A RIPK1-specific PROTAC degrader achieves potent anti-tumor activity by enhancing immunogenic cell death

Abstract

Receptor-interacting serine/threonine-protein kinase 1 (RIPK1) functions as critical stress sentinel that coordinates cell survival, inflammation and immunogenic cell death (ICD). While the catalytic function of RIPK1 is required to trigger cell death, its non-catalytic scaffold function mediates strong pro-survival signaling. Accordingly, cancer cells can hijack RIPK1 to block necroptosis and evade immune detection. We generated a small molecule proteolysis-targeting chimera (PROTAC) that selectively degraded human and murine RIPK1. PROTAC-mediated depletion of RIPK1 deregulated TNFR1- and TLR3/4-signaling hubs, accentuating the output of NF-κB, MAPK and IFN signaling. Additionally, RIPK1 degradation simultaneously promoted RIPK3 activation and necroptosis induction. We further demonstrated that RIPK1 degradation enhanced the immunostimulatory effects of radio- and immunotherapy by sensitizing cancer cells to treatment-induced TNF and interferons. This promoted ICD, anti-tumor immunity and durable treatment responses. Consequently, targeting RIPK1 by PROTACs emerges as a promising approach to overcome radio- or immunotherapy resistance and enhance anti-cancer therapies.

eTOC blurb

Tumour cells frequently hijack RIPK1’s scaffolding function to resist cell death and avoid immune detection. Mannion et al. describe the development of a RIPK1-selective PROTAC degrader that boosts the immunostimulatory and anti-tumour activity of radiotherapy (RT) and immune checkpoint blockade (ICB), heating up tumours and driving long-lasting anti-tumour immunity.

Graphical Abstract

INTRODUCTION

Cell death and inflammation are closely linked arms of the innate immune response to combat infection and drive anti-tumor responses. While the ability to sense danger is integral for any organism to maintain tissue function¹, uncontrolled or excessive inflammatory responses can damage tissue and contribute to the pathogenesis of chronic inflammatory diseases, including psoriasis, neurodegenerative diseases, inflammatory bowel disease and asthma². Further, tumor cells frequently evade immune detection and elimination by suppressing innate immune responses³. Cell death surpasses merely being an endpoint; it facilitates the exchange of information between the dying cell and the surrounding tissue micro-environment. Importantly, alerting and recruiting immune cells to sites of disturbance. Receptor-interacting serine/threonine-protein kinase 1 (RIPK1) is a critical stress sentinel that can positively and negatively regulate cell death, innate immunity, and inflammation^4,5. Located downstream of many cytokine receptors and pattern recognition receptors (PRRs), RIPK1 plays a crucial role in determining the outcome between pro-survival NF-κB signaling and cell death^6–9. While the catalytic function of RIPK1 is required to trigger cell death, its non-catalytic scaffolding function is essential to regulate pro-inflammatory and pro-survival signaling^10–16. Accordingly, controlled activation of RIPK1 contributes to tissue repair and immune surveillance. However, over-exuberant activation of RIPK1 can lead to many immune and autoinflammatory diseases^17–22. Thus, RIPK1 deregulation and subsequent aberrant cell death is a common feature in chronic inflammatory disease^9,23,24. As such, RIPK1 has emerged as a promising therapeutic target²⁴.

While chronic RIPK1 activation is associated with inflammatory disease^9,23,24, data demonstrate that cancer cells often hijack RIPK1, promoting cell survival, insensitivity to TNF and resistance to immunotherapy²⁵. This is because RIPK1’s scaffolding function drives NF-κB signaling and NF-κB-dependent induction of pro-survival genes^26,27. Further, chronic RIPK1-mediated NF-κB signaling can fuel an immunosuppressive chemokine program, resulting in fewer infiltrating NK and CD8⁺ T cells, reducing local TNF super family ligands and profound immune checkpoint blockade (ICB) resistance^25,28,29.^,28,29. Additionally, RIPK1’s scaffold function can sub-lethally activate caspase-8, which in turn cleaves and inactivates RIPK3, CYLD and RIPK1 itself, effectively blocking necroptosis^30–33. Since RIPK1 suppresses necroptosis signaling downstream of TLR3/4 and ZBP1^{10,11,13–16,24,31,34–44}, targeting RIPK1 has the potential to overcome apoptosis-resistance and trigger more immunogenic forms of cell death^8,45,46. At present, most anti-cancer drugs kill cancer cells via apoptosis, which is generally immunologically silent. Necroptosis is a caspase-independent, lytic form of cell death that initiates strong immune responses^2,47. It is thought that this immunogenicity originates from the ability of death receptors or PRRs to activate NF-κB and/or interferon signaling while the cell is dying by necroptosis^45–50. Both interferon signaling and NF-κB activation drive the production of inducible DAMPs (iDAMPs) that, together with constitutive DAMPs, potently alert the immune system of danger⁵¹, ensuring that antigen-presenting dendritic cells simultaneously take up dead cells to sample cancer epitopes and become activated, to promote CD8⁺ T-cell cross-priming⁶. Therefore, developing therapies that drive necroptosis may benefit from both killing tumor cells directly and (re)activating the immune system of a patient against their cancer⁴⁶.

Although RIPK1 has emerged as a therapeutic target, its disease-associated activity cannot easily be neutralized by kinase inhibitors alone²⁵. Therefore, alternative strategies, such as the use of PROTAC (proteolysis-targeting chimera) degraders, that can specifically target RIPK1’s scaffolding function, may more effectively treat RIPK1-driven disease. To improve RIPK1 targeting, we generated a series of bifunctional PROTAC molecules. We found that these pharmacological compounds degraded RIPK1 at nanomolar concentrations in human and mouse. Acutely depleting RIPK1 primed TNFR1- and TLR3/4 pathways to over-activate NF-κB, MAPK, and IFN signaling. Simultaneously, RIPK1 degradation deregulated RIPK3 activation and necroptosis, in a cell and tissue-type dependent manner. We demonstrated that RIPK1-PROTAC degraders synergized with immunostimulatory therapies, such as radiotherapy (RT) and ICB. Mechanistically, RIPK1 depletion sensitized cancer cells to RT-induced TNF and interferons. This promoted immunogenic cell death (ICD) and drove long-lasting anti-tumor immunity. Moreover, acutely depleting RIPK1 in the skin protected animals from RIPK1-driven cell death and skin inflammation. Consequently, targeting RIPK1 by PROTACs is a promising approach to alleviate RIPK1-driven pathologies.

RESULTS

RIPK1 PROTACs achieve selective and potent degradation of RIPK1

RIPK1 regulates cell death and immunity through both its catalytic and non-catalytic scaffolding functions. Therefore, its disease-associated activity cannot easily be neutralized by kinase inhibitors alone²⁵. Moreover, while RIPK1 kinase inhibitors are highly effective in mice^10–16, the selective RIPK1 kinase inhibitors PK68⁵² and GSK’963⁵³ only partially suppressed RIPK1-induced cell death in human cells as they ineffectively blocked RIPK1-induced apoptosis (Suppl. Fig. S1A,B), yet efficiently blocked necroptosis (Suppl. Fig. S1C). To improve RIPK1 targeting, we generated a series of heterobifunctional PROTAC molecules⁵⁴ composed of a RIPK1-binding warhead, a linker and a ligand for the von Hippel-Lindau (VHL) E3 ligase (Fig. 1A and Suppl. Fig. S1D). We used PK68 as a RIPK1-binding entity⁵², which was fused to a VHL ligand⁵⁵ via linkers of various lengths. RIPK1-PROTAC treatment showed a dose dependent decrease in RIPK1 levels in a panel of human and murine cell lines and primary cells (Fig. 1B–E and Suppl. Fig. S1E–J). R1-ICR-5 was the most efficient RIPK1 degrader, depleting RIPK1 at nanomolar concentrations in human and murine cells. RIPK1 degradation by R1-ICR-5 was VHL-mediated as R1-ICR-5S, which has inverted stereochemistry at the proline 4-position that precludes VHL binding⁵⁶, failed to deplete RIPK1 in a similar manner (Fig. 1F). Moreover, PK68 treatment did not affect RIPK1 protein levels (Suppl. Fig. S1E,F). Together, these data indicate that R1-ICR-5 triggers VHL-mediated degradation of RIPK1.

Figure 1. Development of selective RIPK1-PROTAC degraders.

(A) Schematic representation of RIPK1-PROTACs.

(B) HT1080^RIPK1-HiBiT cells were treated with the indicated PROTACs for 24 h. HiBiT measurements are shown. RLU, relative light unit.

(C) Western blot analysis evaluating RIPK1 degradation. Cells were incubated with the indicated concentrations of R1-ICR-5 for 6 h.

(D) Micro-confocal images of the indicated endogenous proteins in cells treated with the indicated PROTACs for 20 h. Scale bars: 100 μm (HT1080/HT29) and 50 μm (EMT6/L929).

(E) Quantification of relative endogenous RIPK1 levels from (D).

(F) Western blot analyzis of cells treated with the indicated agents for 6 h.

(G) Western blot analyzis of primary BMDMs treated as indicated. Where indicated, cells were pre-treated with VH298 (1 μM) or bortezomib (100 nM) for 1 h before R1-ICR-5 exposure. RIPK1 degradation was quantified by densitometry (right panel).

(H) Volcano plot depicting whole proteome of BMDMs treated with R1-ICR-5 (1 μM, 5 h) vs non-treated controls. Log₂ fold change and −log₁₀ adjusted P values (Benjamin-Hochberg procedure) are shown along the x- and y-axis, respectively.

(I) Washout experiment. Mlkl^−/− BMDMs were treated with DMSO or R1-ICR-5 (1 μM) for 24 h before washout. Mlkl^−/− cells were used to avoid loss of cellular material. Samples were harvested at the indicated time points, and analyzed by western blot (left panel) and quantified by densitometry (right panel). Data shown are technical replicates and are representative of 3 independent biological repeats. Data show ± SD and are representative of ≥2 independent biological repeats. Statistical analysis (E,I) was calculated by two-way ANOVA with Sidak’s multiple comparison test and (G) one-way ANOVA with Bonferroni’s multiple comparison test.

At effective RIPK1-PROTAC concentrations, we noticed no apparent changes in protein levels of related kinases, such as RIPK2, RIPK3, BRAF and TrkA, a potential ‘off-target’ binder of PK68 (Fig. 1G and Suppl. Fig. S2A–E)⁵². Co-treatment with PK68 effectively outcompeted RIPK1 PROTACs, preventing RIPK1 degradation (Suppl. Fig. S2F,G). R1-ICR-5 degraded RIPK1 as early as 2 hrs after treatment and was nearly complete at 6 hrs (Fig. 1G). Whole proteome analysis demonstrated that only RIPK1 was significantly degraded by the RIPK1-PROTACs in BMDMs and U937s (Fig. 1H, Suppl. Fig. S2H). Treatment with a VHL or proteasome inhibitor confirmed the requirement for VHL-binding and proteasomal proteolysis for degradation (Fig. 1G, Suppl. Fig. S2A–G). Washout experiments indicated that RIPK1 has a low re-synthesis rate, taking more than 72 hrs to reach pre-treatment levels (Fig. 1I and Suppl. Fig. S2I). Although R1-ICR-5’s lipophilicity is currently unfavorable for systemic treatments in animals, local administration of R1-ICR-5 effectively depleted RIPK1 in vivo (Suppl. Fig. S2J).

RIPK1 degraders deregulate innate immune signaling

Next, we evaluated whether acutely depleting RIPK1 compromised TNF or TLR3 signaling^57,58. Unexpectedly, R1-ICR-5 treatment deregulated the transcriptional response of TNF-treated primary BMDMs and triple negative breast cancer (TNBC) EO771 cells, enhancing TNF-induced cytokine expression (Fig. 2A–C and Suppl. Fig. S3A–C). Most notable among the TNF-stimulated genes were Tnf, Il6 and A20 (Tnfaip3). While acute loss of RIPK1 accentuated TNF-induced NF-κB signaling, we noticed that depleting RIPK1 in resting BMDMs induced the transcription of Tnf, A20 and Sod2 (Fig. 2C and Suppl. Fig. S3C). R1-ICR-5-mediated TNF production depended on MK2-mediated TNF biosynthesis⁵⁹ and subsequent autocrine TNF-mediated signaling (Fig. 2D). Consistent with deregulation of TNF signaling, we noticed enhanced recruitment of TRADD to TNFR1 when RIPK1 was depleted (Fig. 2D and Suppl. Fig. S3D), likely due to the competitive dynamics for TNFR1 binding between TRADD and RIPK1⁶⁰. Enhanced TRADD recruitment was accompanied with accentuated ubiquitylation of TRADD, TRAF2, cIAP1 and HOIP (Fig. 2E). Since ubiquitylation of complex-I components mediates NF-κB activation, it is likely that enhanced ubiquitylation of complex-I results in deregulated TNF signaling in BMDMs, L929, EO771 and human LIM1215 and HT29 cells (Fig. 2F and Suppl. Fig. S3B,C,E–G).

Figure 2. Acute degradation of RIPK1 deregulates TNFR1 and TLR3 signaling.

(A) Schematic representation depicting RIPK1’s regulation of TNFR1 and TLR3/4-induced signaling and cell death.

(B) Relative mRNA expression of NF-κB target genes, in BMDMs pre-treated with DMSO or R1-ICR-5 for 4 h, followed by indicated TNF treatment.

(C) Relative Tnf mRNA expression in BMDMs treated for 4 h with the indicated conditions. MK2i refers to an MK2 inhibitor.

(D) Western blot analysis of TNFR1 signaling complex-I. Cells were treated with DMSO or R1-ICR-5 (overnight).

(E) L929 cells were treated as in (D) before anti-GST-TUBE pulldown to isolate the ubiquitylated proteome. *Indicates non-specific signal.

(F) Western blot analysis of BMDMs pre-treated with DMSO or R1-ICR-5 for 4 h, followed by TNF exposure for the indicated time points.

(G) Relative mRNA expression of NF-κB or IFN target genes. BMDMs were pre-treated with DMSO or R1-ICR-5 for 4 h, followed by stimulation with Poly(I:C).

Data show mean ± SD and are representative of (C,E) two, (B,G) three, (F) four or (D) five independent biological repeats. P values were calculated using (B) two-way ANOVA (Sidak’s multiple comparison test) or (C,G) one-way ANOVA (Bonferroni’s multiple comparison test).

RIPK1-PROTAC treatment also deregulated TLR3 signaling, enhancing the induction of Il-6, Tnf, Ccl2, Cxcl9, Cxcl10 and Ifnβ in BMDMs and EO771 (Fig. 2G and Suppl. Fig. S3H,I). Likewise, acutely removing RIPK1 genetically, using BMDMs^{LysM-Cre;Ripk1fl/fl} cells, also accentuated innate immune signaling (NF-κB and interferon) in response to TLR3 stimulation (Suppl. Fig. S3J). Moreover, R1-ICR-5 treatment did not accentuate innate immune signaling in EO771 Ripk1-KO cells (Suppl. Table. S1 and S2). These data suggest that R1-ICR-5’s impact on signalling is on-target, and that RIPK1 is required for fine-tuned activation of TNFR1 and TLR3 signaling, preventing hyper-activation of these pathways.

RIPK1 PROTACs facilitate TNFR1- and TLR3/4-mediated activation of RIPK3

Acutely depleting RIPK1 also sensitized cells to cell death, accompanied by enhanced phosphorylation and activation of RIPK3 and MLKL (Fig. 3A–C). Likewise, acutely depleting RIPK1 also enhanced TLR3- and TLR4-induced necroptosis that was mediated by TRIF (also known as Ticam1), independent of TNFR1 and ZBP1 (Fig. 3D,E and Suppl. Fig. S3K,L). TLR3/4-mediated necroptosis partly required IFNAR1, as Ifnar1-KO suppressed this death (Suppl. Fig. S3K,L). Therefore, acutely degrading RIPK1 not only causes hyper-activation of innate immune signaling pathways (e.g. TNFR1 and TLR3), but also triggers unhindered activation of cell death.

