-molecule inhibitor of this type of cell death [5], which was subsequently shown to be the inhibitor of RIP1 [6]. Several recent studies demonstrated RIP3 as the downstream mediator of RIP1 [7–9], and Mixed Lineage Kinase Domain-Like (MLKL) as the key player in the execution of cell death [10]. A number of in vivo studies have established the physiological and pathological relevance of necrosis [11]. This type of cell death has now been defined as necroptosis [5], which is regulated by and requires RIP1, RIP3 and MLKL [2]. Accumulating evidence indicates that necroptosis functions as a safeguard mechanism for killing cancer cells that fail to die by apoptosis, suggesting a pivotal role in cancer biology and therapy.

2. Regulation and mechanisms of necroptosis

2.1 Characteristics of necroptosis

Necroptosis has several distinctive characteristics compared to other types of cell death, in particular, apoptosis (Table 1). Necroptotic cells share several morphological features of necrosis, including early loss of plasma membrane integrity, translucent cytosol, and swelling mitochondria (Figure 1) [12]. In contrast, apoptotic cells lack these features, and are characterized by cell shrinkage, plasma membrane blebbing, and condensed and fragmented nuclei and organelles (Figure 1). At the biochemical level, necroptotic cells show marked depletion of cellular ATP and leakage of intracellular contents, in contrast to apoptosis, which is a more energy-consuming process requiring a relatively higher level of cellular ATP. At the molecular level, necroptosis is caspase-independent and signals through RIP1, RIP3 and MLKL, while apoptosis requires caspase activation and is mediated by interplays of the Bcl-2 family proteins or activation of death receptors. Another key feature of necroptosis is that permeabilization of plasma membrane can lead to release of so-called Damage Associated Molecular Patterns (DAMPs), such as high-mobility group box 1 (HMGB1) protein and mitochondrial DNA [13, 14], which can trigger a robust immune response and inflammation [15]. In contrast, corpses of apoptotic cells are engulfed and then dissolved through the process phagocytosis [16], which does not typically incur a strong immune response [17].

Table 1.

Comparison of necroptosis and apoptosis markers

Necroptosis	Apoptosis
Morphological markers
Plasma membrane swelling and rupture	Plasma membrane blebbing
Cytoplasm swelling and vacuolization	Cytoplasm fragmentation; apoptotic body formation
No nuclear fragmentation	Nuclear condensation and fragmentation
Mitochondria and organelle swelling	Organelle fragmentation
Biochemical markers
ATP depletion	ATP increase
ROS production	ROS production
Calcium and sodium influx	Mitochondria outer membrane permeabilization
Molecular markers
Signaling by the Bcl-2 family proteins	Signaling by RIP1, RIP3, and MLKL
Regulation by death receptors	Regulation by death receptors
Inhibition by caspases	Caspase activation
Random DNA degradation	DNA digestion by endonucleases
Extracellular release of DAMPs, such as HMGB1 and mitochondrial DNA	Cytosolic release of mitochondrial cytochrome c, SMAC and AIF

Figure 1. Morphological features of necroptosis and apoptosis in cancer cells.

HT29 colon cancer cells treated with an anticancer drug for 48 hours were analyzed by transmission electron microscopy. The cell undergoing necroptosis shows plasma membrane rupture and permeabilization, compared to the intact plasma membrane with blebbing in the apoptotic cell (red arrowheads). The necroptotic cell exhibits cytoplasm swelling and vacuolization, which are absent in the apoptotic cell (green arrowheads). The necroptotic cell has swelled mitochondria, in contrast to those in the apoptotic cells (yellow arrowheads). The necroptotic cell also lacks condensed and fragmented nuclei seen in the apoptotic cell (blue arrowheads).

Despite these distinctive features, necroptosis is tightly connected with apoptosis and other cell death programs, which presents interesting opportunities as well as challenges to understand the regulation and relative contributions of different cell death forms. Choices of the cell death pathways are dependent upon a variety of factors such as cell type, genetic background, stimuli, and intracellular redox, PH and ion levels. Under most experimental conditions, apoptosis is the default cell death program, and necroptosis serves as a fail-safe mechanism to eliminate stressed cells that fail to undergo apoptosis. However, necroptosis can be predominant under certain circumstances, such as viral infection and exposure to certain natural products [18, 19]. More often, a continuum of necroptosis and apoptosis was observed in cells [20, 21]. Levels of stress and extent of cellular damage also affect choices of cell death programs. Increased stress levels and intensified death signals can switch cell death from apoptosis to necroptosis [22]. The interrelationship between necroptosis and other cell death pathways is complicated and has been a subject of extensive investigations.

