The Role of the Key Effector of Necroptotic Cell Death, MLKL, in Mouse Models of Disease

Necroptosis is an inflammatory form of lytic programmed cell death that is thought to have evolved to defend against pathogens. Genetic deletion of the terminal effector protein—MLKL—shows no overt phenotype in the C57BL/6 mouse strain under conventional laboratory housing conditions. Small molecules that inhibit necroptosis by targeting the kinase activity of RIPK1, one of the main upstream conduits to MLKL activation, have shown promise in several murine models of non-infectious disease and in phase II human clinical trials. This has triggered in excess of one billion dollars (USD) in investment into the emerging class of necroptosis blocking drugs, and the potential utility of targeting the terminal effector is being closely scrutinised. Here we review murine models of disease, both genetic deletion and mutation, that investigate the role of MLKL. We summarize a series of examples from several broad disease categories including ischemia reperfusion injury, sterile inflammation, pathogen infection and hematological stress. Elucidating MLKL’s contribution to mouse models of disease is an important first step to identify human indications that stand to benefit most from MLKL-targeted drug therapies.


Introduction Necroptosis and Disease
Mixed Lineage Kinase Domain-Like (MLKL) was shown to be the essential effector of a pro-inflammatory, lytic form of programmed cell death called necroptosis in 2012 [1,2]. Like other forms of lytic cell death (e.g., pyroptosis), necroptosis is characterised by the release of Damage-Associated Molecular Patterns (DAMPs) and IL-33, IL-1α and IL-1β production, as recently reviewed by [3]. Unlike apoptosis, necroptosis and the canonical signalling pathway RIPK1-RIPK3-MLKL are not essential to the development and homeostasis of multicellular organisms [4]. The two most downstream effectors of necroptosis, MLKL and its obligate activating kinase RIPK3, can be deleted at the genetic level in laboratory mice without any overt developmental consequences in the absence of challenge [5][6][7]. Thus, it is widely held that MLKL and necroptosis have evolved primarily to defend against pathogenic insult to cells and tissues. This important role of necroptosis in pathogen defence is written in the DNA of many bacteria and viruses alike, which together encode several genes that disarm different facets of the necroptotic signalling pathway [8][9][10]. In evolution, genetic deletion of MLKL and/or RIPK3 in the ancestors of modern day carnivora, metatheria (marsupials) and aves (birds) show that complex vertebrates can survive without necroptosis when faced with infectious challenges of the 'real world' [11]. This adds to the precedent for the well-honed flexibility, redundancy and co-operation of different programmed cell death pathways in the defence against pathogens [12] and offers some biological guide that inhibiting MLKL pharmacologically in humans would not [12] and offers some biological guide that inhibiting MLKL pharmacologically in humans would not compromise pathogenic defence. Interestingly, recent studies suggest that the suppression of necroptosis may even reduce the resulting inflammatory response that is often more dangerous than the infection itself [13]. Mlkl gene knockout (abbreviated to both Mlkl −/− and Mlkl KO) mice are distinguishable from wild-type mice in numerous models of disease, a broad sampling of which are presented (Table 1) and illustrated ( Figure  1) here.
The therapeutic potential for necroptosis-targeted drugs lies largely in non-infectious indications. The first human clinical trials of RIPK1-targetted small molecule compounds were conducted in cohorts of rheumatoid arthritis, ulcerative colitis and psoriasis patients [14]. While these showed some promise in phase II, these were returned to the research phase in late 2019 by their licensee GlaxoSmithKline [15]. At the time of writing, there are only three active clinical trials in progress: a phase I trial assessing the safety of a new RIPK1 inhibitor (GFH312), a phase II study utilising RIPK1-binding compound, SAR443122, in cutaneous lupus erythematosus patients and a phase I/II study utilising RIPK1 inhibitor GSK2982772 in psoriasis [16]. There are currently no 'first in human' trials of RIPK3-or MLKL-binding compounds listed on http://clinicaltrials.gov (accessed on 28 May 21).   Wt mice utilised. [17] Renal infarction, no reperfusion Cholesterol crystal embolism.
Mlkl −/− mice protected from infarction (measured by infarct size, kidney injury and neutrophil infiltration). No difference in the extent of acute kidney injury (measured by eGFR) and kidney failure.
A20 MYC-KO Mlkl −/− mice were protected against inflammatory arthritis (thickness of rear ankles, histological scores of inflammation, cartilage and bone destruction), splenomegaly and showed reduced expression of IL-1β and TNFα.
Mlkl −/− and Wt mice showed same survival profile (mortality was monitored from 24 h to 144 h).
Mlkl −/− mice are indistinguishable from Wt in terms of body weight, food intake, ALT/AST, hepatic triglycerides, macrovesicular and microvesicular steatosis (H&E). Mlkl −/− mice have similar levels of CYP2E1, ER stress and hepatocyte apoptosis but mildly reduced levels of some hepatic inflammatory markers.