Figure 3. RIPK1 degraders drive TNFR1-driven necroptosis, independent of TRIF and ZBP1.

(A, B) Quantification of propidium iodide positive (PI⁺) cells, treated with the indicated conditions. Cells were pre-treated with DMSO, R1-ICR-5 or RIPK2 degrader (A,18 h; B,5 h), before stimulation with the indicated conditions (A,5 h; B,7 h).

(C) Western blot analysis monitoring RIPK3 and MLKL activation. Cells were pre-treated for 18 h, before exposure to TE for the indicated time points. *Indicates non-specific signal.

(D, E) Quantification of PI⁺ cells. Cells were pre-treated with DMSO, R1-ICR-3 or R1-ICR-5 for 4 h, followed by the indicated treatments (3 h).

(F) Quantification of PI⁺ BMDMs. Cells were pre-treated with DMSO or R1-ICR-5 for 4 h, followed by the indicated treatment.

(G) Quantification of PI⁺ L929 cells. Cells were untreated or pre-treated with R1-ICR-5 for 12 h, followed by treatment with the indicated agents. RIPK1 kinase inhibitors were added 30 min prior to TE (6.5 h).

(H-K) Quantification of PI⁺ cells. The indicated cells were treated with R1-ICR-5, R1-ICR-5S, or MK2 inhibitor for 24 h.

(L) Quantification of PI⁺ cells. Ripk3^−/− cells were reconstituted with Dox-inducible WT or RHIM-mutant RIPK3 (RHIMm). Cells were induced with Dox and incubated with R1-ICR-5 for 12 h, before treatment with TE for 2.5 h.

(M) Quantification of PI⁺ cells. Cells were pre-treated with DMSO or R1-ICR-5 for 18 h, following treatment with TE (5 h). Trif and Zbp1 deficiency was confirmed by western blot analysis.

(N) Quantification of PI⁺ BMDMs. Cells were treated with the indicated agents for 24 h (left panel). Right panel: western blot analysis of BMDMs following treatment with DMSO or R1-ICR-5 for 6 h.

(O) Quantification of PI⁺ BMDMs. The indicated cells were treated with R1-ICR-5 for 24 h.

Data show ± SD and are representative of three (A,D-O), four (B) or two (C) independent biological repeats. P values were calculated using (F,H,K,O) one-way ANOVA (Bonferroni multiple comparison test) or (A,B,D,E,G,I,J,L-N) two-way ANOVA (Sidak’s multiple comparison test).

TNFR1 but not TNFR2 was required for TNF-induced cell death following depletion of RIPK1, because treatment with human TNF, which only binds to murine TNFR1, was fully capable of inducing necroptosis in RIPK1-depleted BMDMs (Fig. 3F). TNF/R1-ICR-5-induced cell death was entirely RIPK1-independent because co-treatment with numerous RIPK1 kinase inhibitors did not block this death (Fig. 3G). However, when RIPK1 was present, these RIPK1 kinase inhibitors potently blocked TNF-induced cell death, suggesting that the scaffold function of RIPK1 blocks the ability of TNFR1 to trigger cell death.

Depleting RIPK1 also sensitized cells to autocrine-produced TNF. Accordingly, primary BMDMs were sensitive to RIPK1-PROTAC treatment alone, and this effect was prevented by genetic deletion of Tnfr1 or co-treatment with Enbrel, a TNF-neutralizing biologic (Fig. 3H,I and Suppl. Fig. S3M). Inhibition of MK2- mediated TNF biosynthesis⁵⁹ also suppressed R1-ICR-5 killing in primary BMDMs. The inactive control compound R1-ICR-5S, or engagement of VHL using a RIPK2-targeting PROTAC, had no effect on cell viability (Fig. 3B,I,J), indicating that the cytotoxic effect of R1-ICR-5 was specific to RIPK1 depletion and relied on engagement of the TNFR1 pathway.

The cytotoxicity of the PROTACs correlated with their ability to degrade RIPK1, with R1-ICR-5 being the most potent (Suppl. Fig. S1G and S3N). R1-ICR-5 not only killed primary BMDMs, but also mouse dermal fibroblasts and peritoneal macrophages (Suppl. Fig. S3O,P), expanding this observation to other cell types. PROTAC treatment alone predominantly caused necroptosis in sensitive cells (Fig. 3J,K and Suppl. Fig. S3P), indicating that TNFR1 signaling can trigger RIPK3 activation when RIPK1 is acutely depleted (Suppl. Fig. S3Q), via an unknown mechanism, an observation that was noted previously^61–64. Although RIPK1-PROTAC-induced activation of RIPK3 occurred in a range of cell types, the response to acute depletion of RIPK1 was cell and tissue type dependent, as some were unable to drive autocrine TNF-mediated RIPK3 activation, in RIPK1’s absence. For example, naïve CD4⁺ and CD8⁺ T cells and keratinocytes^10,15 resisted cell death triggered by acute depletion of RIPK1 (Suppl. Fig. S3R).

To corroborate the effects of R1-ICR-5, we genetically and conditionally deleted RIPK1, using Doxycycline (Dox)-Cre;Ripk1^fl/fl MDFs and LysM-Cre;Ripk1^fl/fl BMDMs. Akin to R1-ICR-5, Dox-induced deletion of Ripk1 in MDFs^{Dox-Cre;Ripk1fl/fl} triggered RIPK3-dependent cell death (Suppl. Fig. S3S). RNAi-mediated depletion of Ripk1 did not potentiate this death. Yet, in RIPK1-proficient, non-Dox treated MDFs, RNAi-mediated depletion of RIPK1 did indeed cause cell death, indicating that RIPK1 acts as pro-survival factor in these cells. Likewise, BMDMs^{LysM-Cre;Ripk1fl/fl} were more sensitive to TNFR1- and TLR3/4-induced necroptosis than their control counterparts (Suppl. Fig. S3T). Additionally, EO771^{Dox-Ripk1shRNAi} cells, harboring a Dox-inducible Ripk1-targeted shRNA, also became acutely sensitive to TNF-induced necroptosis upon shRNA-mediated depletion of RIPK1 (Suppl. Fig. S3U). Together, these data demonstrate that genetic, acute deletion of RIPK1 fully recapitulates the phenotypes triggered by R1-ICR-5.

TNF/RIPK1-PROTAC-induced RIPK3 activation occurs independent of TRIF and ZBP1

Next, we evaluated how TNF drives RIPK3 activation. The recruitment of RIPK3 to the RIP Homotypic Interaction Motif (RHIM) of either RIPK1, TRIF, or ZBP1 can initiate RIPK3 activation and necroptosis^65,66. To evaluate the significance of RIPK3’s RHIM, we reconstituted RIPK3-deficient cells with wild-type RIPK3 or a RHIM-mutant form of RIPK3⁶⁷. While RIPK3^WT readily reconstituted TNF-induced necroptosis in RIPK1’s absence, the RIPK3^RHIMm did not (Fig. 3L), demonstrating the indispensable role of RIPK3’s RHIM, in its activation.

As TRIF and ZBP1 are the only other RHIM-containing proteins, we investigated whether TRIF and/or ZBP1 mediated TNFR1-induced RIPK3 activation upon RIPK1 depletion. Double deficiency of Trif and Zbp1 in primary lung fibroblast (LF) did not impact TNF-induced cell death upon R1-ICR-5 treatment (Fig. 3M,N), ruling out any involvement of these RHIM-containing proteins. While Trif^−/− partially reduced R1-ICR-5-induced cell death in primary BMDMs, it did not prevent it entirely (Fig. 3N). ZBP1 deficiency did not affect R1-ICR-5-induced cell death. Under more potent cell death conditions, such as in the presence of the pan-caspase inhibitor Emricasan (E) or upon treatment with TNF/Smac mimetic (SM)/E or TNF/cycloheximide (CHX), Trif and Zbp1 deficiency had negligible impact on TNF-induced necroptosis in RIPK1-depleted BMDMs (Fig. 3N and Suppl. Fig. S3V,W). Our findings are unexpected because whole body or skin-specific deletion or RHIM mutation of RIPK1 in mice, leads to necroptosis in the skin and lethality mediated by ZBP1 and, to a lesser extent, TRIF^10,11,15,68. However, we did not observe changes in ZBP1 levels upon acute removal of RIPK1 in primary LFs and BMDMs (Fig. 3M,N), which contrasts with in vivo models where ZBP1 expression is prominently induced at the age of P28¹⁵. This suggests that the ZBP1-dependent in vivo phenotype of Ripk1^E-KO and Ripk1^mRHIM/E-KO mice might be an indirect, tissue-integrated response to long-term inactivation of RIPK1 function, rather than a direct cell-intrinsic response to RIPK1 loss. We noticed that Trif^−/− BMDMs had lower basal protein expression of interferon-responsive gene products, ISG15, MLKL, and ZBP1 (Fig. 3N), suggesting that loss of Trif might reduce the sensitivity to TNF indirectly, perhaps by reducing MLKL and TNF⁶⁹. Consistently, Ifnar1^−/− reduced autocrine-TNF and TLR3/4-induced cell death in RIPK1-depleted BMDMs (Fig. 3O and Suppl. Fig. S3K,L). Taken together, while we find that the RIPK3’s RHIM domain is required for TNF- and TLR3/4-induced necroptosis following R1-ICR-5 treatment, the RHIM-containing proteins TRIF and ZBP1 are dispensable.

The death domain of RIPK1 limits TRADD fibrillation

Next, we tested the involvement of complex-I components in RIPK3 activation. Depletion of TRADD fully protected cells from TNF-induced necroptosis upon PROTAC or genetic depletion of RIPK1 (Fig. 4A). While genetic deletion and RNAi-mediated depletion of Spata2 and Cyld partially protected cells under these conditions, depletion or pharmacological inhibition of no other signaling component downstream of TNFR1 suppressed TNFR1-mediated RIPK3 activation (Fig. 4B,C and Suppl. Fig. S4A,B). This indicates that TRADD plays a critical role in driving R1-ICR-5-mediated activation of RIPK3.

Figure 4. RIPK1 degradation facilitates TRADD fibrillation and TRADD-RIPK3 interaction.

(A, B) Quantification of PI⁺ L929s. Cells were treated with the indicated RNAis for 48 h, followed by treatment with control or R1-ICR-5 for 24 h. Subsequently, cells were treated with TE for 2 h.

(C) Quantification of PI⁺ BMDMs. Cells were pre-treated with DMSO or R1-ICR-5 for 4 h, followed by treatment with TZ or TE. Data represent 2 biological replicates and are representative of three independent experiments.

(D) TNFR1 complex-II analysis using L929^{Dox-Flag-TRADD} cells pre-treated with DMSO or R1-ICR-5 overnight. FLAG-TRADD was induced for 1 h (Dox), followed by treatment with TE (1 h). Lysates were subjected to anti-FLAG IP before western blot and densitometry analyzes.

(E) Western blot analysis of L929^{Dox-Flag-TRADD} cells pre-treated with DMSO or R1-ICR-5 for 18 h, followed by 1 h induction of FLAG-TRADD (Dox). Subsequently, cells were treated with TE for the indicated time points. *Indicates non-specific signal.

(F) Quantification of PI⁺ L929^{Dox-Flag-TRADD} cells. Cells were pre-treated with DMSO or R1-ICR-5 for 18 h, followed by 1 h induction of TRADD (Dox). Subsequently, cells were treated with TE (2 h).

(G) Confocal microscopy images of TRADD fibrils using L929^{Dox-Flag-TRADD}. Cells were treated with DMSO or R1-ICR-5 overnight, followed by 1 h Dox-induction of FLAG-TRADD (in the presence of E). Subsequently, cells were treated with TNF for 10 minutes. i) and ii) are magnifications of the indicated areas. Scale bars: 50 μm and magnification 10 μm. Quantification of the number of TRADD aggregates per cell (~70 cells/condition) (right panel).

(H) Analysis of the ubiquitylated proteome. cells were pre-treated with DMSO or R1-ICR-5 overnight, followed by treatment with TE for the indicated time points. Cells were subjected to TUBE pull-down. Samples were split and boiled with or without β-mercaptoethanol (reducing or non-reducing). *Indicates non-specific signal.

(I) Predicted AlphaFold structure of mTRADD (AF-Q3U0V2-F1, left panel)⁹⁰. C198 has a confidence of 45–58 %, whereas all other cysteines have a >85 % confidence. Right panel, schematic representations of mTradd constructs used to reconstitute L929 Tradd^−/− cells in (J,K). DD, Death Domain.

(J,K) Quantification of PI⁺ cells, reconstituted with the indicated Dox-inducible Tradd transgenes. Cells were treated with DMSO or R1-ICR-5 overnight, prior to treatment with Dox/E for 2 h. Subsequently, cells were treated with TNF for 3h.

(L) Quantification of PI⁺ cells, stably expressing Dox-inducible Ripk1 constructs. Cells were treated with DMSO or R1-ICR-5 overnight, before incubation with Dox/E for 3 h. Subsequently, cells were treated as indicated.

(M) Schematic diagram depicting rewiring of TNFR1-mediated activation of RIPK3 upon R1-ICR-5 treatment. RIPK1, through its DD (Death domain), competes with TRADD for the binding of TNFR1. Upon depletion of RIPK1, TRADD is enriched at the TNFR1 signaling complex-I, ultimately leading to accentuated formation of complex-II, RIPK3 activation and necroptosis.

Data show ± SD and are representative of (a-c,f-h,j-l) 3 or (d,e) 2 independent biological repeats. P values were calculated using (d,k) one-way ANOVA (Bonferroni multiple comparison test) or (a-c,f,j,m) two-way ANOVA (Sidak’s multiple comparison test) or (g) unpaired t-tests.

RIPK1 not only suppressed the recruitment of TRADD to TNFR1 complex-I (Fig. 2D), but also prevented its association with RIPK3 in the death-inducing complex-II (Fig. 4D). Enhanced recruitment of RIPK3 to TRADD correlated with enhanced activation of RIPK3, MLKL and cell death (Fig. 4D–F and Suppl. Fig. S4C). Consistently, proximity ligation assay (PLA) and confocal co-localization analysis indicated that TRADD and RIPK3 come into close proximity upon R1-ICR-5 treatment (Suppl. Fig. S4D,E). Since RIPK3 also interacted with TRADD in RIPK1-deficient cells (Suppl. Fig. S4F), we next evaluated whether TRADD can directly bind to RIPK3. However, reciprocal in vitro binding assays established that recombinant TRADD did not directly bind to RIPK3, suggesting that the interaction between TRADD and RIPK3 is indirect, or might require post-translational modifications (Suppl. Fig. S4F–I).

RIPK1 suppressed TRADD aggregation and fibrillation in response to TNF (Fig. 4G), since treatment with TNF caused the time-dependent accumulation of TRADD aggregates when RIPK1 was degraded. Depletion of RIPK1 by RNAi similarly caused TNF-mediated TRADD fibrillation (Suppl. Fig. S4J,K). These data are consistent with a model whereby the recruitment of RIPK3 to TRADD fibrils causes a conformational change of RIPK3 that allows RHIM-mediated homo-oligomerization, ubiquitylation and auto-activation. Consistently, TNF stimulation markedly enhanced the oligomerization, ubiquitylation and activation of RIPK3 and MLKL, which was preceded by TRADD ubiquitylation and oligomerization (Fig. 4H).