2.2 Molecular mechanisms of necroptosis

Necroptosis is often triggered by stimuli that also induce apoptosis via the extrinsic pathway, such as members of the tumor necrosis factor (TNF) family of cytokines including TNF-α, FAS ligand (FASL; CD95), and TNF-related apoptosis-inducing ligand (TRAIL) [23]. Upon ligand stimulation, activated receptors of these stimuli interact with RIP1 through their respective death domains, and recruit cellular inhibitor of apoptosis proteins (cIAPs), such as cIAP1 and cIAP2, to form a plasma membrane associated complex (Figure 2), which leads to activation of pro-survival NF-κB and mitogen-activated protein kinases (MAPKs) [24]. During this process, RIP1 becomes polyubiquitinated by cIAPs and other E3 ubiquitin ligases. Auto-ubiquitination and subsequent degradation of cIAPs, which is stimulated by second mitochondria-derived activator of caspases (SMAC) or small-molecule SMAC mimetics (also known as IAP inhibitors) [25–27], promotes deubiquitination of RIP1 by deubiquitinases, including cylindromatosis (CYLD) and A20 [28, 29]. This results in RIP1dissociation from the plasma membrane and its conversion from a pro-survival into a pro-death protein (Figure 2). RIP1 binding to FAS-Associated Death Domain (FADD) recruits procaspase-8 [26], leading to activation of caspase-8 and induction of apoptosis (Figure 2). The activated caspase-8 inhibits necroptosis by cleaving the core regulators of necroptosis such as RIP1 and RIP3 [30, 31]. If caspase-8 activity is ablated by caspase inhibitors or genetic knockout, the mode of cell death switches to necroptosis. RIP3 and RIP1 bind to each other through their respective homotypic interaction motif (RHIM) domains to form a functional amyloid signaling complex [7–9, 32], leading to auto-phosphorylation of RIP3 at serine 227 and subsequent recruitment of the pseudokinase MLKL [10]. It is important to note that unlike RIP3, which is consistently pro-death, RIP1 is pleotropic, and can both positively and negatively regulate necroptosis and RIP3 activity. For example, RIP1 was shown to suppress spontaneous RIP3 activation in the cytosol [33].

Figure 2. Necroptosis regulators and pathways.

Upon TNF-α stimulation, the activated TNF receptor (TNFR) interacts with RIP1 and recruits cIAP1 and cIAP2 to form a plasma membrane associated complex, resulting in RIP1 polyubiquitination (Ubs). Inhibition of cIAPs (by SMAC or SMAC mimetics) leads to deubiquitination of RIP1 by CYLD and dissociation of RIP1. RIP1 then binds to FADD and procaspase-8 to form a complex, which activates caspase-8 (Casp-8) and leads to apoptosis induction. If caspase-8 activity is inhibited, RIP1 binds to RIP3 to form necrosome and promotes RIP3 auto-phosphorylation and subsequent activation, allowing RIP3 to recruit and phosphorylate MLKL. This results in oligomerization of MLKL, membrane insertion of MLKL oligomers, disruption of plasma and intracellular membrane integrity, and necroptotic death. Other necroptotic stimuli, including FASL, TRAIL, LPS, dsRNA (such as poly (I:C)) and interferon γ (IFNγ), can stimulate their respective receptors to activate RIP1 and/or RIP3 to promote necroptosis. Viral infection directly activates RIP3 through DAI. Anticancer agents and genotoxic stress can also induce RIP1/RIP3-dependent necroptosis. Several inhibitors of necroptosis have been developed, including the RIP1 inhibitor necrostatin-1 (Nec-1), the RIP3 inhibitors GSK843 and GSK872, and the MLKL inhibitor necrosulfonamide (NSA).

In addition to death-receptor ligands, necroptosis can also be triggered by other stimuli, including engagement of Toll-like receptors (TLRs) 3 and 4 by lipopolysaccharides (LPS) and double-stranded RNA (dsRNA) or analogs such as polyinosinic-polycytidylic acid (poly(I:C)), T-cell receptor stimulation, interferons (IFNs), DNA damage and viral infection (Figure 2) [23]. Depending on the stimulus, RIP3 can be activated through RIP1-dependent or -independent mechanisms that often involve a RHIM-RHIM interaction. For example, TLRs can directly activate RIP3 without RIP1 via another RHIM-containing protein known as TRIF [34, 35]. Upon viral infection, DNA-dependent activator of interferon regulatory factors (DAI), a RHIM-containing cellular protein, can activate RIP3 directly, likely through a RHIM-RHIM interaction [36]. After herpes simplex virus 1 (HSV-1) infection, the viral ribonucleotide reductase large subunit (ICP6) interacts with RIP3 through a RHIM-RHIM interaction to trigger necroptosis and host defense, which do not require CYLD and TNFR [37, 38]. In addition, cells with high levels of RIP3 or enforced oligomerization of RIP3 can bypass the requirement of RIP1 to induce necroptosis [33]. Interferons induce RIP1/RIP3-dependent necroptosis through transcriptional activation of interferon responsive genes [39, 40].

