To study how necroptosis results in the release of pro-inflammatory mediators in vitro, we turned to BLaER1 monocytes as a model system. In the absence of caspase-8, these cells display a strong necroptotic phenotype following TLR4 stimulation (13). To this end, we employed CASP8 ^-/- cells as necroptosis-prone cells and CASP8 ^-/- x MLKL ^-/- cells as necroptosis resistant controls in a CASP4 ^-/- genetic background, to prevent possible non-canonical inflammasome activation upon LPS stimulation ( Figure S1A ). As a proxy for lytic cell death, we measured LDH release, normalizing the values of a given treatment to an untreated control (0%) and a lysis control (100%). This normalization explains that for some conditions LDH release values below 0% were obtained. CASP8 ^-/- cells, but not CASP8 ^-/- x MLKL ^-/- cells, displayed robust necroptosis as evidenced by measuring LDH release in the supernatant ( Figures 1A , S1B ) or when assessing the phosphorylation status of MLKL ( Figure 1B ). To address whether necroptotic cells can trigger inflammatory responses in bystander cells, we devised a system in which we could co-culture cells undergoing necroptosis (donor cells) with cells that were insensitive to necroptosis induction (recipient cells). In this setting, we measured interleukin-6 (IL-6) production in recipient cells as a proxy for pro-inflammatory gene expression ( Figure 1C ). IL-6 was chosen for its high sensitivity and dynamic range as an NF-κB target gene. The IL6 gene was deleted in the donor cell population, so that all IL-6 measured in this setup can be ascribed to recipient cells. Apart from its intended phenotype, IL-6 deficiency had no impact on necroptosis or pro-inflammatory gene expression in donor cells ( Figures S1C–E ). Further, since LPS is used in this model to trigger necroptosis in donor cells, we rendered recipient cells insensitive to LPS by disrupting the TLR4 gene ( Figure 1C ). Donor and recipient cells were then co-cultured and analyzed for LDH release, as well as IL-6 production. As expected, co-cultures containing CASP8 ^-/- donor cells that were additionally deficient for IL-6 displayed a marked release of LDH into the supernatant upon LPS treatment ( Figure 1D , left panel). This response was not seen when CASP8 ^-/- x MLKL ^-/- donor cells were used. Intriguingly, when co-cultured with necroptosing donor cells but not CASP8 ^-/- x MLKL ^-/- cells, a strong IL-6 response was seen in recipient cells ( Figure 1D , right panel). To explore whether this pro-inflammatory activity of necroptosis required cell-to-cell contact, we next conducted experiments, in which we transferred supernatant from necroptosing donor cells or control conditions onto recipient cells. Analogous to the co-culture conditions, a potent IL-6 response was seen when supernatant from stimulated CASP8 ^-/- cells was transferred ( Figure 1E ). In line with these results, transfer of supernatant from necroptosing donor cells led to NF-κB activation in recipient cells as evidenced by p65 and IκBα phosphorylation, as well as IκBα degradation ( Figure 1F ). Altogether, these results indicated that necroptosis-driven inflammatory responses can be recapitulated in an in vitro system, and that cell-to-cell contact is not required to evoke pro-inflammatory activity.