To elucidate how TRADD initiates RIPK3 activation upon R1-ICR-5 treatment, we created several TRADD variants and assessed their capacity to restore TNF-induced RIPK3 activation in TRADD-deficient cells (Fig. 4I). While Dox-induced expression of TRADD-WT or -DD alone had no effect on cell viability, such cells became acutely sensitive to TE when RIPK1 was depleted, demonstrating that the DD of TRADD is sufficient to drive TNF-induced activation of necroptosis (Fig 4I,J and Suppl. Fig. S4L). Next, we assessed the requirement of TRADD oligomerisation for RIPK3 activation. We noticed that TRADD readily formed higher order oligomers (Suppl. Fig. S4I) that were destabilised by reducing agents (Fig. 4H), suggesting that di-sulphide bridges might contribute to the formation of TRADD oligomers. To test this, we generated TRADD C>S mutants with substitution mutations on surface exposed Cysteines (Cys) (Fig. 4I). We also included a TRAF-binding mutant⁷⁰. While such mutants were still recruited to TNFR1 complex-I, they were severely impaired in driving RIPK3-mediated necroptosis upon R1-ICR-5 treatment (Fig 4I,K and Suppl. Fig. S4M). Particularly, TRADD^Call>S, in which all surface exposed Cys were substituted to Ser, was strongly impaired in driving RIPK3 activation, even though this mutant was recruited to TNFR1 as efficiently as the TRAF-binding mutant (TRADD^Y16A/F18A) (Suppl. Fig. S4M). While the TRADD^Call>S and TRADD^Y16A/F18A mutants failed to recruit TRAF2, cIAPs and LUBAC to complex-I, TRADD^Y16A/F18A was proficient in activating RIPK3 upon R1-ICR-5 treatment. Under the same conditions, however, TRADD^Call>S was unable to drive RIPK3-mediated necroptosis. Together, these data are consistent with the notion that di-sulphide bridges are required for TRADD oligomerisation and RIPK3 activation.

Next, to map the domain of RIPK1 that suppressed TRADD fibrillation, we generated Dox-inducible RIPK1 truncation mutants (RIPK1^ΔKDΔRHIM and RIPK1^DD-only), carrying the DD but lacking the RHIM. Both mutants efficiently suppressed TNF-mediated activation of necroptosis yet had no impact on TLR3/TRIF-driven necroptosis upon R1-ICR-5 treatment (Fig. 4L), as expected. Together, our data suggest that RIPK1, through its DD, regulates TRADD by competing for TNFR1 occupancy in complex-I and additionally, may suppress TRADD:TRADD homo-oligomerization, to prevent further accumulation and fibrillation of TRADD and resultant RIPK3 activation (Fig. 4M).

PROTAC-mediated degradation of RIPK1 achieves anti-tumor activity

The coordinated exposure to iDAMPs and dead cells can activate immune cells, driving anti-tumor immunity^45,47. Since depletion of RIPK1 concomitantly enhances NF-κB/interferon signaling and responsiveness to necroptosis, we explored the potential of using PROTAC-mediated degradation of RIPK1 to trigger ICD and boost anti-cancer therapies. Upon R1-ICR-5 treatment, several distinct cancer cell lines were capable of driving TNFR1-, TLR3 and IFNR-mediated necroptosis (Fig. 5A–C, Suppl. Fig. S5A). Ripk1^−/− deficiency caused resistance to TNF-induced necroptosis upon R1-ICR-5 treatment (Suppl. Fig. S5B). The observation that cells become vulnerable to TNF-induced necroptosis following acute pharmacological or genetic removal of RIPK1, yet are resistant to the same stimulus when RIPK1 is permanently deleted, implies that cells ultimately adapt to long-term absence of RIPK1. Treatment with R1-ICR-5 also sensitized cancer cells to IFN-induced necroptosis (Fig. 5C,D), which agrees with a previous report¹¹. TRADD efficiently bound to RIPK3 following R1-ICR-5 treatment, leading to increased activation of RIPK3 and necroptosis (Fig. 5E–G). Similar to L929 cells, RIPK1’s DD effectively suppressed TNFR1- but not TLR3-induced necroptosis, in EO771 TNBC cells, corroborating the notion that these cancer cells are similarly wired (Suppl. Fig. S5C).

Figure 5. RIPK1 protects cancer cells from TNF and IFN-induced cell death.

(A,B) Quantification of PI⁺ cells. Cells were pre-incubated with DMSO or R1-ICR-5 for 4 h, cells were left untreated or stimulated with TNF or TE in the presence and absence of RIPK3 inhibitors (6, 48 and 24 h for EO771, MC38 and MCA-205, respectively).

(C) Quantification of PI⁺ cells. Cells were pre-treated with DMSO or R1-ICR-5 (4 h), before 24 h of IFNβ or IFNγ. MCA-205 and MC38 were pre-treated overnight with IFNγ before incubation with DMSO or R1-ICR-5 for 24 h (MCA-205) or 48 h (MC38).

(D) Quantification of PI⁺ cells. Cells were pre-treatment with IFNβ in the presence of E overnight. Subsequently, cells were treated with R1-ICR-5 for 5 h, before TNF treatment (24 h).

(E) Purification of TNFR1 complex-II from EO771^{Dox-FLAG-TRADD} cells. Cells were treated with DMSO or R1- ICR-5 for 2 h, before induction of Tradd for 2 h (Dox). Subsequently, cells were treated as indicated and analyzed by western blotting. Densitometry was used to quantify P-RIPK3. P-RIPK3 signal was normalized to purified TRADD.

(F) Western blot analysis of the indicated proteins in EO771^{Dox-FLAG-TRADD}. Cells were pre-treated with DMSO or R1-ICR-5 for 4 h. Subsequently, TRADD was induced with Dox and cells were treated with TE for the indicated time points.

(G) Quantification of PI⁺ cells, treated as in (F). Data are triplicate technical repeats and are representative of two independent biological repeats.

(H) IFNβ and TNF gene expression analysis of EO771 cells left untreated or exposed to 8 Gy irradiation. Data show mean ± SD and are representative of three biological repeats. Statistical analysis was performed by unpaired t test.

(I-L) Quantification of PI⁺ cells, subjected to the indicated dose of irradiation. 24 h after treatment, cells were incubated with DMSO or R1-ICR-5 for 24 h. (K) Enbrel was added 30 minutes prior to irradiation.

(m) Quantification of PI⁺ cells, stably expressing the indicated shRNA constructs, treated as in (K) in the presence of Dox.

Data show ± SD and are representative of (A-D,F,H-M) three, (E) two or (G) four independent biological repeats. P values were calculated using two-way ANOVA (Sidak’s multiple comparison).

Since PROTAC-mediated depletion of RIPK1 sensitized cancer cells to TNF and IFN (Fig. 5A–D), we hypothesized that R1-ICR-5 may synergize with therapies that generate a TNF and IFN-rich tumor micro-environment, such as RT and ICB. We focused on TNBC, where RT and ICB are standard-of-care treatments. Exposing EO771 cells to RT caused dose-dependent DNA damage (Suppl. Fig. S5D), which in turn markedly induced NF-κB and IFN-responsive genes, including genes associated with ICD (Cxcl10 & Ifnβ)⁷¹ and TNFRSF signaling (Tnf & Tnfsf10 (Trail)) (Fig. 5H, Suppl. Fig. S5E,F). RT-induced cytokine production was entirely TRIF-dependent, as Dox-induced depletion of Trif abrogated RT-induced NF-κB and IFN target gene expression (Suppl. Fig. S5G). RT not only stimulated NF-κB- and IFN-driven signalling responses, but also caused necroptosis in EO771 and MC38 cells, that was blocked by Mlkl deficiency and enhanced by RIPK1-PROTAC-treatment (Fig. 5I,J). Together, this indicates that RIPK1 acts as a survival factor, suppressing RT-mediated cell death. RT-induced necroptosis was partly mediated by TNF, because treatment with Enbrel or Tnfr1 deficiency suppressed the treatment response (Fig. 5K,L). We also noticed that TRIF contributed to the treatment response, because RNAi-mediated depletion of Trif suppressed RT-induced necroptosis, which was further reduced by concomitantly blocking TNF signaling (Fig. 5M). In contrast, depletion of Zbp1 had a relatively minor impact on RT-induced necroptosis (Suppl. Fig. S5H). Together, these observations suggest that cancer cells can hijack RIPK1 to suppress the cellular response to RT.

Next, we tested whether targeting RIPK1 could boost the ability of radiotherapy to ‘heat up’ tumors in vivo. To this end, EO771 TNBC cells were injected into the mammary fat pad of immune competent mice. Upon tumor formation, image-guided RT (8 Gy) was administered, followed by three intra-tumoral injections of R1-ICR-5 (Fig. 6A and Suppl. Fig. S6A,B,E). While RT and R1-ICR-5 had negligible effects as monotherapies, RT/R1-ICR-5 achieved therapeutic responses in 50% of mice (Fig. 6B–D and Suppl. Fig. S6C–E). While three out of ten mice were tumor-free beyond 70 days, one mouse relapsed at day 76 (Fig. 6B–D). Tumor analysis revealed that RT/R1-ICR-5 reshaped the tumor immune microenvironment, promoting TNF⁺IFNγ⁺ lymphocyte infiltration (Fig. 6E–G and Suppl. Fig. S6F,O and Suppl. Data S1) and favoring the recruitment of adaptive and innate lymphocytes that exhibited key features of activation (CD69⁺, CD44⁺, CD62L⁻) (Suppl. Fig. S6G–J) and anti-tumor function (TNF⁺, IFNγ⁺, Granzyme B⁺ and CD107a⁺) (Fig. 6E,F and Suppl. Fig. S6K–N). The emergence of CD8⁺ and CD4⁺ effector memory (T_EM; CD44^hiCD62L⁻) populations at early time points (Suppl. Fig. S6G,H), suggested that RIPK1 degradation enhanced RT-induced immunity, which was confirmed by cured animals (Fig. 6C,D) being fully resistant to tumor re-challenge (Fig. 6H,I). Together, these data indicate the R1-ICR-5 may extend the survival of RT-treated mice by enhancing the recruitment of activated TNF+IFNy+ lymphocytes to the tumor microenvironment, while simultaneously rendering EO771 cancer cells vulnerable to TNF/IFNy-induced cell death.

Figure 6. RIPK1 functions as a protective factor for cancer cells against the effects of RT.

(A) Schematic depicting the treatment regimen of tumor bearing mice. I.T, Intratumoral injection.

(B) Tumor growth curves of tumor-bearing mice treated as depicted in (A): sham (n = 8), RT (n = 11), RT/R1-ICR-5 (n = 10). Thick lines represent average tumor growth. Curves represent two independent experiments.

(C) Survival curves of EO771 tumor bearing mice treated as depicted in (A,B). Curves represent two independent experiments. Median survival (RT = 24 days; RT/R1-ICR-5 = 35 days) (Logrank (Mantel-cox) P=0.0290).

(D) Pie-charts depicting the response of treated mice from (A,C). Chi-square test – Sham Vs RT/R1-ICR-5, P=0.0186; RT Vs RT/R1-ICR-5, P=0.0382.

(E-G) Flow cytometric analysis of tumors treated as in (A) and harvested on day 13 post-RT. Sham (n=10), RT (n=9) and RT/R1-ICR-5 (n=8). Data are representative of two independent experiments.

(E) The number of TNFα⁺IFNγ⁺ immune (CD45⁺) and non-immune cells (CD45⁻) per gram of tumor.

(F) The number of TNFα⁺IFNγ⁺ lymphocytes per gram of tumor.

(G) The number of CD45⁺, CD8⁺ T, CD4⁺ conventional T (CD4_cv), γδ T, NK and NK T cells per gram of tumor.

(H) Tumor rechallenge experiment. Data show the growth curves of EO771 mammary tumors, in naïve mice (n=7, black lines) or RT/R1-ICR-5-treated tumor-free mice from (A-D) (100 days post-treatment (n=2, red lines). Each line represents 1 animal, thick lines denote average tumor growth. I.P, intraperitoneal injection; I.T, intratumoral injection.

(I) Pie-charts depicting the proportion of mice bearing tumors, 42-days after re-challenge with EO771 cells (H). Chi-square test, P=0.0027.

(J) Schematic depicting the treatment regimen of tumor bearing mice.

(K) Tumor growth curves of mice treated as in (J). Thick lines represent average tumor growth. Treated mice: IgG (n = 5), anti-CTLA4 (n = 7), RT (n = 10), RT/R1-ICR-5 (n=10), RT/anti-CTLA-4 (n=10) and RT/anti-CTLA-4/R1-ICR-5 (n=10).

(L) Tumor growth kinetics (day 0–30) of mice treated as in (J-K), measured by area under the curve (AUC). Each point represents the AUC of individual mice, from (K).

(M) Survival curves of mice treated as in (J,K). Curves are representative of one biological replicate. Median survival RT/anti-CTLA-4 = 70.5 days; RT/anti-CTLA-4/R1-ICR-5 = undefined.

(N) Pie-charts depicting the response of tumor-bearing mice treated as in (I-M).

Data show ± SD. P values were calculated using (C,M) Log-rank (Mantel-Cox), (D,I,N) Chi-square test (Comparing progression Vs complete response), (E-G) or Kruskal Wallis test or (L) one-way ANOVA (Bonferroni’s multiple comparison test).

Next, we tested whether R1-ICR-5 boosted RT/ICB combinations, using lower doses of RT (4 Gy) followed by α-CTLA-4 ± R1-ICR-5 treatment (Fig 6J–N). While RT, α-CTLA-4, and RT/α-CTLA-4 had negligible impact in the first 30 days, combining RT/α-CTLA-4 with R1-ICR-5 strongly inhibited tumor growth (Fig. 6K,L), with nine out of ten tumor-bearing mice responding to therapy and eight being completely cured (Fig. 6M,N).

We also determined whether R1-ICR-5 could enhance the anti-tumor efficacy of anti-PD-1 treatment alone. While neither R1-ICR-5 or anti-PD-1 ICB proved efficacious as single agents, their combination significantly slowed the growth kinetics of treated tumors (Fig. 7A–D). Combination-treated mice benefitted from improved response rates (60% response) and overall survival, with a third of treated mice being completely cured (Fig. 7E).

Figure 7. RIPK1 PROTACs enhance response to immune checkpoint blockade.

(A) Schematic depicting the treatment regimen of tumo-bearing mice. I.P, intraperitoneal injection; I.T, intratumoral injection.

(B) Tumor growth curves of tumor-bearing mice treated as in (A). Thick lines represent average tumor growth. Treated mice: IgG (n = 8), R1-ICR-5 (n = 9), anti-PD-1 (n = 8), R1-ICR-5/anti-PD-1 (n=10).

(C) Tumor growth kinetics (day 0–21) of mice treated as in (A), measured by area under the curve (AUC). Each point represents the AUC of individual mice, from (B).

(D) Survival curves of tumor-bearing mice treated as in (A). Curves are representative of one independent experiment. R1-ICR-5 improves the median survival of mice treated with anti-PD-1 (median survival: anti-PD-1 = 28.5 days; R1-ICR-5/anti-PD-1 = 36.5 days).

(E) Pie-charts depicting the response of tumor-bearing mice treated as in (A-D). Chi-square test – Captisol/IgG Vs PD-1, P=0.0209; Captisol/IgG Vs PD-1/R1-ICR-5, P=0.0073.

(F) Association of RIPK1 copy number alterations (CNA) with disease free survival in TCGA patients (n=2,601) across multiple tumor types.

(G) Overall survival in RIPK1-low/NK-high versus RIPK1-high/NK-low patient groups from SCAN-B TNBC dataset (n=148).

(H) Gene set enrichment analysis (GSEA) of indicated gene sets in RIPK1-low/NK-high versus RIPK1-high/NK-low TNBC patients (n=148). FDR, False discovery rate q-value; NES, Normalised enrichment score.

Data show ± SD. P values were calculated using (C) one-way ANOVA (Bonferroni’s multiple comparison test), (D,F,G) Logrank (Mantel-Cox) or (E) Chi-square test (Comparing progression Vs complete response).