Mlkl Knock-Out (KO) and Constitutively Active (CA) Mice at Steady State
Following the discovery of its essential role in necroptosis [1,2], two Mlkl knockout mouse strains were independently generated by traditional homologous recombination [5] and TALEN technology [7]. More recently, CRISPR-Cas9 engineered Mlkl −/− [17,33,73], constitutively active point mutant Mlkl D139V [40], affinity tagged Mlkl [67], conditional Mlkl −/− strains [29,67] and antisense oligonucleotide (ASO) in vivo Mlkl knockdown [60] mouse models have also been used in the study of necroptosis. Mlkl −/− mice are born at expected Mendelian ratios and are overtly indistinguishable from wild-type littermates at birth and through to early adulthood [5,7]. Full body histological examination of 2 day old Mlkl −/− pups did not reveal any obvious morphological differences, including lesions or evidence of inflammation, relative to wild-type C57BL/6 mice of the same age [40]. Furthermore, no histological differences have been reported for the major organs of young adult Mlkl −/− mice [5,7]. Hematopoietic stem cell populations in the bone marrow [5] and CD4/CD8 T cell, B cell, macrophage and neutrophil mature cell populations in secondary lymphoid organs display no observable differences in adult mice [7]. At steady state, serum cytokines and chemokines are indistinguishable from age-matched wild-type littermates [65]. The genetic absence of Mlkl and thus necroptosis is generally considered to be innocuous at steady state in the C57BL/6 strain of laboratory mice at a typical experimental age (up to 16 weeks) when housed under conventional clean, pathogen free conditions.

Mlkl −/− Mice in Ischemia and Reperfusion Injury (IRI)
While grouped here for simplicity, MLKL and cell death in general has the potential to influence many facets of the physiological response to blood vessel occlusion/recanalisation and resultant end-organ damage. These facets include the aetiology of the infarction itself, the cellular damage incurred due to the deprivation of oxygen and ATP, the generation of reactive oxygen species that occurs following tissue reperfusion and the inflammation that ensues and convalescence after injury [77]. As summarized in Table 1 (see 'ischemia and reperfusion injury'), Mlkl −/− mice appear partially protected from the initial embolic insult [18,21]. For example, Mlkl −/− mice were reported to exhibit reduced infarct size and have better locomotive recovery day 7 post stroke [18]. This protective effect may in part be due to the role of MLKL and RIPK3 in regulating platelet function and homeostasis [72,78]. Mlkl knockout is also protective in models of hepatic and renal IRI [20,79]. Neutrophil activation and inflammation are significant contributors to hepatic IR injury [80]. Despite equivalent levels at steady state, the Mlkl −/− mice liver parenchyma shows significantly lower numbers of neutrophils 24 h post infarct [17]. This reinforces that the absence of MLKL can play a protective role at the initial ischemic stage and/or at later reperfusion stages, depending on the context.