Our results indicated that the necroptosis-driven inflammatory response in bystander cells is largely mediated by a soluble factor. Further, considering the experimental setup that made use of TRL4 ^-/- recipient cells, we could exclude a TLR4-agonistic molecule as the mediator of this response. Prolonged incubation at 75°C resulted in a complete loss of the pro-inflammatory activity of the necroptotic supernatant, indicating that a proteinaceous component might play a critical role in its activity ( Figure S2A ). Since triggering necroptosis in myeloid cells leads to secondary NLRP3 inflammasome activation (28, 29), we asked whether IL-1β would be responsible for IL-6 secretion in recipient cells. To address this, we induced necroptosis in donor cells in the absence or presence of the NLRP3 inhibitor MCC950 (30). As expected, necroptotic cells secreted IL-1β, which was readily blocked by MCC950 (Figure S2B). However, blocking IL-1β secretion did not impact on IL-6 secretion by TRL4 ^-/- recipient cells, excluding a role of IL-1β as an alarmin in our model system ( Figure S2C ). To further explore the possibility that a proteinaceous component might mediate the pro-inflammatory response in recipient cells, we separated the supernatant of necroptotic cells using size exclusion chromatography (SEC). To obtain a large volume of necroptotic supernatant, we switched to THP-1 cells as donor cells. We engineered these cells to undergo necroptosis upon expression of a dominant active MLKL construct (MLKL^1-201) that is sufficient to induce cell death ( Figure S2D ) (31). Such engineered cells underwent necroptosis as expected and upon transfer, the supernatant of these cells triggered IL-6 production in recipient cells ( Figures S2D, E ). We then subjected supernatant of control or necroptosing THP-1 cells to SEC and tested the individual fractions for pro-inflammatory activity on recipient cells ( Figures 2A–E ). These experiments revealed that the IL-6 inducing activity eluted at around 15 - 17 ml of the column volume, which corresponds to proteins of approximately 20-70 kDa in size. As a complementary approach, we also generated recipient cells, in which we deleted essential signaling hubs for various PRR systems. To block TLRs and cytokines of the IL-1 family, we deleted MYD88 and TICAM1 (TRIF), to prevent RLR and cGAS-dependent activation, we deleted MAVS and STING1 (STING), and to block TNF and Lymphotoxin-α (LTA)-dependent pro-inflammatory gene expression we deleted TNFRSF1A (TNFR1) ( Figures 2F , S2F ). Transferring necroptotic supernatant onto these cells in the presence of the TLR4 inhibitor CLI095 to inhibit LPS-driven effects, revealed that TNFRSF1A^-/- cells showed a markedly decreased IL-6 response, whereas the perturbation of the other signaling hubs had no impact ( Figure 2G ). MyD88/TRIF and MAVS/STING deficient cells used in this assay were produced in a wildtype (WT) background, while TLR4 and TNFR1 deficient cells were generated from caspase-4-deficient cells. As such, we additionally conducted the supernatant transfer experiment to compare WT and CASP4 ^-/- cells in their response to the necroptotic supernatant. Caspase-4 deficiency did not account for a different behavior of these recipient cells in our supernatant transfer assay ( Figure S2G ). To elucidate whether this TNFR1-dependent activity required either TNF, LTA, or both, we generated donor cells, in which these cytokines were deleted ( Figure 2H ). None of these perturbations impacted on necroptosis induction ( Figure 2H , left panel). However, TNF-deficient donor cells failed to trigger IL-6 production in recipient cells, whereas LTA deficiency had no impact ( Figure 2H , middle and right panel). Since these results pointed to TNF as the main driver of pro-inflammatory gene expression in bystander cells, we next assessed TNF production of the donor cell population as a function of necroptosis induction. These experiments revealed that TNF secretion was markedly increased in cells that underwent necroptosis, with peak levels being reached already 4 hours after stimulation ( Figure 2I ). Importantly, TNF release by LPS-stimulated CASP8^-/- cells could be reduced by treatment with the RIPK3 inhibitor GSK’872, ruling out that clonal differences may account for the divergence in TNF secretion between CASP8^-/- and CASP8^-/- x MLKL^-/- cells ( Figure S2H ). In summary, these results indicated that necroptosis-driven induction of pro-inflammatory gene expression is largely driven by TNF in this model. Moreover, these data suggested that necroptosis strongly enhanced TNF secretion.

As a key pro-inflammatory mediator, TNF is controlled at multiple levels. Next to being regulated at the level of transcription, the translation of TNF mRNA is heavily controlled at the post-transcriptional level. To this end, several cis-elements within the 3’ UTR of TNF have been identified that are regulated by various trans-activating factors that repress TNF mRNA translation or induce decay of its mRNA (32–36). Moreover, the availability of the protein itself in the extracellular milieu is also regulated. As such, being produced as a type II transmembrane protein, TNF needs to be cleaved by membrane metalloproteases to be released. This step is mainly regulated by ADAM17 and to a lesser extent by ADAM10 (37–40). Comparing CASP8 ^-/- cells with CASP8 ^-/- x MLKL ^-/- revealed no significant difference in TNF mRNA expression between these two cell types ( Figure 3A ), thereby excluding the possibility of enhanced TNF mRNA transcription. However, when we measured TNF protein expression, we observed both higher amounts of pro-TNF levels in cell lysates as well as increased levels of cleaved TNF in the supernatant, suggesting that regulation of TNF expression and shedding may both contribute to enhanced TNF release ( Figure 3B ). Of note, CASP8 ^-/- and CASP8 ^-/- x MLKL ^-/- cells produced similar amounts of IL-6 and IL-8, showing that other pro-inflammatory cytokines transactivated by NF-κB were not affected by these genotypes. As such, both IL-6, as well as IL-8 levels did not show any differences between CASP8 ^-/- and CASP8 ^-/- x MLKL ^-/- cells ( Figures S1C, D ). Speculating that necroptosis impacted on the post-transcriptional regulation of TNF mRNA translation via its 3’ UTR, we generated CASP8 ^-/- and CASP8 ^-/- x MLKL ^-/- cells, in which we deleted the 3´UTR of the TNF gene (TNF Δ 3´UTR) ( Figure 3C ). Doing so, we eliminated all critical cis-controlling elements that negatively affect the stability or translation of TNF mRNA, while we left the poly adenylation signal intact. As a result, pro-TNF expression levels were dramatically increased in both CASP8 ^-/- and CASP8 ^-/- x MLKL ^-/- cells and considerable amounts of pro-TNF expression were already observed in lysates of unstimulated cells ( Figure 3D ). Importantly, removing the 3’ UTR resulted in similar expression levels of pro-TNF for both necroptotic and non-necroptotic cells ( Figure 3D ), indicating that necroptosis indeed impacts on the negative regulation of TNF mRNA translation through its 3’ UTR. In this setting, we could now investigate the impact of necroptosis on TNF shedding. Indeed, despite similar expression of pro-TNF, CASP8 ^-/- cells still exhibited 3-4x higher levels of soluble TNF in the supernatant upon necroptosis induction ( Figures 3D-F ). Altogether, these results suggested that increased secretion of mature TNF during necroptosis depends on higher pro-TNF expression, by alleviating negative regulatory pathways at the post-transcriptional level, as well as enhanced TNF shedding.