Biochemical, chemical biology, and genetic studies have identified MLKL as a major executioner of necroptosis downstream of RIP1/RIP3 [41]. MLKL is normally kept inactive as a monomer in the cytosol by its kinase-like domain [10]. Upon activation, RIP3 binds to MLKL through its kinase domain to promote MLKL phosphorylation at threonine 357 and serine 358 residues [10], which destabilizes the monomeric structure of MLKL and promotes its oligomerization [42, 43]. The oligomerized MLKL, which has affinity for phosphatidylinositol lipids and cardiolipin, translocates from the cytosol to the plasma and intracellular membranes, where it disrupts membrane integrity to promote necroptotic death (Figure 2) [42–44].

Several different mechanisms have been proposed to explain how oligomerized MLKL causes cell death. The complex formed by RIP1, RIP3 and MLKL can promote phosphorylation of the long form of phosphoglycerate mutase family member 5 (pGAM5L), which stimulates auto-phosphorylation of the short form of pGAM5 (pGAM5S) [45]. pGAM5S subsequently binds to the mitochondrial fission factor, dynamin-related protein 1 (Drp1), and activates its GTPase activity to promote mitochondrial fission and cell death [10, 45]. However, the role of the pGAM5-Drp1 axis and mitochondria in necroptosis appears to be context-dependent [46, 47]. It was also suggested that oligomerized MLKL can promote influx of calcium through the calcium channel TRPM7 protein [44], or influx of sodium [48], to promote cell death. Nonetheless, the question still remains as to whether the observed influx of ions in necroptotic cells is a consequence or the cause of cell death [34]. Production of reactive oxygen species (ROS) due to disruption of mitochondrial membrane integrity by MLKL may also contribute to necroptotic death [49]. Therefore, the execution mechanism downstream of MLKL in necroptosis remains to be elucidated.

Genetic studies in mice have demonstrated the roles of RIP1, RIP3, MLKL, and other necroptosis regulators in vivo. Knockout (KO) of RIP1 or RIP3 rescued developmental defects, embryonic lethality, and inflammation caused by KO of caspase-8 or FADD [30, 50–52]. KO of MLKL blocked TNF-induced necroptotic death [41, 53]. A deubiquitinase defective CYLD allele rescued extensive necroptosis and inflammation caused by conditional deletion of FADD in the skin [54]. Consistent with its pleotropic functions, RIP1 sometimes shows pro-survival activity and inhibits both apoptosis and necroptosis. RIP1-KO mice die of systemic inflammation at birth, which is rescued by combined loss of caspase-8 and RIP3 [55–57]. Conditional KO of RIP1 caused lethal intestinal pathology due to FADD/caspase-8-mediated apoptosis in the intestinal epithelial cells [58, 59]. These findings reinforce the notion that necroptosis is actively suppressed by apoptosis or apoptotic regulators to ensure normal development and tissue homeostasis.

2.3 Other forms of regulated necrosis

In addition to necroptosis, several other forms of regulated necrosis have been described. DNA alkylating agents, glutamate toxicity, oxidative stress, hypoxia, and ischemia can trigger necrosis that is dependent on poly(ADP-ribose) polymerase (PARP1), which catalyzes poly(ADP-ribosyl)ation of proteins involved in DNA damage response. PARP1-dependent cell death is mediated by translocation of apoptosis-inducing factor (AIF) from the mitochondrial intermembrane space into the nucleus [60], and activation of calpain proteases [61, 62]. PARP1-mediated necrotic death was due to energy crisis caused by depletion of nicotinamide adenine dinucleotide (NAD), and subsequently that of ATP [63]. Cells using aerobic glycolysis to support their bioenergetics such as cancer cells are more sensitive to this form of death [20].

Another form of regulated necrosis is mediated by the mitochondrial function of the tumor suppressor p53. In response to oxidative stress, p53 accumulates in the mitochondrial matrix and triggers mitochondrial permeability transition pore (PTP) opening and necrosis by interacting with the PTP regulator cyclophilin D [64], leading to ATP depletion and cell death [22]. In this case, the killing activity of p53 is independent of its transcriptional activation function [22]. In addition, several other non-apoptotic cell death modalities, such as pyroptosis and ferroptosis, are also considered to be regulated necrosis [17]. However, it has been difficult to distinguish different forms of necrosis due to lack of distinctive molecular markers. In general, these other forms of necrosis are not considered to be necroptosis due to a lack of evidence for a requirement of RIP1 or RIP3 [63].