Sterile Inflammation
The contribution of MLKL to mouse models of inflammation, which are not borne of pathogenic insult, termed sterile inflammation, is complex and varies according to the initiator, severity, and location within the body. This point is nicely illustrated by systemic inflammatory response syndrome (SIRS). Unlike catalytically inactive RIPK1 or RIPK3 deficiency in mice, Mlkl −/− mice are not protected against SIRS driven by low dose TNFα [20] or A20 deficiency [20]. When wild-type mice are pre-treated with compound 2, a potent inhibitor of necroptosis that binds to RIPK1, RIPK3 and MLKL, hypothermia is delayed [81]. Together these findings suggest protection against SIRS is necroptosis-independent and occurs upstream of MLKL. Mlkl −/− mice, however, are significantly protected against SIRS caused by high dose TNF [19,20,81]. Ripk3 −/− mice are similarly protected against high dose TNFα. Remarkably, in contrast to single knockouts, Ripk3 −/− Mlkl −/− double knockouts resemble wild-type mice and develop severe hypothermia in response to high dose TNFα [19]. This paradoxical reaction is yet to be fully explained by the field. Furthermore, the ablation of MLKL is seen to worsen inflammation induced by non-cleavable caspase-8 (seen in Casp8 D387A/D387A mice) [39] and A20 deficiency [20]. This indicates that necroptosis may serve to limit systemic inflammation in certain scenarios in vivo, a phenomenon also supported by examples of pathogen induced inflammation (see 'Infection', Table 1).
One major area of contention is the role of MLKL in mouse models of inflammatory bowel disease [31,32,34,35] and inflammatory arthritis [28,30]; however, key differences in experimental approach may explain these disparities (See Table 1, 'Mlkl −/− mice and wild-type control' details). Similarly, there have been conflicting reports investigating the role of MLKL-driven necroptosis in liver injury [24,25]. Whilst whole body knock-out of Mlkl confers protection, independent of immune cells [24], hepatocyte specific ablation of Mlkl reveals that necroptosis in parenchymal liver cells is, in fact, dispensable in immunemediated hepatitis [25]. It would therefore be of interest to know the cell type in which Mlkl −/− confers protection now that hepatocytes [25] and immune cells [24] have been ruled out. MLKL-deficiency mediates protective effects in models of more localised sterile inflammatory disease: dermatitis [36], cerulein-induced pancreatitis [7], ANCA-driven vasculitis [27], necrotising crescent glomerulonephritis [27] and oxalate nephropathy [22]. Consistent with these findings, mice expressing a constitutively active form of MLKL develop a lethal perinatal syndrome, characterised by acute multifocal inflammation of the head, neck and mediastinum [40].

Infection: Bacterial
MLKL-dependent necroptosis is thought to have evolved as a pathogen-clearing form of cell death. In support of this theory, 7 of the 10 murine models of bacterial infection examined here led to poorer outcomes in mice lacking MLKL. In response to both acute and chronic infection with Staphylococcus aureus and methicillin-resistant S. aureus (MRSA), Mlkl −/− mice suffer a greater bacterial burden and subsequent mortality [41,42]. This was observed across intravenous, subcutaneous, intraperitoneal, and retro-orbital methods of inoculation [41,42]. Interestingly, despite necroptosis being an inflammatory form of cell death, Mlkl −/− mice were found to have greater numbers of circulating neutrophils and raised inflammatory markers (caspase-1 and IL-1β) [41,42]. This is exemplified in models of skin infection where Mlkl −/− mice suffer severe skin lesions characterised by excessive inflammation [41]. This suggests that necroptosis is important for limiting bacterial dissemination and modulating the inflammatory response [41]. As an example, MLKL-dependent neutrophil extracellular trap (NET) formation was reported to restrict bacterial replication [42] and contribute to the pathogenesis of inflammatory disease, such as rheumatoid arthritis [82]. In addition to the important infection-busting role of MLKL in neutrophils, non-hematopoietic MLKL is also important for protection against gut-borne infections. MLKL-mediated enterocyte turnover and inflammasome activation were shown to limit early mucosal colonisation by Salmonella [49]. MLKL was even shown to bind and inhibit the intracellular replication of Listeria, presaging exploration into a more direct, cell-death independent mode of MLKL-mediated pathogen defence [48].
Through co-evolution, bacteria have developed ways to evade, and in some cases, weaponise MLKL and necroptosis for their own ends. Certain bacteria, such as Serratia marcescens and Streptococcus pneumoniae, produce pore-forming toxins (PFTs) to induce necroptosis in macrophages and lung epithelial cells [47,53]. Mlkl −/− mice are consequently resistant to these infections and survive longer than wild-type controls [47,53]. Although PFT-induced necroptosis was reported to exacerbate pulmonary injury in acute infection, it does promote adaptive immunity against colonising pneumococci [46]. Mlkl −/− mice demonstrate a diminished immune response, producing less anti-spn IgG antibody and thus succumb more readily to secondary lethal S. pneumoniae infection [46]. This suggests necroptosis is instrumental in the natural development of immunity to opportunistic PFTproducing bacteria [46]. Interestingly, Mlkl −/− has been described to be both protective against [44] and dispensable in [7] the pathogenesis of CLP-induced polymicrobial shock. Results derived from the CLP model of shock can be difficult to replicate, given the interfacility and even intra-facility heterogeneity of the caecal microbiota in mice [7]. Cell death pathways occur simultaneously during the progression of sepsis, and there is no conclusive evidence of which pathway plays the most deleterious role [7]. Similarly, MLKL appears dispensable for granulomatous inflammation and restriction of Mycobacterium tuberculosis colonisation [43].