To explore how necroptosis impacts on TNF shedding, we used 293T cells as a heterologous expression system, in which we could dissociate TNF expression from shedding. To induce necroptosis, we expressed the autoactive MLKL^1-201 construct, while using an inactive MLKL variant (MLKL^1-154) or mCherry as a control (31). Of note, this reductionist setting could also rule out RIPK3-dependent effects on necroptosis-dependent TNF secretion. MLKL^1-201-dependent necroptosis induction led to a marked release of LDH from 293T cells, while the other conditions showed little activity ( Figure 4A , left panel). Concomitant with cell death, a strong TNF signal in the supernatant was detected by ELISA ( Figure 4A , right panel). This increase in TNF within the supernatant was due to a markedly enhanced pro-TNF conversion to soluble TNF, as evidenced by immunoblot. Importantly, the release of IL-6, which does not require shedding for its maturation, was not enhanced upon necroptosis induction in these settings ( Figure S3A ). To confirm that necroptosis indeed resulted in increased TNF shedding and not in a passive release of pro-TNF, we additionally conducted experiments in 293T cells lacking ADAM10 and ADAM17 (41). In these cells, TNF release upon necroptosis was severely impaired, while ADAM10/17 deficiency had no impact on cell death following MLKL^1-201 expression ( Figure 4B ). Analogous results were obtained when studying TNF levels in lysates or supernatant by immunoblot ( Figure 4B , lower right panel). Next, we sought to investigate the role of ADAM proteases in the BLaER1 cell model. RNA-seq analysis revealed that ADAM10, ADAM17 and ADAM9 are the main ADAM proteases expressed in these cells and that LPS positively modulates their expression ( Figure S3B ). With ADAM17 and ADAM10 being the main TNF sheddases described, we generated BLaER1 cells lacking either of these two enzymes ( Figure S3C ) to address their role in necroptosis-dependent TNF shedding. Doing so revealed that ADAM17 was the predominant enzyme required for TNF shedding in the context of necroptosis, while neither ADAM17 nor ADAM10 had a significant impact on cell death induction ( Figure 4C , left and middle panel). Consistent with this notion, supernatant from ADAM17 ^-/- but not control or ADAM10 ^-/- cells displayed a markedly reduced activity in the bystander cell activation assay ( Figure 4C , right panel). Additionally, the secretion of another known target of ADAM17, IL-6Rα (42), increased in necroptotic BLaER1 cells compared to control cells ( Figure S3D ). Altogether, these results indicated that necroptosis drives enhanced TNF-dependent inflammatory responses in two ways: on the one hand, necroptosis results in enhanced TNF translation, and on the other hand, necroptosis triggers TNF shedding in a predominantly ADAM17-dependent manner.

To explore the spatiotemporal relationship of necroptosis and TNF shedding, we made use of a TNF shedding reporter construct (C-tag TNF) that allows the monitoring of TNF maturation at the ADAM17 cleavage site in real time (41). Upon cleavage, this construct exposes a neo C-terminus that serves as an epitope for a nanobody that only binds the cleaved version of this reporter. This construct was additionally equipped with an N-terminal mCherry tag, to monitor the subcellular localization of TNF. Cells expressing the C-tag reporter displayed the same gain in TNF shedding upon necroptosis induction as cells expressing unmodified TNF ( Figure S4A ). Using a fluorescently labeled nanobody, we monitored the C-tag signal as a function of MLKL^1-201-induced necroptosis in control cells, as well as in ADAM10 ^-/- x ADAM17 ^-/- deficient cells. In addition, we also measured C-tag positivity in controls not undergoing necroptosis. 4.5 hours following necroptosis induction, we observed a strong signal for the C-tag nanobody at the plasma membrane of cells that also showed signs of necroptosis ( Figures 5A, B , and Supplementary Video S1 ). This signal was completely abrogated in cells lacking ADAM10/17, which still displayed membrane-associated expression of TNF ( Figures 5A, B , and Supplementary Video S2 ). Control cells not undergoing necroptosis, also showed a C-tag reactive signal, yet with marked lower intensity ( Figures 5A, B , and Supplementary Video S3 ). Time-lapse analysis revealed that cells undergoing necroptosis became positive for the C-tag signal within 30-40 minutes ( Figures 5B, C ). Quantifying the C-tag signal and analyzing the slope for several cells individually revealed that cells undergoing necroptosis indeed displayed a rapid, switch-like gain in TNF shedding. On the other hand, the gain in C-tag signal for cells not undergoing necroptosis was only gradual, and it never reached the same level as necroptosing cells in the time frame studied ( Figures 5B, D ). Similar to apoptosis, necroptosis also results in the externalization of phosphatidylserine phospholipids to the outer leaflet of the membrane (43). Measuring Annexin V positivity next to TNF shedding, we observed that these two events were closely correlated, with the individual slopes of the gain in fluorescent signal being comparable ( Figures S4B–D ). However, C-tag positivity preceded Annexin V positivity by a mean value of 24 minutes ( Figure S4E ). Studying PI positivity as a proxy for membrane permeabilization, we observed that cells undergoing necroptosis and TNF shedding displayed a delayed gain in PI signal ( Figures S4F-I ). In comparison to Annexin V, this delay was even more pronounced with a mean difference of 53 minutes ( Figure S4I ). Further, analyzing the slope of the gain in PI positivity revealed that this event occurred more rapidly than the TNF shedding event ( Figures S4G, H ). This was also reflected in the slope of the PI signal which was approximately twice the value of the C-tag signal ( Figure S4H ). Altogether, these results indicated that necroptosis-induced TNF shedding is a rapid event that differs from the steady state shedding process in magnitude and kinetics. As such, it displays a switch-like behavior, in which TNF shedding occurs in a time frame of 30-40 minutes. In the course of cell death, TNF shedding precedes the outward flipping of phosphatidylserine phospholipids, which is followed by membrane permeabilization.

