Transcriptional remodeling shapes therapeutic vulnerability to necroptosis in acute lymphoblastic leukemia Open Access

Despite considerable advances in pediatric acute lymphoblastic leukemia (ALL) treatment,¹ relapses still represent a clinical challenge with poor prognosis and dismal outcomes.² One potential source of treatment failure is apoptosis evasion, owing to insufficient mitochondrial priming.³ Apart from approaches to decrease the threshold for mitochondrial apoptosis,⁴ induction of nonapoptotic cell death has emerged as attractive strategy,⁵ to elicit cell death even under apoptosis-insensitive conditions.⁶ Receptor-interacting protein kinase 1 (RIPK1), a central regulator of cell survival, inflammation, and death,⁷ mediates the nonapoptotic cell death mechanism necroptosis, activating RIPK3 and mixed lineage kinase domain-like pseudokinase (MLKL). Further to the regulation of distinct RIPK1-governed protein-protein interactions and phosphorylation events,⁸ the regulatory circuitries upon necroptosis induction remain to be elucidated. Small-molecule second mitochondria-derived activators of caspase (SMAC) mimetics (SM) potently induce RIPK1-dependent necroptosis in primary human ALL downstream depletion of inhibitor of apoptosis proteins (IAPs),⁹ revealing a distinct vulnerability that relies on tumor necrosis factor (TNF) receptor 1 and 2 signaling, yet is independent of TNF-α.¹⁰ Key features such as posttranslational modification of RIPK1, activation status of caspase-8, RIPK3 homodimerization and MLKL phosphorylation status, or ESCRT (endosomal sorting complex required for transport)–mediated membrane repair contribute to decisions to turn cells into a necroptotic state.^11,12 However, these mechanisms alone cannot explain the potential of cells to undergo necroptosis, and further regulatory events are required for cell death execution downstream of RIPK1. Several in-human studies have demonstrated the tolerability of the bivalent SM birinapant (reviewed in a previous study¹³), yet displaying limited single-agent activity. This illustrates the critical need for combinatorial treatment strategies, as exemplified by the synergistic activity of SM with anti-CD3 immunotherapy in T-cell ALL.¹⁴ Here, we investigate the regulation of RIPK1-dependent cell death in pediatric ALL by combining functional drug response profiling (DRP), transcriptomics, and CRISPR-based genome editing. Integrative analyses uncover the importance of transcriptional activity driven by the master regulators specificity protein 1 (SP1), p300, and histone deacetylase 2 (HDAC2) in necroptosis activation and nominate a combination of SM with HDAC inhibition as a potential antileukemic strategy to augment RIPK1-dependent necroptosis and counteract apoptosis evasion.

Primary human ALL samples originated from cryopreserved bone marrow (BM) aspirates of patients from the AIEOP-BFM 2000, the AIEOP-BFM 2009, and the ALL-REZ 2002 studies. Informed consent was given in accordance with the principles of the Declaration of Helsinki. Approval was granted by the ethics commission of the Kanton Zürich (approval number 2014-0383). Patient characteristics are presented in supplemental Table 3, available on the Blood website.

In vivo experiments were approved by the veterinary office of the Kanton of Zürich. For generating patient-derived xenografts (PDXs), primary human ALL cells were transplanted into 5- to 12-week-old immunodeficient NOD/SCID/IL2rγnull (NSG) mice via tail vein injection. To assess drug activity, engrafted mice were treated with vehicle, birinapant (15 mg/kg, or 5 mg/kg), mocetinostat (Mo; 25 mg/kg), or the combination of birinapant (15 mg/kg) and Mo (25 mg/kg), via intraperitoneal injections, 5 times per week for 3 weeks. Leukemia progression was monitored by staining peripheral blood with ɑ-hCD19-PE (catalog no. 302208, BioLegend), ɑ-hCD45-Alexa Fluor 647 (catalog no. 304018, BioLegend), and ɑ-mCD45-eFluor 450 (catalog no. 48-0451-82, Invitrogen), analyzed by flow cytometry (LSRFortessa, BD Biosciences) and FlowJo software (10.0 8rl). Cells transduced with RFP657-containing plasmids were stained with ɑ-hCD19-PE and ɑ-mCD45-eFluor 450. Engrafted ALL cells were collected from the spleen and BM and analyzed using the same markers. Animals were randomized into treatment arms. No animals were excluded from analyses.

Human telomerase reverse transcriptase mesenchymal stem cells were plated at 50 000 cells per mL in serum-free AIM V medium in 384-well plates. After 24 hours, ALL cells were added on top of the mesenchymal stem cells at 400 000 cells per mL. Cells were treated after 4 to 6 hours as indicated, using an HP D300 Digital Dispenser (Tecan). For matrix-based screens, drug concentrations were used to reach a maximum decrease in ALL viability of ∼50%. A-485 and inobrodib (INB) were used at different concentration ranges, based on their different potencies.^15,16 LD4172, degrader of CREB-binding protein 1 (dCBP1), A-485, and INB were added to the coculture system right after the plating of ALL PDXs, to allow a 6-hour preincubation. After 72 hours of treatment, live cells were stained with CyQuant (catalog no. C35012, Invitrogen), imaged with Operetta CLS High-Content Analysis System and quantified using Biology Image Analysis Software (Single Cell Technologies). Viable ALL numbers were quantified for viability analyses, as described in supplemental Methods.