Infection: Viral
MLKL can either protect against viral infection or contribute to viral propagation and/morbidity, depending on the type of virus. In initial studies, Mlkl −/− mice were indistinguishable from wild-type littermates in response to influenza A (IAV) [13,50,51] and West Nile virus infection [54]. Recent findings, however, suggest that Mlkl knockout confers protection against lung damage from lethal doses of influenza A [13]. Despite equivalent pulmonary viral titres to wild-type littermates, Mlkl −/− mice had a considerable attenuation in the degree of neutrophil infiltration (~50%) and subsequent NET formation and thus were protected from the exaggerated inflammatory response that occurs later in the infection [13]. In line with this finding, Mlkl −/− mice are protected against bacterial infection secondary to IAV [53]. A recent study also finds that MLKL-deficiency mediates protection against cardiac remodelling during convalescence following IAV infection by upregulating antioxidant activity and mitochondrial function [52], indicating the potential utility of MLKL-targeted therapies for both the acute and long term effects of viral infection.

Metabolic Disease
MLKL deficiency has shown diverse effects in at least four separate studies of nonalcoholic fatty liver disease (NAFLD). After 18 hours of a choline-deficient methioninesupplemented diet [55] or 8 weeks of a Western diet [17], Mlkl −/− mice are indistinguishable from wild-type controls. Following 12 weeks of a high fat diet (HFD) [56], however, Mlkl −/− mice appear resistant to steatohepatitis given that MLKL-deficiency promotes reduced de novo fat synthesis and chemokine ligand expression [56]. A similar effect is seen following 12 weeks of a high fat, fructose, and cholesterol diet, where Mlkl −/− mice are markedly protected against liver injury, hepatic inflammation and apoptosis attributed to inhibition of hepatic autophagy [57]. In line with these findings, mice treated with RIPA-56 (an inhibitor of RIPK1) downregulate MLKL expression and were found to be protected against HFD-induced steatosis [83]. These findings offer a tantalising clue that necroptosis may contribute to this disease, depending on the trigger. In contrast to NAFLD, MLKL does not appear to play a statistically significant role in acute or chronic alcoholic liver disease [58].
MLKL deficiency provides variable protection against the metabolic syndrome, depending on the challenge. MLKL deficiency appears to protect against dyslipidaemia with reduced serum triglyceride and cholesterol levels following a high fat diet [56] and Western diet [60], respectively. Whilst Mlkl −/− mice appear to have significantly improved fasting blood glucose levels given improved insulin sensitivity following 16 weeks of a high fat diet [59], there is conflicting evidence on the effect at steady state [56,59]. There is also conflicting evidence on the role of MLKL deficiency in adipose tissue deposition and weight gain [56,59]. Whilst comparable at baseline, after 16 weeks of HFD, Mlkl −/− mice gained significantly less body weight, notably visceral adipose tissue, than their wild-type littermates [59]. Blocking upstream RIPK1, has a similar effect [84]. Yet, in another study, after 12 weeks of HFD there were no significant differences found in body weight between Mlkl −/− and wild-type [56]. Finally, MLKL has been shown to play a role in atherogenesis; MLKL facilitates lipid handling in macrophages, and upon inhibition, the size of the necrotic core in the plaque is reduced [60]. Of the eight broad disease classes covered in this review, the role of MLKL in metabolic disease is arguably the most disputed, owing to the long-term nature of the experiments and the propensity for confounding variables, including genetic background and inter-facility variation in microbiome composition. The field may benefit from a more standardised approach to metabolic challenge and the prioritization of data generated using congenic littermate controls.