Our data thus far indicated that necroptosis triggers both TNF translation and shedding and that the latter takes place in the early stages of cell death, before massive membrane damage. Live cell imaging experiments revealed that TNF shedding occurs as a switch-like phenomenon and appears to be strictly related to physical changes of the cell membrane before membrane permeabilization. We therefore asked whether the mechanism behind increased TNF shedding is specifically related to the necroptotic pathway or if it is promoted by changes in membrane integrity during lytic cell death in general. To address this question, we applied a model in which we can induce lytic cell death by adding an exogenous pore-forming molecule, namely the pore-forming toxin streptolysin O (SLO) from S. pyogenes (44, 45). As for the MLKL gain-of-function studies, we used HEK 293T cells overexpressing pro-TNF that we then challenged with exogenous SLO. As expected, SLO treatment led to lytic cell death, as indicated by LDH release ( Figure 6A , left panel). Similar to the overexpression of MLKL^1-201, SLO-mediated cell death was accompanied by massive TNF secretion due to enhanced shedding, as evidenced by ELISA and immunoblot ( Figure 6A , right panel and Figure 6B ). Importantly, IL-6 secretion was not enhanced upon SLO-mediated cell death, but rather decreased due to cell death during the experiment ( Figure 6C ). In conclusion, these data indicate that enhanced TNF shedding is linked to the membrane dynamics upon lytic cell death rather than being a specific feature of the necroptotic pathway itself.

Thus far our experiments had relied on cell lines to study the relevance and mechanism of necroptosis-induced TNF shedding. To explore whether these events also play a role in primary cells undergoing necroptosis, we turned to primary, human monocyte-derived macrophages (hMDM) as a model. Treating these cells with LPS and the pan-caspase inhibitor zVAD, resulted in a lytic cell death in approximately 25% of the cell population within an 8-hour timeframe ( Figure 7A , left panel). Treatment with the specific RIPK3 inhibitor GSK’872 blocked this cell lysis, indicating that cells died by necroptosis. Analyzing TNF secretion revealed that necroptosis induction doubled the amount of released TNF, while it decreased IL-6 production by half ( Figure 7A , middle and right panel). To employ a system that does not require LPS stimulation to induce necroptosis, we used murine macrophages derived from hematopoietic stem cells (HSCs). As such, we transduced HSCs with a murine RIPK3 construct that can be dimerized to induce necroptosis upon the addition of the small molecule drug AP20187 ( Figure 7B ) (46). After differentiation of these HSCs, these cells displayed a typical macrophage-like morphology and about 30% of them expressed the RIPK3 dimerizing construct, as judged by EGFP expression ( Figures S5A, B ). Indeed, RIPK3 dimerization by treatment with AP20187 induced lytic cell death specifically in transduced cells. Moreover, this synthetic necroptosis-induction resulted in TNF release of these macrophages in the supernatant ( Figure 7C ). In summary, these results suggested that primary cells undergoing necroptosis also display enhanced TNF secretion.

BLaER1 (60) and THP-1 (DSMZ, ACC 16) cells were cultured in RPMI medium 1640, supplemented with 10% (v/v) FCS, 1% sodium pyruvate (v/v), 100 U/ml Penicillin-Streptomycin (all Gibco). BLaER1 cells were trans-differentiated for 5-6 day in medium containing 10 ng/ml IL-3, 10 ng/ml M-CSF (MPI of Biochemistry, Munich) and 100 nM β-estradiol (Sigma-Aldrich) at a concentration of 7 x 10⁴ cells/well in 96 well plates. Unless otherwise stated, CASP4 ^-/- BLaER1 cells were used as genetic background to generate subsequent knock-out clones analyzed in this work. 293T cells (DSMZ, ACC 635) were cultured in DMEM supplemented as above.