Our data indicate that RIPK1 confers resistance to innate immune signaling and ICD, functioning as a protective factor for cancer cells and hampering the immunostimulatory effects of RT. Consistently, analyzes of The Cancer Genome Atlas (TCGA) revealed a strong correlation between increased RIPK1 copy number alterations and poor disease-free survival (Fig. 7F). Moreover, stratifying TNBC patients based on RIPK1 status and NK cell infiltration, found that RIPK1-low/NK cell-high was associated with better overall survival (Fig. 7G). Correspondingly, RIPK1-low/NK-high patients differentially induced genes relating to Type-I/II IFN, cell death and inflammatory responses (Fig. 7H and Suppl. Fig. S6P). Therefore, targeting RIPK1 by PROTACs emerges as a promising approach to heat up tumors and overcome resistance to immunostimulatory therapies.

Acute depletion of RIPK1 suppresses RIPK1-driven skin inflammation

Previous work established that aberrant RIPK1-mediated keratinocyte cell death can lead to skin pathologies²². Since keratinocytes are somewhat resistant to RIPK1-independent activation of RIPK3^10,15, we evaluated whether PROTAC-mediated degradation of RIPK1 in vivo (Suppl. Fig. S7A), could suppress RIPK1-induced skin inflammation. Subcutaneous injection of R1-ICR-3 or R1-ICR-5 alone did not cause skin toxicity (Suppl. Fig. S7B,C), indicating that acute depletion of RIPK1 is well tolerated and did not cause RIPK3 activation in the skin. Consistently, conditional, skin-specific deletion of Ripk1 had little effect in 7–11 weeks old mice, whereas Casp8 deletion caused deep ulceration and epidermal loss, which was rescued by co-deletion of Ripk1 (Suppl. Fig. 7D,E). To test the therapeutic effects of RIPK1-PROTACs, we used two skin inflammation models. First, we caused RIPK1-dependent cell death and tissue injury by subcutaneous injection of SM and E (SM/E), which mimics inflammatory diseases caused by IAP loss^72,73. While subcutaneous SM/E injection caused significant skin ulceration⁷³, co-injection of RIPK1-PROTACs protected animals from tissue injury (Suppl. Fig. S7B,C). The SM/E-treated mice achieved the highest Histological multivariant Lesion Score (HLS) of mean 214 (range 175–257), while RIPK1-PROTAC-treated animals reached an HLS mean 49 (range 39–58) and 55 (range 30–70), respectively (Suppl. Fig. S7C). Single injections of SM or E alone had no effect⁷³. In contrast, subcutaneous injection of SM/E caused deep ulceration of the epidermis and subepidermal layer of dermis and showed formation of dense collagenous fibrotic tissue (Suppl. Fig. 7B,C)⁷³. Animals also displayed cellular necrosis, including the presence of nuclear dust mostly at superficial or mid dermis. Co-treatment with RIPK1-PROTACs blocked the appearance of the aforementioned features (Suppl. Fig. 7B,C).

We also tested whether PROTAC-mediated degradation of RIPK1 suppressed the inflammatory pathology driven by TBK1 inactivation. Human TBK1 deficiency leads to severe inflammation that is driven by TNF-induced RIPK1-dependent cell death^74–76. Since biallelic loss of TBK1 is embryonically lethal in mice^74,77, we pharmacologically inhibited TBK1 in mature skin. Subcutaneous injection of MRT67307⁷⁸ (subsequently referred to as TBK1i) caused deep ulceration and tissue injury in combination with TNF (Suppl. Fig. 7F,G), which was fully protected by R1-ICR-5. Together, these data demonstrate that RIPK1-mediated inflammatory skin disease can be suppressed by RIPK1-PROTACs. Unlike RIPK1 kinase inhibitors, which only efficiently blocked necroptosis in human cells, RIPK1-PROTACs potently suppressed RIPK1-induced caspase activation and apoptosis in a panel of human cell lines (Suppl. Fig. S1A,B and Suppl. Fig. 7H,I). In contrast, the RIPK1 kinase inhibitors PK68 and GSK’963 were ineffective to do so in human cells (and S7H,I). Therefore, RIPK1-PROTACs might be effective for treating human pathologies driven by RIPK1-induced apoptosis and/or necroptosis, particularly in tissues inherently devoid of RIPK3 activation, when RIPK1 is depleted.

DISCUSSION

Although RIPK1 has emerged as a therapeutic target, its successful targeting has proved to be difficult, at least in part, due to its bifunctional scaffolding and kinase signaling properties. Here we report the development of R1-ICR-5, a highly selective and efficacious PROTAC degrader of both human and murine RIPK1. Unexpectedly, acute depletion of RIPK1 leads to abnormal activation of NF-κB/IRF-IFN signaling, and simultaneous induction of necroptosis. In the context of TNFR1 signaling, RIPK1 depletion enables TRADD’s accumulation at the TNFR1, which in turn causes enhanced recruitment of TRAF2, cIAPs and LUBAC. This amplifies Ub-mediated activation of NF-κB/MAPK signaling and cytokine production. Additionally, unhindered accumulation of TRADD also triggers lethal activation of RIPK3. Although this process requires RIPK3’s RHIM domain, it seems to operate independently of RIPK1, TRIF and ZBP1. While our experiments cannot categorically rule out a role for ‘residual’ RIPK1, TNF-induced activation of RIPK3 occurs independent of ‘residual’ RIPK1 kinase activity, as treatment with RIPK1 kinase inhibitors did not impact TNF-induced necroptosis, following R1-ICR-5 treatment. Further, combining RNAi-mediated depletion of RIPK1 with R1-ICR-5 treatment fails to ameliorate TRADD-mediated RIPK3 activation. Therefore, if indeed residual RIPK1 contributes to TNF-mediated RIPK3 activation, it does so in a non-conventional manner. We conclude that PROTAC-mediated depletion of RIPK1 facilitates the formation of TRADD aggregates/polymers that drive RIPK3 activation and necroptosis. Together, our data indicate that RIPK1 licenses tuned activation of TNFR1 and TLRs signaling (TLR3 and TLR4) by functioning as a nucleation barrier for self-templating polymerization of TRADD and TRIF, thereby preventing these signaling hubs from ‘overheating’. Consequently, RIPK1 functions as a universal suppressor of necroptosis triggered by death receptors (this study and^61–64) and PRRs^{15,16,61–64,79,80}.

The combination of enhanced immunostimulatory signaling and necroptosis are hallmark features of "ICD"^8,45,46,81. Targeting ICD has emerged as a promising strategy in cancer therapy, as it simultaneously kills apoptosis-resistant cancer cells, while also stimulating an immune response against tumor cells, potentially enhancing the efficacy and durability of treatments⁴⁶. While RIPK1-PROTACs might be useful to enhance ICD, removal of RIPK1 alone is not sufficient because RIPK1 merely functions as a brake of immunogenic pathways. Therefore, ligands will be required to engage these pathways. Such ligands can be provided by RT and ICB, which inherently cause the production of TNF and IFNs, sensitizing cancer cells to necroptosis when RIPK1 is depleted. Additionally, RT also stimulates the production of CXCL10, an immunostimulatory chemokine that enhances the recruitment and cross-priming of T lymphocytes^82,83. Consistently, we find that RT/R1-ICR-5 reshapes the tumor immune microenvironment, by favoring the recruitment of TNF⁺IFNγ⁺ lymphocytes, exhibiting key features of activation and anti-tumor function (enhanced degranulation and Granzyme-B⁺). Accordingly, RT/R1-ICR-5 achieved durable treatment responses, rendering cured mice resistant to tumor-rechallenge.

While loss of sensitivity to TNF/IFN cytotoxicity perpetuates immune evasion and resistance to immunotherapy^25,84–86, we find that acute depletion of RIPK1 re-sensitizes cancer cells to the effect of TNF/IFN, and in combination with RT, promotes the influx of TNF⁺IFNy⁺ lymphocytes. Correspondingly, stratifying TNBC patients based on tumor RIPK1 and TNF/IFNy-secreting lymphocytes (NK cells) status, revealed that RIPK1-low/NK-high TNBC patients exhibited prolonged overall survival and differentially expressed genes associated with IFN, TNF and regulated cell death signaling, when compared to RIPK1-high/NK-low patients. RT is currently employed to treat around 50% of all cancer patients, and done so with curative intent for head and neck, prostate and lung cancers⁸⁷. RT is usually not an effective treatment for TNBC patients, which are known to face the worst prognosis, compared to other breast cancer sub-types⁸⁸. However, our findings provide potential new therapeutic strategies to improve the effectiveness of RT in the context of TNBC and other cancer types, where RT is not used with curative intent. While the use of immunotherapies has revolutionized the treatment of some cancers, most patients remain refractory to treatment⁸⁹. In concordance with R1-ICR-5’s ability to restore sensitivity to TNF/IFNs, we also provide evidence that RIPK1 PROTACs enhance the efficiency of anti-cancer immunotherapy (ICB). As such, RIPK1 PROTACs could offer a means of dually boosting the number of patients that are eligible for such therapies, while improving the durability of treatment responses.

In summary, we provide pharmacological proof-of-concept that acutely depleting RIPK1 with PROTACs can potentiate the immunostimulatory effects and therapeutic responses caused by RT and ICB. We envisage that targeting RIPK1 with PROTACs may become even more important in the future, as improved PROTAC molecules are advanced into the clinic

Limitation of the study

In this study, we demonstrate the enhanced immunostimulatory and anti-tumor effects of RIPK1-selective PROTACs when used in combination with radiotherapy (RT) and immune checkpoint blockade (ICB). Further, we show that local injection of RIPK1 PROTACs can effectively reduce inflammatory skin lesions. However, to further evaluate RIPK1 degraders as potential therapeutic agents and to assess their systemic tolerability, a tool compound that can sufficiently degrade RIPK1 on systemic dosing is needed. Encouraging results from IHC studies of tumor slices, intra-tumorally treated with R1-ICR-5, show that the degradation of RIPK1 within tumors is localized and incomplete, yet associated with inhibited tumor growth, improved survival, and increased immune cell infiltration when combined with RT and/or ICB. Moreover, our data indicate that the effects of RIPK1 degradation persist for at least 24 hours post compound removal, suggesting that a complete and continuous degradation of RIPK1 may not be essential to initiate an immune response. Nonetheless, a thorough toxicological assessment using a systemically dosed compound in preclinical mouse models will be essential to advance the development of RIPK1 degraders for therapeutic use.

STAR METHODS

Resource availability

Lead contact

Further information and reasonable requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Pascal Meier (pmeier@icr.ac.uk).

Materials availability

Materials generated in this study are available upon reasonable request and will be fulfilled under a material transfer agreement (MTA).

Data and code availability

• Raw RNA-Seq data for the mammary tumor cell line (EO771) have been deposited in the Sequence Read Archive under the accession number PRJNA1094427. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE⁹¹ partner repository with the dataset identifier PXD043560.

• Original immunohistochemistry, microscopy and western blot images have been deposited at Mendeley and are publicly available (DOI: 10.17632/822syk52yw.1). All other raw data reported in this paper will be shared by the lead contact upon reasonable request.

• This paper analyses existing, publicly available data. These accession numbers for the datasets are listed in the key resources table.

• This paper does not report original code.