2.4 Detection of necroptosis

Due to lack of specific molecular markers of necroptosis, a combination of approaches is often required to distinguish necroptosis from other cell death modalities. For cultured cells, transmission electron microscopy (TEM) is commonly used to provide morphological evidence of necrosis (Figure 1). Extracellular release of HMGB1, loss of cell viability, and depletion of ATP can be used as markers of necroptosis, but do not distinguish necroptosis from other types of necrotic death. More definitive evidence for necroptosis is the requirement of RIP1, RIP3, and/or MLKL, which can be established using approaches such as gene targeting, siRNA/shRNA knockdown, kinase-dead or interacting domain deficient mutants, or small-molecule inhibitors. The RIP1 inhibitor necrostatin-1 (Nec-1) has been widely used, which inhibits both catalytic and allosteric functions of RIP1 by inducting conformational changes [65]. The MLKL kinase inhibitor necrosulfonamide (NSA) [10], and the RIP3 inhibitors GSK843 and GSK872 [35, 66], can also be used for this purpose (Figure 2). Due to the pleiotropic effects of RIP1, RIP3 and MLKL serve as more specific markers of necroptotic death.

Biochemical markers are commonly used to study necroptosis. Activation of RIP1, RIP3, and MLKL in necroptosis can be detected by changes in their phosphorylation status or membrane accumulation using immunoblotting or immunostaining [42, 67]. Necrosome formation can be detected by RIP1/RIP3 and RIP3/MLKL interactions or complex formations using immunoprecipitation or other methods [7]. It is more challenging to study necroptosis in vivo. Increased mRNA and protein expression of RIP1, RIP3 and MLKL is indicative of predisposition to necroptosis [68]. Necroptotic death can be verified by decreased cell death or tissue injury in RIP1, RIP3, or MLKL deficient animals, or the effects of their respective small-molecule inhibitors [68]. For human specimens, necroptotic cell death is suggested by cell loss detected by H&E staining, and positive staining of Tdt-mediated dUTP nick end labeling (TUNEL), but negative staining of apoptosis markers such as active caspase-3. HMGB1 release could be measured by ELISA assays in the blood samples from patients [69].

3. Necroptosis as a safeguard mechanism against cancer

Accumulating evidence suggests that similar to apoptosis, necroptosis also functions as a barrier against tumor development. During tumor progression, a variety of genetic, epigenetic, and expression alterations of key necroptosis regulators are involved, which might allow tumor cells to evade the fate of necroptotic death, and possibly immune surveillance.

3.1 Alterations of necroptosis regulators in cancer

A number of studies reported downregulation or mutations in the core necroptotic machinery, including RIP1, RIP3, MLKL and CYLD, in various types of cancer (Table 2). RIP3 and CYLD were found to be markedly downregulated in chronic lymphocytic leukemia (CLL), which compromises TNF-α-induced necroptosis in CLL cells [70]. The downregulation of CYLD in CLL is due to transcriptional repression by lymphoid enhancer-binding factor 1 (LEF1), a downstream effector of the Wnt/β-catenin pathway [70]. RIP1 and RIP3 expression is significantly decreased in colon, breast, and lung cancer cells and primary tumors compared to corresponding normal tissues, which compromises the response of cancer cells to necroptotic stimuli [71–73]. Silencing of RIP1 and RIP3 in cancer cells is due to promoter hypermethylation [71, 73], or transcriptional suppression by hypoxia [72]. Furthermore, population studies indicate that single nucleotide polymorphisms (SNPs) in the genomic region of RIP3 are associated with increased risk of non-Hodgkin’s lymphoma [74].

Table 2.

Necroptosis regulators and their alterations in cancer

Proteins	Role in necroptosis	Function in necroptosis	Alterations in cancer	References
RIP1	Core machinery	Binds to RIP3 to form necrosome	Loss of expression	[8, 72, 119]
RIP3	Core machinery	Binds to RIP1 to form necrosome; phosphorylates MLKL	Loss of expression; promoter methylation	[7–9, 71, 72]
MLKL	Core machinery	Phosphorylation by RIP3; oligomerizes and inserts into membranes	Downregulation; mutations	[10, 42, 44, 75, 76]
CYLD	Signaling	Deubiquitinates RIP1; triggers ripotosome formation	Frequent mutations; downregulation	[28, 70, 78, 81]
cIAP1, 2	Inhibition	Polyubiquitinate RIP1; block ripotosome formation	Overexpression; amplification	[25–27, 82, 102]
Caspase-8	Inhibition	Cleaves RIP1, RIP3 and CYLD	Mutations	[30, 31, 120]
FADD	Inhibition	Inhibits necrosome formation	Rare mutations	[30, 51, 52, 121]
pGAM5	Execution	Activates Drp1 GTPase activity		[45]
Drp1	Execution	Promotes mitochondrial fission		[45]
Bcl-XL	Inhibition	Binds to pGAM5	Overexpression	[84, 85, 122]

Mutations in highly conserved residues located in the C-lobe of the MLKL pseudokinase domain, including F398I and L291P, have been identified in human stomach adenocarcinomas [75]. These loss-of-function mutations abrogate the ability of MLKL to promote necroptosis [41]. Downregulation of MLKL in pancreatic adenocarcinomas was associated with decreased patient survival [76].