The following reagents were purchased from commercial suppliers: Alexa Fluor^® 647 Annexin V (Biolegend, Cat#640911), B/B homodimerizer (Takara Bio, Cat#635058), b-estradiol (Sigma-Aldrich, Cat#E8875-250MG), CLI-095 (In vivoGen, Cat# tlrl-cli95), Collagen R solution 0.2% (Serva, Cat#47254.02), DNAse I (Thermo Fisher Scientific, Cat#EN0525), Doxycycline hyclate (Sigma-Aldrich, Cat#D9891), eBioscience™ Propidium Iodide Staining Solution (Thermo Fisher Scientific, Cat# 00-6990-50), GeneJuice^®Transfection Reagent (Merck Chemicals GmbH, Cat#70967), GeneRuler 1 kb DNA Ladder (Thermo Scientific, Cat# SM0311), GSK’872 (Aobious, Cat#AOB4887), Hepes 1M (Sigma-Aldrich, Cat#H0887-100ML), HT-DNA (Sigma-Aldrich, Cat#D6898), Lipofectamine 2000 (Life Technologies, Cat#11668019), LPS-EB Ultrapure (In vivoGen, Cat#tlrl-3pelps), MCC950 (Sigma-Aldrich, Cat#5381200001), MEM Non-Essential Amino Acids Solution (100x) (Thermo Fisher Scientific, Cat#11140050), PageRuler™ Prestained Protein Ladder (Thermo Fisher Scientific, Cat#26617), Penicillin-Streptomycin (Life Technologies, Cat#15140122), Phorbol 12-myristate 13-acetate (Enzo Life Sciences, Cat#BML-PE160-0005), R848 (In vivoGen, Cat# tlrl-r848), Recombinant Human TNF (Peprotech, Cat#300-01A), RetroNectin^® Recombinant Human Fibronectin Fragment (Takara Bio, Cat#T100A), Recombinant Murine Flt3-Ligand (Peprotech, Cat#250-31L), Recombinant Murine IL-3 (Peprotech, Cat#213-13), Recombinant Murine IL-6 (Peprotech, Cat#216-16), Recombinant Murine SCF (Peprotech, Cat#250-03), Recombinant Murine TPO (Peprotech, Cat#315-14), Retronectin (Takara Bio Cat#T100A), Sodium Pyruvate (Thermo Fisher Scientific, Cat#11360088), Streptolysin O (Sigma-Aldrich, Cat# S5265-25KU), Z-VAD-FMK (Peptanova, Cat#3188-v).

The following antibodies were purchased from commercial suppliers: Alexa-488-conjugated secondary antibody (Thermo Fisher Scientific, Cat# A-11001, RRID : AB_2534069 and Cat# A-11008, RRID : AB_143165), anti-b-actin (C4) HRP (Santa Cruz Biotechnology, Cat# sc-47778 HRP, RRID : AB_271418), anti-Caspase-8 (1C12) (Cell Signaling Technology, Cat# 9746, RRID : AB_2275120), anti-ADAM10, CT (Millipore, Cat# AB19026, RRID : AB_2242320), anti-ADAM17 (D22H4) (Cell Signaling Technology, Cat# 6978, RRID : AB_1082838), anti-IκBα (L35A5) (Cell Signaling Technology, Cat# 4814, RRID : AB_390781), anti-phospho-IkBa (14D4) (Cell Signaling Technology, Cat# 2859, RRID : AB_561111), anti-TNF (D5G9) (Cell Signaling Technology, Cat# 6945, RRID : AB_10859375), CaptureSelect™ Alexa488 anti-C-tag conjugate (Thermo Fisher Scientific, Cat#7213252100), MLKL polyclonal antibody (Thermo Fisher Scientific, Cat# PA5-34733, RRID : AB_2552085), phospho-NF-κB p65 (Ser536) (93H1) rabbit mAb (Cell Signaling Technology, Cat# 3033, RRID : AB_331284), recombinant anti-MLKL (phospho S358) antibody (Abcam, Cat# ab187091, RRID : AB_2619685).

PBMCs were obtained from healthy volunteers. Informed consent was obtained from all subjects according to the Declaration of Helsinki and approval by the responsible ethical committee (project number 19-238, Ethics committee of the medical faculty of the Ludwig Maximilian University Munich). Monocytes were isolated using the CD14 positive selection method of MACS technology (Miltenyi Biotec, Cat#130-050-201) according to manufacturer’s protocol. Purified monocytes were plated in 96 well plates (0.1 x 10⁶ cells/well) and differentiated into monocytes-derived macrophages (MDMs) for 7-8 days in RPMI medium 1640 supplemented with 10% (v/v) FCS, 1% (v/v) sodium pyruvate, 1% (v/v) non-essential amino acids (Gibco), 100 U/ml Penicillin-Streptomycin and 100 ng/ml M-CSF (Peprotech).

Cloning of genes of interest into pLI, pRP, pFUGW and pEF-BOS backbones was performed by conventional restriction enzyme cloning. hTNF, MLKL^1-154 and MLKL^1-201 were amplified from cDNA of BLaER1 cells. The codon-optimized sequence of hIL6 was ordered from IDT. pMDLg/pRRE was a gift from Didier Trono (Addgene plasmid #12251; http://n2t.net/addgene:12251; RRID : Addgene_12251), pRSV-Rev was a gift from Didier Trono (Addgene plasmid #12253; http://n2t.net/addgene:12253; RRID : Addgene_12253), pLIX_403 (herin referred to as pLI) was a gift from David Root (Addgene plasmid #41395; http://n2t.net/addgene:41395; RRID : Addgene_41395).