• Any additional information required to reanalyse the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
α-cleaved-caspase-3	Cell Signaling Technology	Cat #9664S
α-phospho-ERK1/2	Cell Signaling Technology	Cat#9101S
α-FLAG	Cell Signaling Technology	Cat#8146
α-γH2AX	Cell Signaling Technology	Cat#9718P
α-IKKβ	Cell Signaling Technology	Cat#8943S
α-phospho-IKK α/β	Cell Signaling Technology	Cat#2697S
α-ISG15	Cell Signaling Technology	Cat#2743
α-MK2	Cell Signaling Technology	Cat#3042
α-phospho-MK2	Cell Signaling Technology	Cat#3007
α-p38	Cell Signaling Technology	Cat#9212
α-phospho-p38	Cell Signaling Technology	Cat#4511
α-p65	Cell Signaling Technology	Cat#8242s
α-phospho-p65	Cell Signaling Technology	Cat#3033
α-TRADD	Cell Signaling Technology	Cat#3694S
α-TRAF2	Cell Signaling Technology	Cat#4724S
α-phospho-mRIPK3 T231/S232	Cell Signaling Technology	Cat#91702
α-mouse RIPK1 N-terminal	Cell Signaling Technology	Cat#3493
α-human RIPK1 N-terminal	Cell Signaling Technology	Cat#73271
α-mouse RIPK2	Cell Signaling Technology	Cat#4142S
α-mouse RIPK2	Cell Signaling Technology	Cat#4928
α-mouse RIPK3	Cell Signaling Technology	Cat#15828
α-human RIPK3	Cell Signaling Technology	Cat#13526S
α-TrkA	Cell Signaling Technology	Cat#2505S
α-B-raf	Santa Cruz Biotechnology	Cat#sc-5284
α-IkBa	Santa Cruz Biotechnology	Cat#sc-371
α-Ubiquitin	Santa Cruz Biotechnology	Cat#sc-8017
α-phospho-MLKL S345	Abcam	Cat#ab196436
α−TNFR1	Abcam	Cat#19139
α-RIPK1	BD Bioscience	Cat#610459
α-RIPK2	BD Bioscience	Cat#612348
α-MLKL	Millipore	Cat#MABC604
α-ZBP1	Adipogen	Cat#AG-20B0010-C100
α-FADD	Assay designs	Cat#AAM-212
α-HOIP	Bethyl laboratories	Cat#A303–560A
α-Sharpin	Proteintech	Cat#14626–1-AP
α-ERK1/2	Gift from Prof. Chris Marshall	N/A
α-cIAP1	Enzo	Cat#ALX-803–335-C100
α-mouse IgG (H+L)-HRP	Jackson Immuno Research	Cat #115–035-003
α-rabbit IgG (H+L)-HRP	Jackson Immuno Research	Cat#111–035-003
α-rat IgG (H+L)-HRP	Jackson Immuno Research	Cat#112–035-003
Phalloidin Alexa 633	Thermo Fisher Scientific	Cat#A22284
Alexa 488 α-rabbit secondary antibody	Thermo Fisher Scientific	Cat#A-11034
DAPI	Sigma Aldrich	Cat#10236276001
Alexa 488 α-rabbit secondary antibody	Thermo Fisher Scientific	Cat#A21206
α-IgG2a (in vivo)	2B scientific	Cat#BE0085
α-DYKDDDDK tag	Thermo Fisher Scientific	Cat#8146
α-CTLA-4 (in vivo)	2B scientific	Cat#BE0164
α-PD-1 (in vivo)	2B scientific	Cat#BE0146
α-mouse CD107a-BUV395	BD	Cat#565533; Clone#1D4B
α-mouse CD107a-PE	Biolegend	Cat#121611; Clone#1D4B
α-mouse CD44-BUV395	BD	Cat#740215; Clone#IM7
α-mouse CD25-BV650	Biolegend	Cat#102037; Clone#PC61
α-mouse CD4-BV496	BD	Cat#741050; Clone#RM4–4
α-mouse MHCII-PE-Dazzle-594	Biolegend	Cat#107647; Clone#M5/114.15.2
α-mouse CD3-BUV563	BD	Cat#741319; Clone#17A2
α-mouse TCRδ-BUV615	BD	Cat#751183; Clone#GL3
α-mouse TCRδ-PerCpCy5.5	Biolegend	Cat#118117; Clone#GL3
α-mouse CD11c-BUV615	BD	Cat#751222; Clone#N418
α-mouse CD11c-PECy7	Biolegend	Cat#117317; Clone#N418
α-mouse Singlec-F-AF488	Biolegend	Cat#55523; Clone#S17007L
α-mouse Singlec-F-BV605	BD	Cat#121612; Clone#E50–2440
α-mouse CD69-PECy7	Biolegend	Cat#104511; Clone#H1.2F3
α-mouse CD69-BV785	Biolegend	Cat#740388; Clone# H1.2F3
α-mouse CD19-BUV661	BD	Cat#612971; Clone#1D3
α-mouse NK1.10BUV737	BD	Cat#741715; Clone#PK136
α-mouse NK1.1-PECy5	Biolegend	Cat#108715; Clone#PK136
α-mouse CD45-BUV805	BD	Cat#748370; Clone#30F11
α -mouse F4/80-BV650	Biolegend	Cat#123149; Clone#BM8
α-mouse FoxP3-eFlour450	Thermo Fisher Scientific	Cat#17–5773-82; Clone#FJK-16s
α-mouse Ly6G-BV510	Biolegend	Cat#127633; Clone#1A8
α-mouse CD62L-PECy5	Biolegend	Cat#104410; Clone#MEL-14
α-mouse CD8α-BV570	Biolegend	Cat#100739; Clone#53–6.7
α-mouse Ly6C-AF700	Biolegend	Cat#128023; Clone#HK1.4
α-mouse CD11b-BV750	Biolegend	Cat#101267; Clone#M1/70
α-mouse TNFα-APC	Biolegend	Cat#506307; Clone#MP6-XT22
α-mouse Grzmb-AF647	Biolegend	Cat#515405; Clone#GB11
α-mouse IFNγ-AF488	Biolegend	Cat#505815; Clone#HM61.2
α-mouse TCRβ-PerCP-Cy5	Biolegend	Cat#109228; Clone#H57–597
α-mouse CD8α-BUV395	BD Biosciences	Cat#563786; Clone#53–6.7
α-mouse CD4-BV650	Biolegend	Cat#100555; Clone#RM4–5
α-mouse CD44-BV785	Biolegend	Cat#103059; Clone#IM7
α-mouse CD25-EF450	Thermo Fisher Scientific	Cat#48–0251-82; Clone#PC61–5
LIVE/DEAD™ Fixable Near-IR	Thermo Fisher Scientific	Cat#L34976
α-FLAG tag	Cell Singalling Technology	Cat#8146
α-mouse HRP	Abcam	Cat #ab131368
Mouse Fc Block (CD16/CD32)	BD Biosciences	Cat #553141
Bacterial and virus strains
BL21(DE3) E. Coli	New England Bioscience	Cat #C2527
One Shot TOP10 E. Coli	Thermo Fisher Scientific	Cat #C404010
Biological samples
Chemicals, peptides, and recombinant proteins
Propidium Iodide	Sigma-Aldrich	Cat#P4864
Hoechst-33342, Trihydrochloride, Trihydrate	Thermo Fisher Scientific	Cat#H3570
Glutathione Sepharose 4B beads	GE Healthcare	Cat#17075601
Recombinant mouse M-CSF	Peprotech	Cat#300–25
Gentle Collagenase/Hyaluronidase	Stemcell Technology	Cat#07919
RIPK1 inhibitor GlaxoSmithKline GSK’963	Stratech Scientific	Cat#HY-103028A; CAS#:2049868–46-2
RIPK1 inhibitor (PK68)	Insight Bio	Cat#HY-128348; CAS#:2173556–69-7
RIPK3 inhibitor GlaxoSmithKline GSK’872	Stratech Scientific	Cat#S8465; CAS: 1346546–69-7
Necrostatin-1s	MedChemExpress	Cat#HY-103028A; CAS#: 4311–88-0
RIPK2 PROTAC degrader 2	Cambridge Bioscience	Cat#HY-111866–10 mg; CAS#1801547–16-9
SM-164	Caltag Medsystems	Cat#TAR-T12932L; CAS#957135–43-2
ASTX660	Insight Bio	Cat#HY-109565; CAS#1799328–86-1
Recombinant human TNFa	Enzo	Cat#ALX-522–008-C050
Recombinant mouse TNFa	Enzo	Cat# ALX-522–009-C050
Emricasan, pan-caspase inhibitor	Medkoo	Cat#510230; CAS#254750–02-2
Z-VAD-FMK, pan-caspase inhibitor	Apex Bio	Cat#A1902; CAS #187389–52-2
Bortezomib, proteasome inhibitor	Stratech	Cat#S1013; CAS#179324–69-7
VH298, VHL inhibitor	Abcam	Cat#ab230370
Human recombinant INFβ	Peprotech	Cat#300–02BC
Murine recombinant INFβ	Biomol	Cat#RP0487M
Murine recombinant IFNγ	Peprotech	Cat#315–05
Poly(I:C) HMW	Invivogen	Cat#tlr-pic; CAS#31852–29-6
Riboxxol	Riboxx	Cat#A-00102; CAS#63231–63-0
LPS EK Ultrapure	Invivogen	Cat#tlrl-peklps
Enbrel (Etanercept)	Cambridge Bio	Cat#HY-108847–10
5Z-7-Oxozeaenol, TAK1 inhibitor	Bio-techne	Cat#3604
BX795, TBK1/IKKε inhibitor	Invivogen	Cat#tlrl-bx7–2; CAS#1472611–45-2
p38 inhibitor (Ralimetinib LY2228820 dimesylate)	Selleckchem	Cat#862507–23-1; CAS#862507–23-1
PF-3644022 (MK2 inhibitor)	Bio-Techne	Cat#HY-107427; CAS #1276121–88-0
TPCA-1 (IKK2 inhibitor)	Sigma-Aldrich	Cat#T1452; CAS#507475–17-4
NIK SMI1 (NIK inhibitor)	Stratech	Cat#S8941-SEL; CAS#1660114–31-7
Cycloheximide	Tocris	Cat#0970–100; CAS#66–81-9
HOIPIN-8 (HOIP inhibitor)	Axon MedChem	Cat#2972; CAS#2519537–69-8
Tpl2 kinase inhibitor	Insight biotechnology	Cat#HY-12358; CAS#871307–18-5
XIAP degrader	Insight biotechnology	Cat#HY-115865
DL-Dithiothreitol (DTT)	Sigma-Aldrich	Cat#10708984001; CAS#3483-12-3
PR-619 (DUB inhibitor)	2BScientific	Cat#SI-9619–0005
Recombinant human USP2 Catalytic domain protein	Bio-Techne, R&D	Cat#E504050
Anti-FLAG M2 affinity agarose beads	Sigma-Aldrich	Cat#A2220
Strep-Tactin Sepharose resin beads	IBA lifesciences	Cat#2–1201-025
Lenti-X concentrator	Clontech Takara	Cat#621231
Polybrene	Merk	Cat#TR-1003-G
Lipofectamine RNAiMAX Transfection Reagent	Thermo Fisher Scientific	Cat#13778150
Blasticidin	Invivogen	Cat#ant-bl-05
Collagenase type VI	Sigma Aldrich	Cat#7919
Tryton X-100	Thermo Fisher Scientific	Cat#28314
PR-619 DUB inhibitor	2B Scientific	Cat# SI-9619–0005
PhosSTOP easypack	Sigma Aldrich	Cat#4906837001
Complete protease inhibitor cocktail	Roche	Cat# 11836153001
GSH beads	Sigma Aldrich	Cat#GE17–0756-01
Protein G beads	Sigma Aldrich	Cat#P3296
Clean-Blot IP Detection Reagent (HRP)	Thermo Fisher Scientific	Cat#21230
Dispase	Sigma Aldrich	Cat#D4693–1G
DNase type I	Sigma Aldrich	Cat#DN25
Corning Matrigel GFR	SLS	N/A
Ionomycin from Streptomyces Conglobatus	Sigma Aldrich	Cat#I9657
MRT3767, TBK1i	Sigma Aldrich	Cat#SML0702
HALT Protease and Phosphatase Inhibitor Cocktail	Thermo Fisher Scientific	Cat#78447
Dako REAL peroxidase block	Agilent	Cat#S202386–2
Dako rabbit EnVision polymer-HRP	Agilent	Cat#K400311–2
Dako DAB+	Agilent	Cat#K346811–2
Dako FLEX Haematoxylin	Agilent	Cat#K800821–2
Rabbit ImmPRESS polymer-HRP	Vector Laboratories	Cat#MP-7401–50
Dako Target Retrival Solution	Agilent	Cat#K800421–2
Target retrieval solution (TRS)	Agilent	Cat#K800521–2
Critical commercial assays
Effectene Transfection Kit	Qiagen	Cat#301427
Nano-Glo HiBiT Lytic Detection KitP	Promega	Cat#N3030
NaveniFlex MR	Bethyl Laboratories	Cat#NF.MR.100
CellTiter-GLO	Promega	Cat#G9681
RNeasy RNA extraction kit	Qiagen	Cat#74106
QuantiTec Reverse Transcription kit	Qiagen	Cat#205314
Pierce™ BCA protein assay kit	Thermo Fisher Scientific	Cat#23227
Deposited data
Bulk RNA-seq (Sham Vs RT-treated EO771 cells)	Sequence read archive	ID: PRJNA1094427
Mass spectrometry: Human and mouse whole proteome analysis	PRIDE	ID: PXD043560
TCGA Pan-cancer analyses	cBioPortal	N/A
SCAN-B TNBC patient data	GEO Database	ID: GSE96058
GSEA analyses (hallmark gene set collection)	The Molecular Signatures Database (MSigDB)	N/A
Raw western blots and image from this paper	Mendeley	DOI: 10.17632/822syk52yw.1
Experimental models: Cell lines
HEK293T	ATCC	Cat#CRL-3216
HT1080	ATCC	Cat#CCL-121
U937	ATCC	Cat#CRL-3253
HT29	ATCC	Cat#HTB-38
MDA-MB-231	ATCC	Cat#CRM-HTB-26
MDA-MB-361	ATCC	Cat#HTB-27
MDA-MB-468	ATCC	Cat#HTB-132
HaCaT	ATCC	Cat#PCS-200–011
LLC	ATCC	Cat#CL-101
L929	ATCC	Cat#CRL-2148
THP-1	ATCC	Cat#TIB-202
EMT6	ATCC	Cat#CRL-2755
Lim1215	Laboratory of Prof. Chris Marshall	N/A
A431	ATCC	Cat#CRL-1555
F3II	Laboratory of Prof. Daniel Alonso	N/A
D2A1	Laboratory of Prof. Clare Isacke	N/A
Kym-1	Laboratory of Prof. John Silke	N/A
Immortalized MDFs	Charles River (Cells generated in house)	N/A
MCA-205	Sigma Aldrich	Cat#SCC-173
EO771	CH3 Biosystems	Cat#94A001
MC38	Kerafast	Cat#ENH204-FP
Wildtype BMDMs and Lung fibroblasts (C57BL/6)	Charles River (Cells generated in house)	N/A
Spata2-/- BMDMs (C57BL/6)	Laboratory of Prof. Mads Gyrd-Hansen’s	N/A
Trif-/- BMDMs (C57BL/6) (Ticam1-/-)	Laboratory of Prof. Manolis Pasparakis	N/A
Zbp1-/- BMDMs (C57BL/6)	Laboratory of Prof. Manolis Pasparakis	N/A
Trif-/-Zbp1-/- BMDMs and LFs (C57BL/6) (Ticam1-/-Zbp1-/-)	Laboratory of Prof. Manolis Pasparakis	N/A
L929 Tradd -/-	This Manuscript	N/A
L929 Tradd-/- reconstituted with TRADD WT	This Manuscript	N/A
L929 Tradd-/- reconstituted with TRADD DD-only	This Manuscript	N/A
L929 Tradd-/- reconstituted with TRADD Y16A/F18A	This Manuscript	N/A
L929Tradd-/- reconstituted with TRADD C239S	This Manuscript	N/A
L929 Tradd-/- reconstituted with TRADD Call>S	This Manuscript	N/A
L929 GFP-RIPK1ΔKDΔRHIM	This Manuscript	N/A
L929 GFP-RIPK1 DD-only	This Manuscript	N/A
E0771 GFP-RIPK1ΔKDΔRHIM	This Manuscript	N/A
E0771 GFP-RIPK1 DD-only	This Manuscript	N/A
293T-Ripk1-/-	This Manuscript	N/A
E0771-Ripk1-/-	This Manuscript	N/A
E0771-Tnfr1-/-	This Manuscript	N/A
E0771-Mlkl-/-	This Manuscript	N/A
Experimental models: Organisms/strains
C57BL/6-Mlkl-/-	Laboratory of Prof. Henning Walczak	N/A
C57BL/6-Casp8-/-Ripk3-/-	Laboratory of Prof. Henning Walczak	N/A
C57BL/6-Casp8tm1Hed/J	The Jackson Laboratory	RRID:IMSR_JAX: 027002
C57BL/6.Cg-Ndor1Tg(UBC-cre/ERT2)1Ejb/1J	The Jackson Laboratory	RRID:IMSR_JAX: 007001
C57BL/6-Tnfrsf1atm1Imx/J	The Jackson Laboratory	RRID:IMSR_JAX: 003242
C57BL/6-Ifnar1tm1.2Ees/J	The Jackson Laboratory	RRID:IMSR_JAX: 028288
C57BL/6-Ripk1fl/fl	Laboratory of Prof. Manolis Pasparakis	N/A
Oligonucleotides
SMART vector inducible non-targeting (Nt) shRNA	Dharmacon	VSC11502
SMART vector inducible Ripk1 shRNA	Dharmacon	V3SM11253–237881699
SMART vector inducible Trif shRNA (Ticam1)	Dharmacon	V3SM11253–235355483
SMART vector inducible Zbp1 shRNA	Dharmacon	V3SM11253–235538171
See Table S2 for PCR primers, Taqman probes, siRNAs & CRISPR gRNAs
Recombinant DNA
pTRIBZ	This Manuscript	N/A
pTIBZ-Cre	This Manuscript	N/A
pTIBZ-N2xFLAG-mTradd-WT	This Manuscript	N/A
pTIBZ-N2xFLAG-mTradd-DD	This Manuscript	N/A
pTIBZ-N2xFLAG-mTradd-Y16A/F18A	This Manuscript	N/A
pTIBZ-N2xFLAG-mTradd-Call>S	This Manuscript	N/A
pTIBZ-N2xFLAG-mTradd-C239S	This Manuscript	N/A
pMA-RQ-N2xFLAG-mTradd-WT	This Manuscript	Synthetic, Invitrogen
pMA-RQ-N2xFLAG-mTradd-DD	This Manuscript	Synthetic, Invitrogen
pMA-RQ-N2xFLAG-mTradd-Y16A/F18A	This Manuscript	Synthetic, Invitrogen
pMA-RQ-N2xFLAG-mTradd-Call>S	This Manuscript	Synthetic, Invitrogen
pMA-RQ-N2xFLAG-mTradd-C239S	This Manuscript	Synthetic, Invitrogen
pTIBZ-mRipk3-WT	This Manuscript	N/A
pTIBZ-mRipk3-RHIMm	This Manuscript	N/A
pTIBZ-EGFP-mRipk1 ΔKDΔRHIM	This Manuscript	N/A
pTIBZ-EGFP-mRipk1 DD-only	This Manuscript	N/A
pMA-RQ-mRipk1 ΔKDΔRHIM	This Manuscript	Synthetic, Invitrogen
pMA-RQ-mRipk1 DD-only	This Manuscript	Synthetic, Invitrogen
pTRIPZ	Open Biosystems	Cat#RHS4696
pTIPZ-hRipk1 ΔRHIMΔDD-HiBiT	This Manuscript	N/A
pTYB1	New England BioLabs Inc.	Cat#N6701
pTYB1-FLAG-mTradd WT	This Manuscript	N/A
pT7CFE1-CHis	ThermoFisher	88860
T7CFE-mRipk3	This Manuscript	N/A
pcDNA3.1	Invitrogen	Cat#V79520
pcDNA3-GFP-2xHA/2xStrep	This Manuscript	N/A
pcDNA3.5-N2xHA-mRipk1	This Manuscript	N/A
pcDNA3-N3xHA-mZbp1	This Manuscript	N/A
pcDNA3-N3xHA-mRipk3	This Manuscript	N/A
pcDNA3-mTrif-C3xHA (Ticam1)	This Manuscript	N/A
pcDNA3-N2xFLAG-mTradd	This Manuscript	N/A
pLC-EGFP	Gift from B. Bornhauser	Addgene Cat# 75159
pLC-Cherry	Gift from B. Bornhauser	Addgene Cat# 75161
pSpCas9(BB)-2A-GFP	Addgene	48138
pLC-EGFP-mTradd gRNA (CGCAACTGGACGATGAGCTG)	This Manuscript	N/A
pLC-EGFP-mTnfr1 gRNA (CGGACAGTCACTCACCAAGT)	This Manuscript	N/A
pLC-Mlkl gRNA-CGM25 (CTGGCAGCAGGAAGATCGAC)	This Manuscript	N/A
pLC-Ripk1 gRNA-CGM19 (AGGGAACTATTCGCTGGTGA)	This Manuscript	N/A
pMA-hRipk1-gRNA (GCTCGGGCGCCATGTAGTAG)	This Manuscript	Synthetic, Invitrogen
psPAX2	Addgene	Cat# 12260
pMD2.G	Addgene	Cat# 12259
pBABE SV40 (Δ89–93)	Gift from Parmjit Jat (Cotsiki et al, 2004)	https://doi.org/10.1073/pnas.0308006100
Software and algorithms
Dotmatics	Dotmatics	N/A
Harmony High Content Analysis Software	Perkin Elmer	N/A
ImageLab (v6.1.0)	Bio-Rad	N/A
GraphPad Prism (v10.2.1) (339)	Dotmatics	N/A
Adobe Illustrator (v28.0) (2024)	Adobe	N/A
R studio package (v4.3.0)	Limma	N/A
Fiji (v2.9.0)	ImageJ	N/A
FlowJo (v.10)	BD Biosciences	N/A
FACSDIVA software (v6.1.3)	BD Biosciences	N/A
Other

Experimental model and study participant details

Cell lines

293T, HT1080, U937, HT29, MDA-MB-231, MDA-MB-361, MDA-MB-468, HaCaT, LLC, L929, THP-1 and EMT6 were from ATCC. F3II cells (kindly provided by Daniel Alonso), D2A1 cells (kindly provided by Clare Isacke), Kym-1 (kindly provided by John Silke), and MC38 (Kerafast). Aforementioned cell lines were cultured in DMEM. E0771 cells (CH3 BioSystems^™) and MCA-205 (Sigma) were cultured in RPMI 1640 medium (Thermo Fisher Scientific). Culture media were supplemented with 10% FBS (Sigma Aldrich, #A2153) and 1% penicillin/streptomycin.MCA-205 were additionally cultured with 2-mercaptoethanol (50 μm) and MC38s and EO771s were additionally cultured in the presence of Hepes (5mM). All cell lines were maintained at 37°C, in 10% CO₂.