Neuromuscular
Evidence is rapidly accumulating for the role of MLKL in mouse models of neurological disease. Mlkl −/− mice were reported to be protected in one model of chemically-induced Parkinson's disease, with a significantly attenuated neurotoxic inflammatory response contributing to higher dopamine levels [69]. Strikingly, recent evidence suggests that MLKL may be important for tissue regeneration following acute neuromuscular injury [67,70]. In a model of cardiotoxin-induced muscle injury, muscle regeneration is driven by necroptotic muscle fibres releasing factors into the muscle stem cell microenvironment [70]. Mlkl −/− mice accumulate massive death-resistant myofibrils at the injury site [70]. Furthermore, in a model of sciatic nerve injury, MLKL was reported as highly expressed by myelin sheath cells to promote breakdown and subsequent nerve regeneration [67]. Overexpression of MLKL in this model is found to accelerate nerve regeneration [67] speaking to the potential of MLKL enhancing rather than blocking drugs in this area. However, contraindicating the use of MLKL activating drugs to mitigate neurological disease is the observation that MLKL accelerates demyelination in a necroptosis-independent fashion and thereby worsens multiple sclerosis pathology [68]. Finally, the role of necroptosis in murine amyotrophic lateral sclerosis remains contentious in the field. Wang et al. (2020) reported that MLKL-dependent necroptosis appears dispensable in the onset, progression, and survival of SOD1 G93A mice. Yet, there have been robust studies suggesting RIPK1-RIPK3-MLKL drives axonal pathology in both SOD1 G93A and Optn −/− mice [85,86]. While opinions in the field remain split on the relative contribution of MLKL in hematological (which express high levels of MLKL) vs. non-hematological (i.e., neurons express low levels of MLKL at baseline) cells in many of these models, disorders of the neuromuscular system have clearly come to the fore in commercial necroptosis drug development efforts.

Hematological
Mlkl −/− mice are hematologically indistinguishable from wild-type at steady state [5,79,87]. This trend continues as the mice age to 100 days [65]. Properly regulated necroptosis, however, is indispensable for hematological homeostasis. Mlkl D139V/D139V mice (which encode a constitutively active form of MLKL that functions independently of upstream activation from RIPK3) have severe deficits in platelet, lymphocyte, and hematopoietic stem cell counts [40]. Mice expressing even one copy of this Mlkl D139V allele are unable to effectively reconstitute the hematopoietic system following sub-lethal irradiation or in competitive reconstitution studies [40]. Like Ripk3 −/− mice, Mlkl −/− mice display a prolonged bleeding time and thus unstable thrombus formation [72,78]. MLKL, however, appears to play an additional, RIPK3-independent, role in platelet formation/clearance. In a model of lymphoproliferative disorder, Casp8 −/− Mlkl −/− double knockout mice develop a severe thrombocytopaenia that worsens with age (measured at 50 and 100 days) [65]. This phenotype is not observed in age-matched Casp8 −/− Ripk3 −/− mice [65]. Finally, MLKL has also been shown to function in neutrophil NET-formation at the cellular level [42], and Mlkl −/− mice are seen to be protected from diseases that implicate NET-formation, for example ANCA associated vasculitis [27] and deep vein thrombosis [71].