Unless otherwise indicated, CASP8 ^-/-, CASP8 ^-/- x MLKL ^-/- and subsequently derived gene-deficient BLaER1 macrophages were stimulated with 2 ng/ml LPS from E. coli (In vivogen) for 18 h. In case of supernatant transfer experiment, the supernatant was then collected, centrifuged at 600 g for 5 minutes and used undiluted to stimulate TLR4 ^-/- recipient BLaER1 cells for 24 h. When recipient cells were not in a TLR4 ^-/- background, the TLR4 inhibitor CLI095 was added to the collected supernatant at the final concentration of 1 µg/ml. For co-culture experiments, CASP8 ^-/- or CASP8 ^-/- x MLKL ^-/- BLaER1 macrophages were mixed 1:1 with TLR4 ^-/- cells and stimulated for 18 h with 2 ng/ml of LPS. For heat shock experiments, the supernatant of donor cells was heated at 75°C for 10 minutes before stimulation of TLR4 ^-/- BLaER1 macrophages. FCS was replaced to a final concentration of 10% after heat shock treatment. The NLRP3 inhibitor MCC950 was used at the final concentration of 5 µM. Stimulation of BLaER1 macrophages with other immune ligands was performed as follows: hTNF (Peprotech) 100 ng/ml, R848 (In vivogen) 1 µg/ml for 16 h. 200 ng/well of IVT4 or herring testes DNA (HT-DNA, Sigma-Aldrich) were complexed with Lipofectamine 2000 (Life technologies) according to manufacturer’s protocol and cells were stimulated with the mix for 16 h. Primary MDMs were stimulated with 0.2 ng/ml LPS for 8 h. The Caspase-8 inhibitor Z-VAD-FMK (Peptanova) and the RIPK3 inhibitor GSK’872 (Aobious) were added to the cells together with LPS at the final concentration of 20 µM and 3 µM, respectively. For the size exclusion chromatography experiment, THP-1 cells previously transduced with pLI-MLKL^1-201 were differentiated with 100 ng/ml PMA (Sigma-Aldrich) for 24 h before stimulation. Differentiation medium was removed, and cells were washed in PBS before incubation in serum-free medium in presence or absence of doxycycline (1 µg/ml) for 16 h. Doxycycline was used to stimulate 293T cells at the final concentration of 1 µg/ml for 7 h on the day after transfection. In case an immunoblot had to be performed, 293T and BLaER1 cells were stimulated respectively in 1% FCS or 3% FCS containing medium. To induce necroptosis in HSCs-derived macrophages transduced with the dimerizing construct (mRIPK3-2xFv-T2A-EGFP), cells were stimulated with indicated concentrations of B/B homodimerizer (Takara Bio) for 15 h. Streptolysin O was used at the concentration of 8 µg/ml for 7 h to stimulate 293T cells.

Hematopoietic stem cells (HSCs) were obtained from 8-weeks old wild-type C57BL/6 mice. Bone marrow was flushed from femurs and tibias and HSCs were isolated by negative selection with the Direct Lineage Depletion Kit (Miltenyi Biotec, Cat#130-110-470) using MACS technology. Isolated HSCs were cultured in StemPro™-34 SFM (1X) media (Gibco) supplemented with StemPro^®-Nutrient Supplement, 100 U/ml penicillin-streptomycin and the following cytokines (all PeproTech): IL-3 (10 ng/ml), FLT3-L (10 ng/ml), SCF (100 ng/ml), TPO (20 ng/ml), IL-6 (20 ng/ml). The day after isolation, HSCs were transduced as described. 24 h later, HSCs were plated in 96 well plate at 0.7 x 10⁵/well and differentiated into macrophages in fully supplemented DMEM with 100 ng/ml M-CSF for 5 days. Medium with fresh M-CSF was added on day 2 or 3 after plating. All mice were handled according to institutional guidelines approved by the animal welfare and use committee of the government of Upper Bavaria.

BLaER1 and 293T cells deficient for specific genes were generated via CRISPR/Cas9 mediated gene ablation as previously described (61). Briefly, sgRNAs were designed to target early codons of selected genes. pLK0.1-gRNA-CMV-GFP and a Cas9 expression plasmid (pRZ-BFP-Cas9) were co-electroporated in BLaER1 cells using a Biorad GenePulser device or co-transfected in 293T cells in combination with GeneJuice transfection reagent (Merck). 24 h later, cells positive for respective fluorescent markers were sorted and subjected to limiting dilution cloning. After 3-4 weeks, monoclones were identified and subjected to deep sequencing (Illumina Miseq platform). Clones bearing frameshift mutations were utilized for subsequent experiments. For the generation of TNF Δ 3´ UTR clones, 2 gRNAs targeting the very beginning and the very end of the 3´UTR of the TNF gene were cloned into pMini-U6-gRNA-CMV-BFP-T2A-Cas9 or pMini-U6-gRNA-CMV-mCherry-T2A-Cas9 and co-electroporated in BLaER1 parental cells. Double fluorescent cells were sorted and subjected to limiting dilution cloning. Clones lacking the 3´ UTR were selected by PCR on the target region using the following primers: fwd 5´CCTGGTATGAGCCCATCTATCTG, rev 5´ TTCTTTTCTAAGCAAACTTTATTTCTCGCC (long product = WT, short product = Δ 3´UTR)