CYLD was initially identified as a tumor suppressor gene mutated in Familial Cylindromatosis [77]. Germline and somatic mutations of CYLD have been detected in a variety of tumors [78], and are associated with poor clinical outcomes following chemotherapy [79]. CYLD was also found to be negatively regulated at the transcriptional level by Snail, which promotes tumor progression in malignant melanoma [80]. CYLD mRNA expression was shown to be downregulated in colon and hepatocellular carcinoma cells and tumor samples [81].

Another mechanism leading to compromised necroptosis in cancer cells is overexpression of pro-survival proteins such as IAPs. IAPs, including cIAP1, cIAP2 and XIAP, are often aberrantly expressed in a variety of human tumors due to genomic amplification, elevated mRNA and protein expression, or loss of their endogenous inhibitors such as SMAC [82]. These changes in IAPs have been linked to poor prognosis of cancer patients [82]. Aberrant activation of IAPs was thought to primarily affect apoptosis in cancer cells by inhibiting caspases and promoting the pro-survival NF-κB signaling [83]. It is now clear that cIAP depletion facilitates TNF-α-induced necroptosis [8]. Furthermore, Bcl-X_L, a pro-survival Bcl-2 family member often overexpressed in cancer [84], could also suppress necroptosis [85]. To what extent necroptosis is affected by alterations of IAPs, Bcl-X_L, and other pro-survival factors remains to be established by ruling out the effects on apoptosis.

3.2 Necroptosis in immune surveillance of cancer

The immune system is believed to function as a primary defense system against cancer by identifying and destroying nascent tumors via a collective action of several immune cell types and effector molecules [86]. In contrast to apoptosis, which seems to be immunologically quiescent, necroptosis is often a more potent inducer of immune response through the release of DAMPs into tissue microenvironment [17]. As a consequence, necroptotic cells can recruit inflammatory cells to survey the extent of tissue damage, remove associated necroptotic debris, and aid tissue remodeling and repair [87]. Necrosis was also used to explain adaptive immunity in seemingly infection-free situations, such as in spontaneous and therapy-induced tumor rejection [88].

Necroptosis has been shown to trigger activation of specific types of immune cells and adaptive immunity against tumor challenge. Necroptotic cells can provide both antigens and inflammatory stimuli for dendritic cells (DCs), which in turn activate CD8⁺ T cells through a process called antigen cross-priming [89]. CD8⁺ T cells primed by immunization with necroptotic cells produced multiple effector cytokines, exhibited in vivo cytolytic activity, and protected mice from tumor challenge [90]. Robust CD8⁺ T cell cross-priming and subsequent tumor protection are dependent on RIP1 signaling and NF-κB-mediated transcription [90]. RIP3 was recently shown to be critical for the function of natural killer T (NKT) cells by regulating cytokine expression via its effects on pGAM5 and Drp1, and RIP3 KO mice had reduced NKT-cell-mediated immune responses to metastatic tumor cells [91]. These findings suggest that necroptosis may elicit antitumor immunity by activating CD8⁺ T cells or NKT cells.

On the other hand, inflammation, which can be triggered by necroptosis, is considered to be a tumor-promoting mechanism [1]. Inflammation and NF-κB activation are often detected in early-stage tumors, and facilitate tumor progression by supplying bioactive molecules in the microenvironment to promote tumor cell proliferation and survival, as well as tumor angiogenesis, invasion and metastasis [87, 92, 93]. Mutagenic chemicals released by inflammatory cells, such as ROS, can accelerate genetic evolution of tumor cells [87]. The crosstalk between tumor cell necroptosis and inflammation, which remains to be elucidated, is likely to be significant in tumor development and progression.