Cancer and Cancer Treatment
Cisplatin is a common chemotherapy that treats solid cancers, although its utility is limited by its nephrotoxic effects [23]. Strikingly, Mlkl −/− mice are reported to be largely resistant to cisplatin-induced tubular necrosis compared to wild-type controls [23]. However, mechanistically it is still not understood how systemic delivery of a DNAdamaging agent such as cisplatin could induce renal necroptosis and whether it is particular to kidney tissue. Hematopoietic stem cells derived from Mlkl −/− mice play a key role in studies showing that apoptosis-resistant acute myeloid leukemia (AML) could be forced to die via necroptosis [61]. By adding the caspase inhibitor IDN-6556/emricasan, AML cells are sensitised to undergo necroptosis in response to the known clinical inducer of apoptosis SMAC mimetic, birinapant [61]. IDN-6556/emricasan is well-tolerated in humans [88] and is an excellent example of a therapeutic approach designed to enhance rather than block MLKL activity in vivo. Finally, the role of MLKL in intestinal tumorigenesis remains unclear, with studies reporting that MLKL is dispensable in both sporadic intestinal or colitis-associated cancer [35], and yet MLKL has been reported to have a protective effect in Apc min/+ mice [62] by suppressing IL-6/JAK2/STAT3 signals [63].

Reproductive System
There are two studies that assess the role of MLKL in age-induced male infertility. Mlkl −/− mice aged to 15 months demonstrate significantly reduced body, testicular and seminal vesicle weight alongside increased testosterone levels and fertility [73] when compared to age matched wild-type mice. A recent follow up study suggests that CSNK1G2, a member of the casein kinase family, is co-expressed in the testes and inhibits necroptosismediated aging [89]. This phenomenon is also seen in human testes [89]. Another study that uses congenic littermate controls, however, finds that male mice aged to 18 months are indistinguishable from wild-type littermates with regards to total body, testicular and seminal vesicle weight [74].

Important Experimental Determinants in MLKL-Related Mouse Research
There are several high profile examples of genetic drift [90], passenger mutations [90], facility-dependent variation in the microbiomes of mice [91], sex [92] and age [93] acting as important modifiers of the innate immune response. One notable example is the significant differences between commonly used wild-type control C567BL/6J and C57BL/6NJ (also known as C57BL/6N) sub-strains (separated by 60+ years of independent breeding) in morbidity and survival following LPS-and TNFα-induced lethal shock [20,90]. Of similar interest, male mice are reported to be more susceptible than females to invasive pneumonia and sepsis [94]. Generously powered cohorts of sex-segregated, co-housed, congenic littermates (mice derived from a heterozygous cross) are the gold standard for controlling these confounding variables when comparing wild-type and Mlkl KO/mutant mice (or any other mutant) [91]. All scientists that work with mice will attest that while this approach is certainly time, resource, and mouse-number intensive, it should nonetheless be prioritised by experimenters and peer reviewers alike wherever possible. We congratulate the efforts of scientists who proceed one step further by restoring the wild-type phenotype through ectopic expression of MLKL [67]. When the use of littermate controls is not feasible, clear, and detailed descriptions of mouse age, sex and provenance provided in publications act as important caveats for discerning reviewers and readers to consider. In Table 1, we summarize the outcomes of these comparisons in 80+ contexts and have included a column that provides details of caveats where available.

Concluding Remarks
Genetic and experimental models of disease provide a strong rationale that MLKL and necroptosis are important mediators and modifiers of infectious and non-infectious disease. The study of MLKL in mouse models has indicated a bidirectional role of necroptosis in disease. Inhibiting the function of MLKL may confer protection in diseases characterised by hematopoietic dysfunction and uncontrolled inflammation borne of (and/or perpetuating) the failure of epithelial barriers throughout the body. Enhancing MLKL-induced cell death may prove beneficial in the treatment of malignancies and nerve injury. The therapeutic potential of MLKL as a druggable target in infectious disease is highly nuanced and will require careful tailoring to the pathogen and infection site/stage in question. Importantly, any extrapolation of these observations in Mlkl knockout and mutant mice to human disease must be tempered by our knowledge of key differences in both the structure and regulation of mouse and human MLKL [95][96][97]. This point is particularly poignant considering withdrawals of certain RIPK1 inhibitor drugs from phase I (pancreatic cancer) and II (chronic inflammatory diseases) clinical trials due to lack of efficacy [98,99]. Rapid advances in the 'humanisation' of mouse models of disease and the use of large human clinico-genetic databases will further enrich the extensive body of murine data supporting the role of MLKL in human disease.