The following genomic regions were targeted to generate knockout cell lines (PAM sequence is depicted in bold letters): ADAM17 GAGCAGAACATGATCCGGATGG, ADAM10 TTTCAACCTACGAATGAAGAGG, LTA ACGTTCAGGTGGTGTCATGGGG, TNF TGCAGCAGGCAGAAGAGCGTGG, MLKL GAGCTCTCGCTGTTACTTCAGG, TNFRS1A GCAGTCCGTATCCTGCCCCGGG, TICAM1 (TRIF) GGCCCGCTTGTACCACCTGCTGG, MAVS ACTTCATTGCGGCACTGAGGGG, STING1 GCGGGCCGACCGCATTTGGGAGG, MYD88 CTGCAGGAGGTCCCGGCGCGGG, CASP4 CTCATCCGAATATGGAGGCTGG, IL6 TGTGGGGCGGCTACATCTTTGG, TLR4 GATAAAGTTCATAGGGTTCAGG, CASP8 GCTCAGGAACTTGAGGGAGG, TNF UTR (1) AGGGGGTAATAAAGGGATTGGGG, TNF UTR (2) TTACAGACACAACTCCCCTGGGG.

For the generation of pseudotyped lentiviral particles, 293T cells were transfected by calcium phosphate co-precipitation with the lentiviral plasmid expressing the desired construct (in a pLI backbone) and the following packaging plasmids: pMDLg/pRRE, pRSV-REV and pCMV-VSVG. 8-14 h after transfection, medium was changed, and cells were incubated for 48-72 h before virus-containing supernatant was harvested and used to transduce target cells. Puromycin selection was applied for 3-4 days before further experimental use. To transduce HSCs with the dimerizing construct (mRIPK3-2xFv-T2A-EGFP expressed in the lentiviral vector pFUGW), 0.5 x 10⁶ cells were plated in a 24-well pre-coated with retronectin (Takara Bio). Virus-containing medium was concentrated by ultracentrifugation and added to cells for 8 h. HSCs were then washed with PBS and transferred to a new 24-well with fresh culture medium. The percentage of transduced cells was evaluated by flow cytometry analysis of EGFP expression on the day of the experimental stimulation.

Cell pellets were lysed in cold RIPA buffer containing a cocktail of protease inhibitors (Roche, Cat# 11697498001). Quantification of proteins was conducted via the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Cat#23227). Cell lysates and cell supernatants were denatured in Laemmli Buffer for 10 minutes at 95°C, followed by separation by denaturing SDS-PAGE. Equal amounts of proteins from lysates (15-20 µg) or equal volumes of supernatant were loaded in each lane. In the case of BLaER1 supernatant, proteins were precipitated with methanol/chloroform before denaturation in Laemmli buffer. Proteins were blotted onto a 0.2 µm nitrocellulose membrane, blocked in 5% milk and incubated with indicated primary antibody overnight at 4°C. The day after, membranes were washed and incubated in respective HRP-coupled secondary antibodies for 1 h at room temperature. Upon need, membranes were re-probed with multiple antibodies. Chemiluminescent signal was detected via a CCD camera (Fusion Fx, Vilber).

293T cells were transfected with plasmid DNA complexed with GeneJuice following provider’s instructions. For experiments where both immunoblotting and ELISA had to be performed, 0.65 x 10⁶ cells of indicated genotypes were plated in each well of a 6 well plate and transfected the day after, unless otherwise indicated. Co-transfection with 2 plasmids was performed by using 1.5 µg of each plasmid/well. Cells were used for the experiment on the day after transfection. When only ELISA analysis had to be performed, 2 x 10⁴ cells/well were plated in 96-well format and transfected the day after with 200 ng/well of the needed plasmids (100 ng/plasmid in case of co-transfection with 2 plasmids or in case of transfection followed by streptolysin O stimulation). For live cell imaging experiments, 1.7 x 10⁴ 293T cells of the indicated genotypes were plated in collagen R (Serva) coated Ibidi µ-Plate 96 well black (Ibidi) using FluoroBrite DMEM medium (Life Technologies) supplemented as described, with the addition of 10 mM Hepes (Sigma-Aldrich). The day after, cells were transfected with a total of 100 ng DNA/well (50 ng of pLI-C-tag-TNF linker mCherry or pLI-C-tag-TNF + 50 ng pEF-BOS-nBFP, the latter for the expression of a nuclear BFP). 24 h after transfection, cells were stimulated with doxycycline as described.