3.3 Necroptosis as a defense mechanism against viral infection

Viral infection is a significant etiological factor of tumor development. It is estimated that one in five cancer cases worldwide is caused by viral infection [94]. Up to now, 7 tumor-causing viruses have been identified, including human papillomavirus (cervical carcinoma), Epstein-Barr virus (B-cell lymphoproliferative disease and nasopharyngeal carcinoma), Kaposi's sarcoma herpesvirus (Kaposi's Sarcoma and primary effusion lymphomas), hepatitis B and hepatitis C viruses (hepatocellular carcinoma), and Human T-cell leukemia virus-1 (T-cell leukemia) [94]. Killing and removal of infected cells is a key defensive mechanism against viral infection and subsequent cancer development [95]. Viruses have evolved a variety of mechanisms to suppress cell death and prolong survival of host cells to ensure reliable viral replication and dissemination.

Viruses express caspase 8 inhibitors, Bcl-2-like proteins, IAP-like proteins or other factors to inhibit apoptosis of host cells [96]. Emerging evidence suggests that necroptosis may become paramount for host defense when apoptosis is compromised, and suppressing necroptosis is necessary for effective viral dissemination. For example, cells infected with murine cytomegalovirus (MCMV) can be eliminated by necroptosis mediated by the interaction of RIP3 and DAI, an intracellular sensor of viral DNA, via their respective RHIM domains [36]. MCMV encodes a RHIM-containing protein called viral inhibitor of RIP activation (vIRA), which facilitates MCMV infection by disrupting RIP3 activity and necroptosis [36]. Infection of vaccinia virus induces RIP1/RIP3-dependent necroptosis, and mice with RIP3 KO or knock-in of a RIP1 kinase-dead mutant failed to control viral replication in vivo [97], suggesting that RIP1/RIP3-dependent necroptosis contains vaccinia virus infection. Infection with HSV-1 was restricted by RIP3/MLKL-dependent necroptosis via the interaction between RIP3 and ICP6, the viral ribonucleotide reductase large subunit, through a RHIM-RHIM interaction [37, 38]. In this case, TNF signaling was unnecessary for necroptosis induced by HSV-1 infection. However, another study showed that ICP6 inhibits TNF-α-induced and RIP3-dependent necroptosis by disrupting the RIP1/RIP3 interaction [98]. ICP6 also suppresses apoptosis by inhibiting caspase-8 activity [96], suggesting simultaneous targeting of both apoptosis and necroptosis by viral proteins. Furthermore, human cytomegalovirus (HCMV) blocks TNF-induced necroptosis in RIP3-expressing human cells following RIP3 activation and MLKL phosphorylation [99]. Most of the viruses that were found to induce necroptosis are not considered to be oncogenic, except for HCMV [100]. Direct evidence supporting the role of necroptosis against tumor-causing viruses is still lacking, and warrants further investigation. On the other hand, interferon-induced necroptosis of immune cells allows immune evasion and facilitates infection of some pathogens such as Salmonella Typhimurium [39].

4. Necroptosis in anti-cancer therapies

4.1 Induction of necroptosis by anticancer agents

Commonly used chemotherapeutic drugs, in particular DNA damaging agents, have been shown to induce necroptosis in cancer cells (Table 3). This is probably not surprising because DNA-damage-induced apoptosis is defective in most cancer cells due to p53 mutations and genetic and epigenetic alterations in other key apoptotic regulators [101]. Etoposide, a topoisomerase inhibitor drug, was shown to suppress cIAP expression, thereby inducing necroptosis through a RIP1-containing complex called ripoptosome [102]. The antimitotic drug taxol can induce RIP1-dependent necroptosis mediated through FADD double phosphorylation by the mitotic kinases Aurora-A and Plk1 [103]. 5-Fluorouracil (5-FU), etoposide, and camptothecin induced RIP1/MLKL-dependent, but RIP3-independent necroptosis in caspase-3-deficient colorectal cancer cells [104]. Cisplatin caused RIP3-dependent necroptosis in apoptosis-resistant esophageal cancer cells through necrosome formation and autocrine TNF-α signaling [105]. Epigenetic silencing of RIP3 contribute to resistance to 5-FU, cisplatin, camptothecin, and etoposide in colon and breast cancer cells [71].

Table 3.

Necroptosis-inducing anticancer agents

Anticancer agents	Mechanisms of necroptosis induction	References
Clinically used drugs
Cisplatin	RIP3-dependent; necrosome formation; autocrine TNF-α	[71, 105]
Etoposide	XIAP, cIAP1 and cIAP2 depletion; ripoptosome activation	[71, 72, 102]
5-FU	RIP1 activation; autocrine TNF-α	[71, 104, 116, 123]
Taxol	FADD depletion; double-phosphorylation of FADD by Aurora-A and Plk1; RIP1 activation	[103]
Sorafenib	ROS production; RIP1activation	[106, 107]
Interferons	TLR3/TLR4 activation; necrosome formation; independent of RIP1 kinase activity	[34, 124]
Investigational agents
TRAIL	TNFR1 signaling; RIP1/RIP3-dependent	[8, 53, 111]
Poly(I:C)	Secretion of IFNs; TLR3/TLR4 activation; Necrosome formation	[34, 110]
IAP antagonists	IAP depletion; ripoptosome activation	[7–9]
ABT-737, obatoclax	Atg5-dependent necrosome formation on autophagosome	[113]
Natural products
Shikonin	ROS production; RIP1/RIP3 activation	[19]
Chal-24	cIAP1 and cIAP2 degradation; ripoptosome activation	[109]
Staurosporine	RIP1/RIP3/MLKL-dependent	[108]