Primary cells

Primary bone marrow-derived macrophages (BMDMs) and mouse dermal fibroblasts (MDFs) of indicated genotypes were isolated as previously described^73,76. In brief, bone marrows were isolated form tibia and femur of 8–12-week-old mice and plated in non-coated Petri dishes and cultured for 6 days in DMEM (+10% FCS and Penicillin and Streptomycin) supplemented with recombinant M-CSF (20 ng/mL) or L929 mouse fibroblast conditioned medium (20%, v/v). Primary MDFs were isolated from the tail of adult 2 months old mice (Laura’s paper). In brief the skin of the tail was cut in small pieces and digested with 3ml of trypsin for 45min at 37°C. Medium was added and digested skin pieces were filtered through a 100μm strainer. The strained medium was placed in 10cm dish and incubated at 37°C/5% CO2 for 2–3 days. Next MDFs were immortalized with pBABE SV40 (large T antigen Δ89–97-missing Bub1⁹²). Lung fibroblasts (LFs) were obtained by homogenizing lungs of 8–12-week-old mice in neat DMEM and 1X Gentle Collagenase/Hyaluronidase, before shaking (2000 rpm) for 45 min at 37°C. Suspensions were filtered (70 μm cell strainer), before pelleting (250G). Suspensions were cultured in DMEM (+10% FBS and 1% Penicillin/Streptomycin). Indicated LFs were immortalized as above with pBABE SV 40. All cells were cultured in DMEM supplemented with 10% FBS (Sigma Aldrich, #A2153) and 1% penicillin/streptomycin and maintained at 37°C, in 10% CO₂. Primary naïve T-cells (CD4⁺ and CD8⁺) cells were ⁹³.

Mice

C57BL/6 Wt (Charles River), UBC-Cre-ER^T2 (referred to as CreER^T2), Casp8^fl/fl, Tnfr1^−/− and Ifnar1^−/− were from The Jackson Laboratory, Mlkl^−/− and Casp8^−/−Ripk3^−/− (Kind gifts from Prof. Henning Walczak) and Ripk1^fl/fl (Kind gifts from Prof. Manolis Pasparakis) (See key resources table for more information). All mice were maintained on C57BL/6 background. For tumor treatments, 5 – 6 weeks old female C57BL/6 mice were purchased from Charles River and were enrolled in the study post an acclimatization period of at least one week. Mice were randomly assigned to treatment groups and studies were performed in a blinded fashion. All animal procedures were conducted within the guidelines of UK Home Office in accordance with Animals (Scientific Procedure) Act (ASPA) 1986, amended 2012 and the institutional guidelines of the Institute of Cancer Research. The Animal Welfare Ethical Review Body (AWERB) reviewed the protocols within the project license.

Method details

PROTAC degrader chemical synthesis

Synthesis of R1-ICR-3 [cyclohexyl (5-(2-((S)-15-((2S,4R)-4-hydroxy-2-((4-(4-methylthiazol-5-yl)benzyl)carbamoyl)pyrrolidine-1-carbonyl)-16,16-dimethyl-13-oxo-4,7,10-trioxa-14-azaheptadecanamido)benzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate]

To a solution of cyclohexyl (5-(2-aminobenzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate (115 mg, 0.30 mmol) and (S)-15-((2S,4R)-4-hydroxy-2-((4-(4-methylthiazol-5-yl)benzyl)carbamoyl)pyrrolidine-1-carbonyl)-16,16-dimethyl-13-oxo-4,7,10-trioxa-14-azaheptadecanoic acid (50 mg, 0.075 mmol) in DMF (5 mL), was added DIPEA (0.039 mL, 0.226 mmol) followed by COMU (64.6 mg, 0.151 mmol). The reaction was heated to 50 °C for 18 hours then purified via HPLC (C18, 35–55% acetonitrile in 10mM ammonium bicarbonate, 0.1% ammonia) to give the title compound as an off white solid (37.7 mg, 0.037 mmol, 48.6%; m/z 1027.4422, expected 1027.4416 for C₅₂H₆₆N₈O₁₀S₂ [M+H]⁺).

Synthesis of R1-ICR-5: cyclohexyl (5-(2-(12-(((S)-1-((2S,4R)-4-hydroxy-2-((4-(4-methylthiazol-5-yl)benzyl)carbamoyl)pyrrolidin-1-yl)-3,3-dimethyl-1-oxobutan-2-yl)amino)-12-oxododecanamido)benzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate

Step 1: cyclohexyl (5-(2-aminobenzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate (0.4 g, 1.05 mmol), 12-(tert-butoxy)-12-oxododecanoic acid (0.31 g, 1.1 mmol) and TCFH (0.88 g, 3.14 mmol) was dissolved in DMF (10 mL) then diluted with acetonitrile (60 mL). 1-Methylimidazole (0.74 mL, 9.4 mmol) was then added and the resulting solution was stirred at 23 °C for 17 h. Volatiles were removed by evaporation under reduced pressure. The resulting mixture was quenched with water (20 mL) and precipitate collected via suction filtration. The solid was redissolved in DCM, dried, filtered and concentrated to yield tert-butyl 12-((6-(5-(((cyclohexyloxy)carbonyl)amino)-6-methylpyridin-3-yl)benzo[d]thiazol-2-yl)amino)-12-oxododecanoate (R1-ICR-5a, 680 mg, 100%, 1.05 mmol; m/z: 651.4).

Step 2: Trifluoroacetic acid (4 mL, 52 mmol) was added to a solution of R1-ICR-5a (680 mg, 1.05 mmol) in DCM (10mL) and the resulting solution was stirred at 23 °C for 1h then concentrated under reduced pressure to yield 12-((6-(5-(((cyclohexyloxy)carbonyl)amino)-6-methylpyridin-3-yl)benzo[d]thiazol-2-yl)amino)-12-oxododecanoic acid (R1-ICR-5b, 621 mg, 100%, 1.04 mmol; m/z: 595.3).

Step 3: HATU (1.19 g, 3.13 mmol) was added to a 0 °C solution of R1-ICR-5b (0.62 g, 1.04 mmol) and DIPEA (1.82 mL, 10.4 mmol) in DMF (10mL) and the resulting mixture was maintained at 0 °C. After 15 min (2S,4R)-1-((S)-2-amino-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (0.49 g, 1.15 mmol) was added and the resulting solution was stirred at 23 °C for 17 h, then cooled to 0 °C before water (50 mL) was added to precipitate the product. The white solid was collected via suction filtration and the filter cake was washed with water (3 × 50 mL), then dissolved in DCM, dried over magnesium sulfate, filtered and concentrated. Sequential purification using Isolute-NH₂ column, reverse phase chromatography (C18, 40–100% methanol in water, 0.1% formic acid) then SCX column gave R1-ICR-5 as a white solid (887 mg, 83%, 0.87 mmol, m/z 1007.4894, expected 1007.4887 for C₅₄H₇₁N₈O₇S₂ [M+H]⁺).

Synthesis of R1-ICR-6 [cyclohexyl (5-(2-(3-(3-(3-(((S)-1-((2S,4R)-4-hydroxy-2-((4-(4-methylthiazol-5-yl)benzyl)carbamoyl)pyrrolidin-1-yl)-3,3-dimethyl-1-oxobutan-2-yl)amino)-3-oxopropoxy)propoxy)propanamido)benzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate]

Step 1: To 3-(3-(3-ethoxy-3-oxopropoxy)propoxy)propanoic acid (33 mg, 0.13 mmol) in DMF (0.5 mL) was added sequentially DIPEA (35 μL, 0.20 mmol) then HATU (55 mg, 0.15 mmol). After 15 minutes cyclohexyl (5-(2-aminobenzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate (45 mg, 0.12 mmol) in DMF (0.7 mL) was added. The resulting mixture was stirred at room temperature for 5 days. Added water (0.5mL) and evaporated under reduced pressure at 50 °C, then purified by reverse-phase flash chromatography (C18, 70–100% methanol in water, 0.1% formic acid) to give 3-(3-(3-((6-(5-(((cyclohexyloxy)carbonyl)amino)-6-methylpyridin-3-yl)benzo[d]thiazol-2-yl)amino)-3-oxopropoxy)propoxy)propanoic acid (R1-ICR-6a, 13.6 mg, 20%, 0.0233 mmol; m/z 585.2375 expected 585.2377 for C₂₉H₃₇N₄O₇S [M+H]⁺)

Step 2: To a mixture of R1-ICR-6a (5 mg, 0.0086 mmol), (2S,4R)-1-((S)-2-amino-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide hydrochloride (4 mg, 0.0086 mmol) in DMF (0.3 mL) at 0 °C under nitrogen was added DIPEA (3 μL, 0.017 mmol) then COMU (4 mg, 0.0093 mmol) in DMF (0.2 mL). The mixture was stirred for 2 days at rt. Additional DIPEA (0.01 mL) and COMU (4 mg) was added and stirred for 3 hours then purified by reverse phase flash chromatography (C18, 10–100% methanol in water, 0.1% formic acid) to give R1-ICR-6 (4 mg, 47%, 0.004 mmol; m/z 997.4341 expected 997.4310 for C₅₁H₆₅N₈O₉S₂ [M+H]⁺)

Synthesis of R1-ICR-7 [cyclohexyl (5-(2-((S)-16-((2S,4R)-4-hydroxy-2-((4-(4-methylthiazol-5-yl)benzyl)carbamoyl)pyrrolidine-1-carbonyl)-17,17-dimethyl-14-oxo-3,6,9,12-tetraoxa-15-azaoctadecanamido)benzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate]

Step 1: To 3,6,9,12-tetraoxatetradecanedioic acid (120 mg, 0.45 mmol) in DMF (0.5 mL) was added sequentially DIPEA (35 μl, 0.20 mmol) then HATU (65 mg, 0.17 mmol). After 15 minutes cyclohexyl (5-(2-aminobenzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate (45 mg, 0.12 mmol) in DMF (0.7 mL) was added. The resulting mixture was stirred at room temperature for 5 days, and was then purified by reverse-phase flash chromatography (C18, 10–100% methanol in water, 0.1% formic acid) to give 14-((6-(5-(((cyclohexyloxy)carbonyl)amino)-6-methylpyridin-3-yl)benzo[d]thiazol-2-yl)amino)-14-oxo-3,6,9,12-tetraoxatetradecanoic acid (R1-ICR-7a, 21.5 mg, 29%, 0.0341 mmol, m/z 631.2434 expected 631.2432 for C₃₀H₃₉N₄O₉S [M+H]⁺).

Step 2: Starting from R1-ICR-7a (5.4mg, 0.0086 mmol), followed method used for synthesis of R1-ICR-6 to obtain R1-ICR-7 (5 mg, 56%, 0.0048 mmol; m/z 1043.4432 expected 1043.4370 for C₅₂H₆₇N₈O₁₁S₂ [M+H]⁺.)

Synthesis of R1-ICR-5S [cyclohexyl (5-(2-(12-(((S)-1-((2S,4S)-4-hydroxy-2-((4-(4-methylthiazol-5-yl)benzyl)carbamoyl)pyrrolidin-1-yl)-3,3-dimethyl-1-oxobutan-2-yl)amino)-12-oxododecanamido)benzo[d]thiazol-6-yl)-2-methylpyridin-3-yl)carbamate]

Prepared from R1-ICR-5b (36 mg, 0.06 mmol) using the same method as for R1-ICR-5, except purified by HPLC (C18, 40–100% methanol in water, 0.1% formic acid) to obtain R1-ICR-5S (24 mg, 39%, 0.0236 mmol; m/z 1007.4899, expected 1007.4887 for C₅₄H₇₁N₈O₇S₂ [M+H]⁺).

Generation of lentiviral particles

To generate lentiviral particles, HEK293T cells were transfected with packaging construct psPAX, pMD2.G and the relevant lentiviral plasmid employing the Effectene transfection kit (301427, Qiagen). Viral supernatants were harvested, passed through a 0.45 μm filter, concentrated with the Lenti-X concentrator (621231, Clontech Takara) and supplemented with Polybrene (10 μg/ml).

Constructs

SMART vector inducible lentiviral shRNAs targeting Trif (Also known as Ticam1, V3SM11253–235355483), Zbp1 (V3SM11253–235538171, Ripk1 (V3SM11253–237881699) and non-targeting (NT) shRNA control (VSC11502) were purchased from Dharmacon. Cells were infected and stable cells were selected with Puromycin (1 μg/ml). Mouse Flag-TRADD constructs were cloned into Dox-inducible lentiviral pTIBZ vector and pool of stable cells was generated after selection with Blasticidin (10 μg/ml). Mouse RIPK3 and RHIM domain mutant of RIPK3 were cloned into the lentiviral vector pTIBZ and stable cell lines were created after infection of transformed LF-Ripk3^−/− and selection with Blasticidin (10 μg/ml). For generation of inducible Ripk1 deletion, immortalized MDFs-Ripk1^fl/fl were infected with lentivirus expressing inducible CRE (pTIBZ-Cre) and stable clones were isolated after Blasticidin (10 μg/ml) selection. Human RIPK1^ΔRHIMΔDD-HiBiT construct was generated by PCR and cloned in lentiviral vector (pTIPZ) and expressed under a Dox-inducible promoter. Stable HT1080 cells were isolated after infection and selection with Puromycin (1 μg/ml). Mouse cDNAs of Ripk1, Ripk3, Tradd, Trif (also known as ticam1) and Zbp1 were cloned in mammalian expression vector pcDNA3 with corresponding tags and sequence verified. Mouse RIPK3 was generated by coupled in vitro transcription/translation reactions using HeLa extracts following synthesis of a PCR fragment which included a 5’ T7 promoter, IRES and kozak sequence and a 3’ poly A tail according to the application notes of the manufacturer’s protocol (Thermofisher Inc.). For bacterial expression, mouse Flag-TRADD was cloned into pTYB1 vector (NEB).