Imaging was performed starting 4 h after doxycycline treatment to induce the expression of MLKL^1-201 and/or C-tag TNF/C-tag TNF mCherry. Before imaging, the CaptureSelect™ Alexa Fluor™488 anti-C-tag conjugate (Thermo Fisher Scientific) was added to the cell supernatant at a final dilution of 1:480. When necessary, propidium iodide (eBioscience) was added to a final dilution of 1:250 and Annexin V Alexa Fluor™ 647 (Biolegend) was used at 1:125 dilution. Imaging was performed over a period of time of 4h 30 on a Leica DMi8 inverted microscope equipped with a HC PL APO 63x/1.20 W CORR CS2 objective, with a picture taken every 6 minutes. For image analysis, we used FIJI (https://imagej.net/) to select regions of interests in individual cells (their motilities were negligible over the observation time) and extract the fluorescence intensity trajectories for the individual fluorescent channels. Different fields of view of at least 3 independent experiments were utilized for the analysis. The mean fluorescence intensities were then plotted as a function of time for single cells and analyzed using OriginPro2021 (OriginLab). In the case of induction of MLKL^1-201, an inflection point was visible in the fluorescence intensity trajectories corresponding to the Alexa488 channel (green lines in Figure 5B ). We computed the maximal slopes at the inflection point time as a simple measurement of expression kinetics. In the other cases (WT control and ADAM10 ^-/- x ADAM17 ^-/- MLKL^1-201 cells), we calculated an average slope by performing a linear regression of the whole fluorescence intensity trajectories.

THP1 cells were stimulated as described above in serum-free medium. Supernatant from doxycycline-induced or uninduced cells (40ml per condition) was collected and concentrated to 500 µL using Amicon Ultra-15 filters with 10 kDa cutoff (Merck Millipore). The concentrated samples were loaded on a Superose 6 10/300GL column (GE Healthcare) and an isocratic run (1.2 column volume) was conducted on an ÄKTA Basic instrument with a flowrate of 0.3 ml/min. Degassed and filtered PBS was used as running buffer. 500 µL fractions were collected and utilized for 14 h stimulation of TLR4 ^-/- BLaER1 recipient cells. Upon stimulation, fractions were diluted 1:1 with fully supplemented RPMI medium, adjusting the FCS content to 10% v/v.

Differentiated BLaER1 cells were left untreated or LPS stimulated for 2 h. Next, 3x10⁶ cells were lysed in 1ml Trizol and library preparation performed as described elsewhere (62). The libraries were sequenced on an Illumina HiSeq1500 device. Reads were aligned to the human reference genome GRCh38 using STAR v.2.5.1 (63) and counted using the GenomicAlignments R package (64). Post mapping analysis was conducted in R (65). Normalization and scaling were performed using DESeq2’s (66). GGplot2 (67) was used to visualize results.

RT-qPCR analysis of LPS stimulated BLaER1 cells of indicated genotypes was performed as follows. RNA from BLaER1 cells was purified via the Total RNA purification mini spin kit (Genaxxon, Cat#S5304.0250) and DNA digestion was performed with DNAse I (Thermo Fisher Scientific) incubation for 30 minutes. 700 ng of RNA was used for cDNA synthesis by RevertAid Reverse transcriptase (Thermo Fisher Scientific, Cat#EP0442) with oligo dT primers. qPCR reaction was performed with the PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific) in presence of primers specific for genes of interest: TNF_fwd: CAGGGACCTCTCTCTAATCAG, TNF_rev: TTGAGGGTTTGCTACAACATGG, TBP1_fwd: TATAATCCCAAGCGGTTTGC, TBP1_rev: CTGTTCTTCACTCTTGGCTCCT GUSB_fwd: CGCCCTGCCTATCTGTATTCATTGGAGGTG, GUSB_rev: GATGAGGAACTGGCTCTTGGTGACAGCC. The average of GUSB and TBP1 housekeeping genes was used to calculate ΔCt, followed by calculation of the ΔΔCt using the untreated sample of each clone as control.

hIL-6 (BD Biosciences, Cat#555220), hTNF (BD Biosciences, Cat#555212), hIL-8 ELISA (BD Biosciences, Cat#555244), hIL-1β ELISA (BD Biosciences, Cat#557953), hIP-10 (BD Biosciences, Cat#550926), mmTNF (BD Biosciences, Cat#558534) and hIL-6Rα ELISA (R&D Systems, Cat#DY227) were performed according to provider’s protocol.

Pierce LDH Cytotoxicity Assay Kit (Thermo Fisher Scientific, Cat#C20301) was performed according to manufacturer’s instructions. Relative LDH release was calculated as LDH release [%] = 100 * (measurement – unstimulated control)/(lysis control – unstimulated control). Because of the use of this equation, negative data can be obtained when LDH release is lower in a tested sample compared to the untreated control.

Statistical significance was determined as explained in the figure legends with the respective post hoc corrections for multiple testing, if relevant. If multiple comparisons are depicted with one comparison bar, the major tick of the comparison bar indicates the reference data to which the statements regarding the level of significance are made. Statistical analysis was performed with GraphPad Prism 8.