Necroptosis can also be induced by kinase inhibitors, natural products, or immune modulators. Sorafenib, an FDA-approved multi-kinase inhibitor drug, can cause RIP1-dependent necroptosis in prostate cancer cells through induction of autophagy and interaction of RIP1 with p62/sequestosome 1 [106]. A combination of sorafenib with the histone deacetylase inhibitor Givinostat induced RIP1-dependent necroptosis through ROS production and activation of the BH3-only Bcl-2 family protein Bim [107]. Staurosporine, a pan-kinase inhibitor and a widely used apoptosis inducer, was found to promote RIP1/MLKL-dependent necroptosis in U-937 lymphoma cells under caspase-compromised conditions [108]. Chalcone-24, a chalcone derivative, induced RIP1/RIP3-dependent necroptosis through autophagy induction and cIAP degradation [109]. Shikonin, a naturally occurring naphthoquinone, could promote RIP1-dependent cell death in cancer cells [19]. Interestingly, necroptosis induced by these agents is often accompanied by autophagy, which may be responsible for suppression of apoptosis and bias toward necroptosis. It has been suggested that the viral dsRNA analog Poly (I:C) can be used as an adjuvant for cancer immunotherapies. A recent study demonstrated that stimulation of cervical cancer cells with Poly (I:C) caused RIP3-dependent necroptosis, which promotes activation of dendritic cells (DC) to produce IL-12, a cytokine critical for anti-tumor immunity [110]. This finding suggests that the status of RIP3 in cervical cancer cells determines the response to Poly-(I:C)-based immunotherapy, and needs to be assessed prior to therapy [110].

4.2 Manipulating necroptosis for improving anti-cancer therapies

Alterations of RIP1, RIP3, CYLD and MLKL likely explain defective necroptosis in some cancers, suggesting its restoration as a therapeutic strategy. In contrast to a plethora of necroptosis inhibitors, necroptosis activators have yet to be developed. Several strategies might be useful for restoring necroptosis in cancer cells.

Proapoptotic agents, including TRAIL, IAP inhibitors, and Bcl-2 inhibitors, can induce necroptosis in cancer cells when apoptosis is blocked. In preclinical studies, TRAIL, alone or in combination with other anticancer agents, has shown promising activities. It induces apoptosis in a wide range of cancer cells, while lacking significant toxicity to normal cells [22]. However, TRAIL has failed to show a significant benefit in clinical studies [22]. TRAIL induces necroptosis in cancer cells upon caspase inhibition [8, 53]. Changes in extracellular pH can switch TRAIL-induced cell death from apoptosis to RIP1/RIP3-mediated necroptosis in colon and liver cancer cells [111]. Induction of necroptosis may help to improve the anti-cancer activity of TRAIL in tumors with deregulated apoptosis.

cIAPs are commonly overexpressed in cancers and represent attractive therapeutic targets. Well-characterized structures of cIAPs and the binding interface with their endogenous inhibitor SMAC have facilitated pharmacological interventions. A number of small-molecule mimetics of the IAP-binding AVPI motif in SMAC have been identified, some of which have advanced into clinical trials for treating solid tumors and lymphomas [82]. cIAP inhibition seems to confer susceptibility to necroptosis, as exemplified by the use of SMAC mimetics to induce cIAP degradation and RIP1/RIP3-dependent necroptosis in cancer cells when caspase activation is inhibited [8].

Agents mimicking the BH3 domains of the proapoptotic Bcl-2 family members can restore apoptosis and selectively kill cancer cells that are addicted to antiapoptotic Bcl-2-like proteins for survival [112]. Obatoclax (GX15-070), a small-molecule inhibitor of antiapoptotic Bcl-2-family proteins, triggers necroptotic cell death by promoting necrosome formation on autophagosomal membranes [113]. Obatoclax was also found to induce autophagy-dependent necroptosis and overcome resistance to glucocorticoids in childhood acute lymphoblastic leukemia (ALL) [114]. BH3 mimetics, such as ABT-737 and ABT-263, could stimulate secretion of TNF-α [115], a prototypic necroptosis inducer. It is likely that necroptosis contributes to the cell-killing effect of BH3 mimetics, which are often not strong inducers of apoptosis when used as single agents. Therefore, targeting the Bcl-2 family proteins may provide another means for enhancing necroptosis in cancer cells.