CRISPR gene targeting

Guide RNAs were designed using crispr.mit.edu portal (sequence can be obtained upon request). sgRNA for mouse genes Mlkl, Tnfr1, Tradd and Ripk1 were cloned into pLC-GFP plasmid (plasmid expressing Cas9 and GFP, kind gift from B.C. Bornhauser). For targeting human Ripk1 gene, co-transfection was done with two plasmids, Cas9 plasmid (Addgene #48138) and plasmid expressing sgRNA for human Ripk1 (Invitrogen synthesis). Control cell lines were created by transfection with corresponding Cas9 expressing plasmids without sgRNA. Cells were transfected with electroporation and 72 h after, GFP positive cells were FACS sorted and single clones were screened for gene knockout by western blot for respective proteins. See Key resources table and Supplementary Table S3 for more information.

mTradd gRNA: CGCAACTGGACGATGAGCTG

mRipk1 gRNA: AGGGAACTATTCGCTGGTGA

hRipk1 gRNA: GCTCGGGCGCCATGTAGTAG

mMlkl gRNA: CTGGCAGCAGGAAGATCGAC

mTnfr1 gRNA: CGGACAGTCACTCACCAAGT

RNA interference

All siRNAs were purchased from Qiagen or Dharmacon (product information can be found in supplementary table S3). In brief, transfection of cells was performed in 96 or 384 well plates using RNAiMAX transfection reagent (Thermo Fisher Scientific). Cells were incubated with transfection mix for 18h and medium was replaced. All experiments were done between 48h-72h post transfection.

Bioluminescence-based RIPK1^ΔRHIMΔDD-HiBiT degradation assay

Protein degradation parameters DC₅₀ values (compound concentration at which 50% of RIPK1^ΔRHIMΔDD-HiBiT protein is degraded) and Dmax (maximal degradation level relative to positive control R1-ICR-5) were determined in HT1080 cells (ATCC) stably expressing Dox-inducible RIPK1^ΔRHIMΔDD-HiBiT in a bioluminescence-based assay. Compounds were dispensed in 384 well Cell Culture Microplates (Greiner catalogue 781091) using an ECHO550 acoustic dispenser (Beckman Coulter). HT1080-RIPK1^ΔRHIMΔDD-HiBiT cells were subsequently plated on top of the compounds. Cells were incubated with compounds for 24 h at 37°C/10%CO2. RIPK1^ΔRHIMΔDD-HiBiT was detected using the Nano-Glo HiBiT Lytic Detection Kit (Promega catalogue N3030) according to the manufacturer’s instructions. The % response at each concentration was calculated by normalizing RIPK1^ΔRHIMΔDD-HiBiT expression in the presence of the compound to the appropriate high (DMSO) and low (1.5μM R1-ICR-5) controls. The compound DC₅₀ and Dmax values were determined using Dotmatics (Bishops Stortford, UK) software by fitting the normalized data to a sigmoidal four-parameter logistic (removed “fit”) equation.

Immunofluorescence

Cytoplasmic RIP1 protein degradation parameters were quantified in an immunofluorescence-based assay using an Opera Phenix Plus High Content Imaging System (Perkin Elmer). Briefly, 40 μL of cells cultured in DMEM-10% FBS (Thermo Fisher Scientific) were plated in 384-well Phenoplates (Perkin Elmer) and treated with R1-ICR-5 or DMSO control. After 20 hours at 37°C/CO2 incubator, cells were fixed in 2% formaldehyde at room temperature for 15 mins, then washed in PBS (Phosphate Buffered Saline) using a Multidrop Combi (Thermo Fisher Scientific). Fixed cells were permeabilized for 15 mins at room temperature in 1xPBS/0.2% Triton X-100 (Thermo Fisher Scientific #28314), before 1hr blocking in PBS/0.5% BSA (Sigma Aldrich, A2153). After washing with PBS (Multidrop Combi), cells were incubated overnight at 4°C in PBS/BSA with anti-hRIPK1 (Cell Signaling Technology, #73271) or anti-mRIPK1 (Cell Signaling Technology, #3493) primary antibodies, in the presence of DAPI (Thermo Fisher Scientific, D3571) and Phalloidin-Alexa 633 (Thermo Fisher Scientific, A22284). After washing with PBS, cells were incubated for 1 h at room temperature in PBS/BSA plus donkey anti-rabbit Alexa 488 (Thermo Fisher Scientific, A21206). To visualize TRADD fibrillation, L929^{Dox-FlagTRADD} cells were plated as described above, and protein was induced with 100ng/ml Dox 1h prior to treatment. TRADD was visualized using anti-DYKDDDDK tag (Cell Signaling Technology, #8146) and a Leica SP8 Point Scanning Confocal microscope equipped with a 63X objective. Quantification of TRADD aggregates was performed using Harmony High Content Imaging and Analysis software (Perkin Elmer).

Proximity ligation assay (PLA) and co-localisation assay

L929 ^{Dox-Flag-TRADD} cells were treated overnight with R1-ICR-5 (1 μM). The next day, cells were treated with Doxycycline (100 ng/ml) for 2h and TNF (10 ng/ml) for 30min. Cells were fixed with 2% paraformaldehyde (Santa Cruz, # sc-281692) for 20 min, permeabilized with 0.5 % Triton x-100 for 5 min and PLA was performed according to the manufacture’s protocol, using the NaveniFlex MR (Navinci Diagnostics AB, USA) kit. Primary antibodies against RIPK3 (Cell Signaling Technologies, D8J3L) and FLAG-tag (Cell Signaling Technologies, 8146) were used at a dilution of 1:500 and 1:1000 respectively, and cells were counter-stained with DAPI. Z-stacks of six to nine fields per well were acquired in the ImageXpress Micro Confocal (Molecular Devices) with an CFI Plan Fluor 40x (0.75NA) objective and analyzed with a custom module editor within MetaXpress (Molecular Devices). Each dot on the graph represents the total number of PLA foci (sum intensity projection of the FITC channel) divided by the total number of nuclei (maximum intensity projection of the DAPI channel) per field of a representative experiment of two independent repeats. Representative images of one experiment are shown as an overlay of the sum intensity projection obtained on the FITC channel and the maximum intensity projection obtained on the DAPI channel. Scale bars represent 20 μm.

Caspase activity and cell viability assays

Caspase activity was measured by DEVDase assay and was performed as previously described⁹⁴. In brief, 1.0×10⁴ cells were plated in 96-well plates and after treatment, medium was removed, and plates were placed at −80C for few hours. Next, 20μl of DISC lysis buffer was added to each well and after 20min on ice, 180μl of DEVDase assay mix (add info) was added. Plates were incubated at room temperature for several hours and DEVDase activity was read at 380nM excitation/460nM emission. Cell viability was measured by CellTiter-GLO assay (Promega) and was performed according to the manufacturer’s instructions. In brief, cells were plated in 96 well plates and treated with indicated conditions and cell survival was measured using Victor X plate reader (PerkinElmer). Drug concentrations used in both assays: Emricasan (5 μM), RIPK1 inhibitors PK68 or GSK’963 (100 nM), R1-ICR-5 (1 μM) SMAC mimetic (SM164, 100 nM) and TNF (10 ng/ml).

In vitro treatments (Cell death & signaling assays)

To measure cell death, cells were seeded in 96 well plates (#655090, Greiner) or 384 well plates (#781091, Greiner), before indicated treatments for specified times. Hoechst (0.5 μg/ml) and PI (1 μg/ml) were added and the % of dead cells was measured using the Celigo S Cell Imaging Cytometer (Nexcelom Bioscience). Cell death measurement of Naïve CD4⁺ and CD8⁺ T cells was performed as previously described⁹³. Drug concentrations used: Cycloheximide (25 μg/mL), Doxycycline (100 ng/mL), Emricasan (1–5 μM), Enbrel (50 μg/mL), HOIPIN-8 (30 μM), mIFNβ/γ (100 ng/mL), hIFNβ (25 ng/mL), IKK2 inhibitor (TPCA-1, 250 nM), LPS (0.1–10 μg/mL), MK2 inhibitor (PF3644022, 250 nM), NIK1 inhibitor (NIK-SMI1, 1 μM), Poly(I:C) (10 μg/mL), P38 inhibitor (LY2228820, 250 nM), RIPK1 inhibitors (Nec1, PK68 and GSK’693: 100 nM), RIPK3 inhibitor (GSK’872, 10 μM), RIPK2 PROTAC (1 μM), R1-ICR-3 (1 μM), R1-ICR-5S (Non-VHL binder, 1 μM), R1-ICR-5 (Murine cells, 1 μM; Human cells, 200 nM), TAK1i (5Z-7-Oxozeaenol, 1 μM), TNFα (10 ng/mL), TPL2 inhibitor (Cot inhibitor-2, 1 μM), XIAP degrader-1 (10 μM), and zVAD (20 μM). To analyze TNFR1 and TLR3-induced inflammatory and necroptotic signaling, cells were seeded into 6-well plates overnight, before treatment and subsequent western blot or RT-qPCR analysis, as indicated. Drug concentrations used: Doxycycline (100 ng/mL), Emricasan (1 μM), Enbrel (50 μg/mL), MK2 inhibitor (PF3644022, 500 nM), SMAC mimetic (ASTX660, 1 μM), TNFα (10 ng/mL), Poly(I:C) (10 μg/mL), Riboxxol (10 μg/mL), hIFNβ (25 ng/mL), RIPK2 PROTAC (1 μM), R1-ICR-5 (Murine cells, 1 μM; Human cells, 200 nM).

In vitro irradiation

Cells were plated overnight before irradiation (AGO 250 kV X-ray machine, 1.62 Gy/min, single doses of 4 or 8 Gy). Enbrel (50μg/mL) was added 30 minutes before irradiation. R1-ICR-5 (1 μM) was added 24 hours after irradiation. Doxycycline (Dox)-inducible shNt, shTrif (Also known as Ticam1) and shZbp1-expressing EO771 cells were Dox-induced 36 h before irradiation. Cell death was measured 48 – 72 h after irradiation. To measure DNA damage by immunofluorescence, EO771 cells (2 × 10³) were seeded overnight in a 96 well plate (#655090, Greiner) and subjected to single doses of irradiation, as indicated. After 30 minutes, cells were washed (PBS) and fixed (4% paraformaldehyde, 10 minutes). Cells were washed three times (PBS), incubated for 1 hour with blocking solution (0.3% Triton X100, 5% normal goat serum, 1% BSA in PBS) and then with α-γ-H2AX overnight at 4°C. Cells were washed three times with PBS and incubated with the α-rabbit secondary antibody and DAPI for 1 hour at room temperature. Cells were washed and imaged by a Zeiss 710 confocal microscope (Zeiss 40X objective). The average γ-H2AX intensity per nucleus was quantified by Fiji software. For bulk RNA-seq, EO771 cells were seeded overnight before exposure to a single fraction of 8 Gy. 48 hours later, total RNA was extracted using RNeasy kit (Qiagen).

RT-qPCR

RNA was extracted from cells using RNeasy (Qiagen) and converted to cDNA using QuantiTec Reverse Transcription Kit (Qiagen), according to manufacturer’s instructions. RT-qPCR and data analysis was performed as previously described⁹⁵. For EO771 cells, mRNA was extracted 48 h after irradiation (0, 8 Gy). The relative mRNA expression of indicated genes was measured using Taqman probes (See Key resources table and supplementary table S3 for more details).

Mass Spectrometry

Sample preparation and TMT labelling - For total proteome analysis, U937 cells were treated with DMSO or R1-ICR-3 (100nM) for 6 h by trypsin digestion and TMTpro 15-plex labelling⁹⁶. For total proteome analysis of BMDMs, cells were treated with DMSO or R1-ICR-05 (1μM) for 5 h. Cell lysates were digested⁹⁷ and peptides were labelled with TMTpro 12-plex according to manufacturer’s instructions. Basic reverse-phase peptide fractionation and LC-MS analysis - The TMTpro labelled peptides were fractionated with high-pH Reversed-Phase (RP) chromatography using the XBridge C18 column (2.1 × 150 mm, 3.5 μm, Waters) on a Dionex UltiMate 3000 HPLC system. LC-MS analysis was performed on a Dionex UltiMate 3000 UHPLC system coupled with the Orbitrap Lumos Mass Spectrometer (Thermo Fisher Scientific). Peptides were loaded onto the Acclaim PepMap 100, 100 μm × 2 cm C18, 5 μm, trapping column at flow rate 10 μL/min and analyzed with an Acclaim PepMap (75 μm × 50 cm, 2 μm, 100 Å) C18 capillary column connected to a stainless-steel emitter on an EASY-Spray source via a PSS2 adapter (MSWIL). Mobile phase A was 0.1% formic acid and mobile phase B was 80% acetonitrile, 0.1% formic acid. For total proteome analyses, a 95 min gradient 5%-38% B was used. MS scans were acquired in the range of 375–1,500 m/z with mass resolution of 120K, AGC 4×10⁵ and max IT 50 ms. Precursors were selected with the top speed mode in 3 sec cycles and isolated for HCD fragmentation with quadrupole isolation width 0.9 Th (IMAC) or 0.7 Th, collision energy at 36% at 50K or 30K resolution. Targeted precursors were dynamically excluded for further fragmentation for 30 seconds or 45 sec with 7 ppm mass tolerance. Database search and protein quantification - The mass spectra were analyzed in Proteome Discoverer 2.4 (Thermo Scientific) with the SequestHT search engine for peptide identification and quantification. The precursor and fragment ion mass tolerances were 20 ppm and 0.02 Da respectively. Spectra were searched for fully tryptic peptides with maximum 2 missed cleavages. TMTpro at N-terminus/K and Carbamidomethyl at C were selected as static modifications. Spectra were searched against reviewed UniProt Mus musculus (BMDMs) or Homo Sapiens (U937 cells) protein entries, peptide confidence was estimated with the Percolator node and peptides were filtered at q-value<0.01 based on target-decoy database search. The reporter ion quantifier node included a TMTpro quantification method with an integration window tolerance of 15 ppm. Only peptides with average reporter signal-to-noise>3 were used for quantification. Differential protein expression analysis was performed using the R package limma (v3.50.1), R version 4.1.0.

Purification of TNFR1 complex I & II

The composition of TNFR1 signaling complex I and -II was analysed as previously described^98,99. In short, complex I was purified from cells using FLAG-hTNF (800 ng/mL) or STREP-hTNF (1 μg/mL). Cells were lysed in DISC buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 10% glycerol) supplemented with cOmplete protease (Roche) and phosSTOP phosphatase (Sigma) inhibitors, and PR619 DUB inhibitor (2B Scientific, 10 μM). Lysates were harvested and clarified by rotating (20 minutes) then centrifuging (15 minutes, 14,000 rpm) at 4 °C. Untreated control samples were treated with 800ng/mL 3xFLAG-TNF or 1 μg/mL 2xSTREP-TNF, post-lysis. Complex I was purified overnight, by incubating lysates with anti-FLAG M2 (Sigma) or Strep-Tactin (IBA lifesciences) beads. Immunocomplexes were washed with DISC buffer supplemented with PR619 (10 μM), before eluting with 1X SDS sample buffer. For USP2 digestion, after the final wash, beads were additionally washed with DUB reaction buffer (50mM Tris pH 7.5 and 150nM NaCl). 20 μL DUB reaction buffer (50mM Tris pH 7.5 and 150nM NaCl) + 5mM DTT + 2uM USP2) was added to the dried beads and digestion was performed for 1 h at 37°C. Complex II was purified from cell lines harbouring a Dox-inducible FLAG-TRADD transgene, which was induced (L929^{Dox-Flag-TRADD}, 1h; EO771^{Dox-Flag-TRADD}, 2 h) before TE treatment (L929^{Dox-Flag-TRADD}, 1h; EO771^{Dox-Flag-TRADD}, 2 h). Complex II was purified from cell lysates, with anti-FLAG M2 beads (Sigma) overnight at 4°C, before washing with lysis buffer supplemented with PR619 DUB inhibitor (10 μM) and eluting with SDS sample buffer. Samples were analysed by SDS-PAGE.

Ubiquitin pulldown (TUBE assay)

Stimulation-induced ubiquitylation of indicated proteins was assessed by total ubiquitin pulldown, using GST-TUBE (Tandem Ubiquitin Binding Entities), as previously described⁷³. In brief, after treatments cells were washed with ice cold PBS and lysed with DISC lysis buffer (50mM Tris pH 7.8, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, and 10% glycerol) supplemented protease and phosphatase inhibitors (as in complex-I/II), 1 mM DTT, PR619 (10 μM) and GST-TUBE (50 μg/ml; 50 μg TUBE/mg protein lysate). Lysates were cleared by centrifugation (4°C, 16000 g, 15 min) and immunocomplexes were purified by overnight rotation at 4 °C, in the presence of 20μL GSH beads (Sigma). Beads were washed four times with TUBE wash buffer (PBS and 1% Triton X-100) supplemented with PR619 (10 μM). To visualize higher order proteins, beads were split in two prior to elution, and incubated with either reducing or non-reducing sample buffer (with or without β-mercaptoethanol, respectively) to enable higher order oligomers to be visualised, by SDS-PAGE.

Immunoprecipitation and immunoblot analysis

Co-purification was conducted as previously described¹⁰⁰. In brief, cells were lysed in 1% Triton lysis buffer (50mM Tris pH7.5, 150mM NaCl, 1% Triton X-100 and protease/phosphate inhibitors (Roche). Lysates were centrifuged (16000 g, 20 min) and supernatants were incubated with either α-HA or α-FLAG beads (SIGMA). Subsequently, beads were washed three times with IPPG150 wash buffer (0.1% Triton X-100, 50 mM Tris pH 7.5, 150 mM NaCl,). Beads were boiled in Laemmli buffer for 10 min and proteins were resolved by SDS-PAGE.

In vitro binding assay

N-terminal FLAG-tagged murine TRADD was amplified by PCR using pcDNA3-FLAG-mTradd as a template and subcloned in frame with chitin binding protein using the NdeI and EcoRI sites of pTYB1 (New England BioLabs Inc.). Transformed E. Coli BL21(DE3) cells were induced overnight at 15°C with 0.1 mM IPTG to express FLAG-mTRADD-CBP fusion protein. FLAG-mTRADD was subsequently isolated by sequential purification using FLAG (M2) agarose and chitin beads followed by Intein-mediated cleavage/elution with Intein Lysis Buffer containing 50 mM DTT. Protein purity was assessed by silver staining and identity confirmed by immunoblotting with both FLAG (M2) mAb and TRADD antibodies (Cell Signaling Technology). mRIPK3 was generated by coupled in vitro transcription/translation reactions using HeLa extracts following synthesis of a PCR fragment which included a 5’ T7 promoter, IRES and kozak sequence and a 3’ poly A tail according to the application notes of the manufacturer’s protocol (Thermofisher Inc.). Purified FLAG-mTRADD and mRIPK3 were incubated with either anti-FLAG (M2) conjugated agarose or alternatively, anti-RIPK3 (Cat # 15828, Cell Signaling Technology) and Protein G dynabeads in 30 mM TrisHCl, pH7.4, 150 mM NaCl, 10% glycerol, 1 % Triton X-100 overnight at 4°C. Beads were washed with 4 × 1 ml buffer and aliquots resolved on 10 % precast TGX gels (Bio-Rad Inc.). Immunoblotting was performed using FLAG (F3165) mAb and rabbit anti-RIPK3 (Cat # 15828, Cell Signaling Technology) combined with either anti-mouse-HRP (veriblot ab131368, Abcam) or clean-blot IP detection reagent (#21230, Thermofisher) respectively. See supplementary Table S3 for primers used in PCR reactions.

Size exclusion chromatography

FLAG-mTRADD (50 ml) expressed and purified from E. Coli as described was resolved on a Superdex 200 increase 3.2/300 size-exclusion column in 30 mM TrisHCl, pH7.4, 150 mM NaCl, 10 % glycerol, 1% (w/v) Triton X-100 using an AKTA Pure protein purification system (Cytiva) and 100 ml fractions collected, essentially as described previously¹⁰¹. Aliquots from the indicated fractions, denatured in Laemmli sample buffer, were resolved on 18-well Criterion TGX gels and immunoblotted with FLAG (M2) mAb. The column was calibrated using a high molecular weight gel filtration calibration kit (Cytiva # 28403842).

Western blot analysis

Western blot analyses were performed as previously described⁷³. Briefly, cells were lysed with DISC lysis buffer (1 M Tris pH 7.8, 5 M NaCl, 0.5 M EDTA, 40% Glycerol and 1% Triton X-100) supplemented with protease inhibitor cocktail and phosphatase inhibitors before clearing by centrifugation and quantification by Pierce^™ BCA protein assay kit (ThermoFisher). Whole cell lysates were harvested with 1X SDS sample buffer (6x sample buffer: 3 ml 20% SDS, 4 ml 100% glycerol, 3 ml β-ME and 1 g bromophenol blue). To assess RIPK1 depletion in vivo, tumor and skin samples were collected 24 h after intra-tumoral or subcutaneous injection. Single cell suspensions were prepared from tumors as described below and from skin by homogenization, using 2.4mm bead mill tubes (ThermoFisher) and the Precellys evolution tissue homogenizer (Bertin Technologies). Lysates and homogenates were collected with RIPA buffer (10mM Tris(-HCL) pH8, 1mM EDTA, 0.5 mM EGTA, 1 % Triton, 0.1 % Sodium deoxycholate, 0.1 % SDS & 140 nM NaCl) + HALT protease & phosphatase inhibitor cocktail (ThermoFisher) before centrifugation (20000 xg, 20 mins at 4°C). All samples were boiled with Laemmli buffer (100°C, 10 min), then resolved by SDS-PAGE (NuPAGE Novex 4–12% Bis-Tris 1.0 mm gels in MOPS buffer (ThermoFisher)), transferred onto polyvinylidene difluoride (PVDF) membranes, blocked (5% BSA or milk in TBST) and probed with indicated antibodies (See key resource table). All primary antibodies were diluted in 5% BSA-TBST + 0.01% NaN₃ and secondary antibodies diluted 1:10000 in 5 % non-fat milk-TBST. Densitometry was performed using ImageJ software.

Flow cytometry

Mammary tumors were harvested in ice-cold PBS from mice 13 days after irradiation. Tumors were mechanically dissociated with scissors and enzymatically digested (30 min, 37°C), in PBS containing Trypsin-Versene, 1 mg/ml collagenase type VI (Sigma-Aldrich), 100 μg/ml dispase (Sigma-Aldrich), 1 mg/ml DNase type I (Roche). Tumor suspensions were passed through a cell strainer (70 μm) into PBS (2% FBS & 2 mM EDTA), before centrifugation (1200 rpm, 5 min, 4°C). Pellets were resuspended in PBS-FBS 2% + Fc block (CD16/CD32) for 10 min (4°C), before surface staining (30 min, 4°C). To detect intracellular cytokines and surface CD107a, cells were restimulated in IMDM containing Golgi Plug (BD Biosciences), 100 ng/ml PMA (Thermo Fisher Scientific) and 1 μg/ml ionomycin (Thermo Fisher Scientific) ( 4 h at 37°C). After, cells were fixed by IC fixation buffer (Thermo Fisher Scientific), permeabilised with permeabilization buffer (Thermo Fisher Scientific) and stained with α-IFN-γ, α-TNFα, α-Granzyme B, α-Foxp3 and α-Ki67 antibodies. Flow cytometric analyses were carried out with the LSR II FACSymphony A5 (BD Biosciences) with FACSDiva software. Data were analysed with FlowJo V.10 software. The full list of antibodies used in this study can be found in the key resources table.

In vivo tumor treatments

For tumor treatments, 5 – 6 weeks old female C57BL/6J mice were purchased from Charles River and were enrolled in the study post an acclimatization period of at least one week. Mice were injected in the 4^th right mammary fat pad with 1×10⁵ EO771 cells in RPMI supplemented with 10% FBS, 5 mM Hepes, penicillin/streptomycin, and containing 50% Matrigel matrix (Corning). Tumor growth was monitored twice a week using a digital caliper, and volumes (0.52 × length × width × width) were expressed in mm³. Mice were culled before the average tumor size reached 15 mm. Tumor treatment was started when tumors reached an average of 50 – 100 mm³. Tumors were irradiated with the Small Animal Radiotherapy Research Platform (SARRP, Xstrahl), employing a parallel opposed beam arrangement with a 10 mm × 10 mm collimator for each beam. In all experiments, beams were equally weighted, and the radiation dosage was 4 or 8 Gy, administered in a single fraction. R1-ICR-5 was solubilized in captisol (Ligand’s Pharmaceutical) and administered by intra-tumoral injection (40 μg/injection) at indicated intervals (See Fig. 6,7 and Suppl. Fig. S6). Anti-IgG2a, anti-CTLA-4 and anti-PD-1 antibodies were injected intraperitoneally (200μg) at indicated intervals (See Fig. 6 & 7). (Sourced from 2B Scientific, see key resources table). Response rates of treated mice were classified as complete response (tumor free for at least 2 months), partial response (tumor volume drops to <100 mm³) and progression (no response to treatment). P values were calculated using the Chi-square test, comparing progression (those that progressed after treatment) against response (those that showed both complete and partial responses to treatment). To assess immune memory, mice that were tumor free for at least two months (cured mice) were re-challenged together with naïve mice by injecting EO771 in the 4^th left mammary fat pad, as described above. To measure RIPK1 depletion by WB (24h after a single injection) and IHC analysis, R1-ICR-5 was injected intra-tumorally on days 0, 2, 4 and 6 and samples were collected 24 hours post final injection. To measure CASP3 cleavage by IHC, EO771 tumors were irradiated and treated with R1-ICR-5 every 3 days as described above. Tumors were harvested 13 days after irradiation and fixed before staining.

Chemical and genetic skin inflammation models

The skin injection experiments were performed as previously described⁷³. Briefly, mice were injected subcutaneously in the flank with 100 μL of SM (ASTX660, 3 mM) and E (Emricasan, 1 mM) or TNF (1 μg) and TBK1i (MRT67307, 150 μg), combined with R1-ICR-3 (10 mM), R1-ICR-5 (10 mM) or vehicle. Mice were culled 3 days after the injection and the regions of injections were macroscopically assessed. Hair removal cream was used to remove the fur in the lesion area before taking a skin biopsy. Biopsies were processed as previously reported⁷³. Genetically induced skin inflammation models employed UBC-CreRT^T2 mice to locally induce loss of Casp8^fl/fl, Ripk1^fl/fl or Casp8^fl/fl/Ripk1^fl/fl on the nape, by topical treatment with 4-OHT (250 μg per dose) on d0 and d3. Skin biopsies were harvested on d23 and were processed and analyzed as below.

Immunohistochemistry

For IHC analyses, staining was carried on the Dako Autostainer Link48 platform (Agilent Technologies) with epitope retrieval carried out using the Dako PT Link module at 97°c for 20 minutes using Dako Target Retrieval Solution pH9 (Agilent, K800421–2) for RIPK1 and TRS pH6 (Agilent, K800521–2) for cleaved CASP3, according to manufacturer's instructions. Endogenous peroxidases were blocked using Dako REAL peroxidase block (Agilent, S202386–2) then primary antibody RIPK1 (D94C12, Cell Signaling #3493) was diluted 1/50 and applied (60min at room temperature and detected using Dako rabbit EnVision polymer-HRP (Agilent,K400311–2)). The reaction was visualized using Dako DAB+ (Agilent, K346811–2) and nuclei counterstained using Dako FLEX haematoxylin (Agilent, K800821–2) according to manufacturer's instructions. CASP 3 cleavage was measured with anti-Cleaved Caspase 3 (ASP175) primary antibody (Cell Signaling #9664, clone 5A1E, 1/100 dilution). Detection using Rabbit ImmPRESS polymer-HRP (Vector Laboratories, MP-7401–50).

Quantification and statistical analysis

Survival analyses

For TCGA patient data analysis, RIPK1 DNA copy number and RNA-seq data from ACC, BRCA, CESC, STAD, LIHC, LUSC, SKCM, SARC, and UCEC patients of TCGA were downloaded from cBioPortal (https://www.cbioportal.org/). To analyse the SCAN-B patient data, we obtained RNA-sequencing and patient survival data from the GEO database using the accession ID GSE96058. The R package org.Hs.eg.db (v3.7.0) was utilized to map HGNC gene symbols to EntrezIDs. Survival differences among groups were evaluated using the Kaplan-Meier method, and the p-value was determined using the log-rank test, using the R package survival (v3.5.5).

Gene set enrichment analyses (GSEA)

Using the ConsensusTME R package (v0.0.1.9000) on the SCAN-B dataset, patient groups were categorized by assessing the expression differences of RIPK1 and the levels of NK cell infiltration. Gene Set Enrichment Analysis (GSEA)¹⁰² was then carried out on these groups to investigate the differential expression of immune response gene sets. These gene sets were obtained from the hallmark gene set collection of The Molecular Signatures Database (MSigDB)¹⁰³.

Bulk RNA-seq analysis

Total RNA sequencing was performed by Azenta (Genewiz), employing Illumina NovaSeq, 2 × 150 base pairs as sequencing configuration. Samples and genes were clustered using agglomerative hierarchical clustering with '1-Pearson correlation coefficienť as distance measure followed by complete-linkage clustering.

Histopathological assessment of murine skin samples

Histopathological sample analysis was carried out as previously described⁷³. Skin sections were scanned at x40 (0.23 μm/pixel) using a NanoZoomer-XR Hamamatsu Photonics (Japan). The analysis of digitized skin samples involved assessment of the vicissitudes in epidermal and dermal region of the skin identifiable on haematoxylin and eosin (H&E) stained samples. First, a Histopathological (multivariate) Lesion Score (HLS) was used to assess the proportion and severity of histopathological changes. This assessment was performed on the entire length of the skin sample (epidermal and immediate dermal level). To increase precision of the assessment, each sample was divided using a grid with equal size of fields of view (FOV) (0.04mm²/800μm perimeter). Each FOV was assessed for presence of regular epidermis or any pathological changes to regular epidermis and dermis. A final calculation took into account the proportion [%] of given skin lesions within each sample, multiplied by a power score ranging from 0 to 4. Score 0 was associated with regular epidermis with no changes to any of the strata. Score 1 with thickening of the epidermis represented by any of the strata. Score 2 with epidermal erosion (partial loss of the epidermis), with the stratum basale left intact. Score 3 with ulcer-loss of epidermis, including the stratum basale. Score 4 with deep ulceration of the epidermis and subepidermal dermis. The final score (HLS) indicates the severity following the formula: [(0x % Score 0) + (1× % Score 1) + (2× % Score 2) + (3× % Score 3) + (4x % Score 4)]. The lowest possible HLS is ‘0’ which is equal to 100% of normal/regular epidermis in the whole sample, and a maximum HLS of ‘400’ equal to 100% of deep ulceration.

Statistical analysis

Graphs and statistical analyses were generated and performed using GraphPad Prism V9.4.1. The statistical analysis performed for each data set is described in the corresponding figure legend. Error bars indicate standard deviation (SD). No data were excluded.

Highlights.

• RIPK1’s scaffolding function protects cancer cells from immune detection and death

• RIPK1-specific PROTAC degraders enhance innate immune signalling and necroptosis

• RIPK1 PROTACs boost the immunostimulatory and anti-tumour effects of RT and ICB

• PROTAC-mediated RIPK1 depletion suppresses skin inflammation