As RIP3 is frequently silenced in cancer cells through promoter hypermethylation, demethylating agents such as decitabine (5-aza-2′-deoxycytidine) can be used to restore RIP3 expression and enhance chemotherapy-induced necroptosis [71]. Chemotherapy-induced necroptosis could also be enabled in p53-deficient colon cancer cells through inhibition of GSK3β [116]. Stimulating the expression and dimerization of RIP3 may be useful for enhancing necroptosis in cancer cells. Recovering CYLD function, either by increasing its expression using inhibitors of Snail or Notch-Hes1, or by inhibiting its SUMOylation to enhance deubiquitinase activity, promoted necroptosis of neuroblastoma cells [117]. Furthermore, elucidating the structural basis of the RHIM-RHIM interaction interfaces may help develop small molecules or peptide mimetics for modulating the key signaling event in necroptosis [118].

5. Future directions and conclusions

The research on necroptosis has been rapidly evolving in recent years, while its role in cancer remains to be a nascent area. Several critical issues remain to be solved. First, the interdependent relationship and crosstalk mechanisms of necroptosis and other forms of cell death need to be better elucidated. Dissecting the precise functional role of necroptosis and underlying mechanisms in tumor development and therapy-induced cell killing will provide new opportunities for more specific manipulation of necroptosis. Second, more specific markers of necroptosis need to be identified. Analysis of necroptosis, especially in vivo and in human tissue samples, has been hindered by the lack of molecular markers. An expanding and improved tool box will likely provide more definitive evidence on aberrant regulation of necroptosis in cancer. Third, more physiologically relevant systems are necessary to study the physiological and pathological roles of necroptosis. Up to now, relatively artificial conditions, such as caspase inhibition and KO of a necroptosis suppressor or apoptosis activator, are necessary for the manifestation of necroptosis phenotypes. Last but not least, the connection between necroptosis and the immune system needs to be clarified in the setting of cancer development and therapies. Recent success of cancer immunotherapy highlights a key role of the immune system in preventing tumor progression, likely more important than previously thought. Clarifying how DAMPs and other signaling molecules released during necroptosis elicit specific types of immune response may be critical for understanding the seemingly opposite functions of immune system in cancer immune surveillance and tumor promotion.

In summary, recent studies have defined RIP1/RIP3/MLKL-dependent necroptosis as a new form of cell death responsible for killing stressed cells that are unable to die by apoptosis. Accumulating evidence suggests that necroptosis functions as an alternative barrier against cancer development, and necroptosis signaling can be explored to develop new anti-cancer therapies.

Highlights.

• Necroptosis is a form of regulated necrosis that kills apoptosis-deficient cells.

• Necroptosis is governed by RIP1, RIP3 and MLKL.

• Alterations of necroptosis regulators are frequently found in human tumors.

• Anticancer agents can promote necroptosis in addition to apoptosis.

• Unlike apoptosis, necroptosis elicits a more robust immune response.

Abbreviations

AIF

apoptosis-inducing factor

ALL

acute lymphoblastic leukemia

cIAP

cellular inhibitor of apoptosis protein

CLL

chronic lymphocytic leukemia

CYLD

cylindromatosis

DAMPs

Damage Associated Molecular Patterns

DCs

dendritic cells

Drp1

dynamin-related protein 1

dsRNA

double-stranded RNA

FADD

FAS-associated death domain

FASL

FAS ligand

5-FU

5-fluorouracil

HCMV

human cytomegalovirus

HMGB1

high-mobility group box 1

HSV-1

herpes simplex virus 1

IFN

interferon

KO

knockout

IL-1α

interleukin-1α

LPS

lipopolysaccharides

MAPK

mitogen-activated protein kinase

MCMV

murine cytomegalovirus

MLKL

mixed lineage kinase domain-like

NAD

nicotinamide adenine dinucleotide

Nec-1

necrostatin-1

NSA

necrosulfonamide

PARP1

poly(ADP-ribose) polymerase 1

pGAM5

phosphoglycerate mutase family member 5

poly(I:C)

polyinosinic-polycytidylic acid

PTP

permeability transition pore

RHIM

respective homotypic interaction motif

RIP

Receptor-Interacting Protein

ROS

reactive oxygen species

SMAC

second mitochondria-derived activator of caspases

SNP

single nucleotide polymorphism

TLR

Toll-like receptor

TNF

tumor necrosis factor

TLR

Toll-like receptor

TUNEL

terminal deoxynucleotidyl transferase mediated dUTP nick end labeling