PID0117 PDXs were plated at a concentration of 1 million cells per mL in serum-free AIM V and treated with 50 nM birinapant (S7015, Selleckchem) for 2.5 or 5 hours or left untreated. RNA was extracted using RNeasy Plus Mini Kit (catalog no. 74134, Qiagen) following the manufacturer’s instructions. Quality control, library preparation, paired-end sequencing at 25 million clusters per sample, and bioinformatic analysis of RNA samples were executed using GalSeq srl. Detailed information about downstream analysis is described in the supplemental Methods.

Five million ALL cells were transduced with the single guide RNA (sgRNA) library (multiplicity of infection of 0.3 and coverage of ∼500 transduced cells per sgRNA) and injected into 1 mouse. Every mouse represented a singular condition and replicate. Treatment was executed as described under “PDXs and in vivo experiments” section. The experiment was conducted in triplicate. At the end of treatment window, genomic DNA from ALL cells collected from mice’s spleens was extracted using a column-based method (GenElute HP Plasmid Maxiprep Kit, catalog no. NA0310-KIT; Sigma-Aldrich). Further processing and analysis were performed as described in supplemental Methods by the Functional Genomic Center Zürich.

Statistical significance of differences was calculated using a 2-tailed Wilcoxon matched-pairs signed rank test for paired samples, to account for inter-PDX variations, when comparing multiple PDXs. The 2-way analysis of variance was used for multiple comparisons within the same PDX. The Mann-Whitney U test was used to calculate significant differences between independent PDX groups. The Pearson correlation coefficient (r) was calculated to detect linear correlations between drug sensitivities and synergy scores. The log-rank (Mantel-Cox) test was used to detect significant differences in the survival distributions between treatment groups. All statistical analyses were performed using Prism 10.0.3 (GraphPad software; not significant, P > .05; ∗P ≤ .05; ∗∗P ≤ .01; ∗∗∗P ≤ .001; ∗∗∗∗P ≤ .0001).

To identify the regulatory mechanisms that drive RIPK1-dependent cell death, we designed a combinatorial drug screen to improve SM action and thus RIPK1-dependent necroptosis in ALL. A library of 106 compounds (supplemental Table 1) was used in combination with fixed doses of SM (IC₂₀ and IC₅₀) on a primary B-cell precursor (BCP) patient-derived xenograft (PDX) ALL, using our established coculture DRP pipeline.¹⁷ Evaluation of response parameters identified 24 candidates with improved antileukemic effect in combination with SM (supplemental Figure 1). We subsequently used these compounds to perform a matrix-based synergy screen with SM in 5 different BCP-ALL and 1 T-cell ALL (T-ALL) PDX (Figure 1A-B). Several apoptosis modulators and HDAC inhibitors (HDACis) were among the classes of compounds with remarkable positive mean synergy scores (zero interaction potency [ZIP] scores) compared with SM, with Mo holding the highest average in mean ZIP scores across the samples (supplemental Table 2). Thus, activation of apoptosis through targeted B-cell lymphoma 2 (BCL2) antagonism potentiates the necroptotic SM response as shown earlier for glioblastoma¹⁸ and inhibition of transcriptional activity by HDACs is similarly effective in enhancing SM-driven cell death (Figure 1C). HDACis are of growing interest in the cancer treatment field because they restore normal acetylation patterns that are often disrupted during tumorigenesis.¹⁹ This results in the modulation of chromatin accessibility and transcription factor (TF) activities, both essential for transcriptional activity^20,21 and for SM-induced cell death.

We further validated the synergistic potential among our top 3 HDACi candidates (Mo, entinostat, and panobinostat) and SM in a larger cohort of pediatric ALL PDXs (n = 18), including diagnosis and relapse cases of both BCP-ALL and T-ALL (supplemental Figure 2; supplemental Table 3). Although panobinostat broadly inhibits the entire HDAC family, entinostat selectively targets HDAC1 and HDAC3,²² and Mo preferentially targets HDAC1, HDAC2, HDAC3, and HDAC11.²³ Our screen highlighted a positive synergy pattern more pronounced in BCP-ALL than T-ALL, with Mo being overall more synergistic with SM than the other HDACis (Figure 2A). To further confirm the synergistic potential of this combination in BCP-ALL, we expanded our PDX cohort (final n = 47) to comprise different leukemia subtypes (supplemental Figure 2; supplemental Table 3). Notably, 45 of 47 BCP-ALL PDXs showed consistently positive synergy scores for at least 2 of the 4 scoring methods evaluated (ZIP, Loewe, highest single agent, and Bliss) indicating a robust synergistic effect²³ (supplemental Figure 3A; supplemental Table 4). Sensitivities to single-agent SM vary among subtypes, given that TCF3::HLF⁺ and TCF3::PBX1⁺ BCP-ALL showed lower responses to treatment than the more SM-sensitive hyperdiploid or philadelphia-like leukemias. Conversely, sensitivity to Mo did not follow a subtype-specific trend. Interestingly, the synergistic potential of the SM and Mo combination was not determined by the cytogenetic subtype (Figure 2B), the disease stage (diagnosis/relapse), or single-agent sensitivities (supplemental Figure 3B-D). Considering the central role of RIPK1 in SM-driven cell death,⁹ we hypothesized that RIPK1 depletion attenuates the synergistic potential of HDACi with SM. Therefore, we used the RIPK1 proteolysis-targeting chimera (PROTAC) LD4172,²⁴ which effectively led to RIPK1 depletion within 6 hours, while also protecting the PDXs from SM-induced cell death (Figure 2C). Depletion of RIPK1 decreased the synergy scores between SM and the HDACi Mo, highlighting RIPK1’s importance for the synergistic activity of this combination (Figure 2D). Downstream of RIPK1, the pseudokinase MLKL is a key necroptosis executor.²⁵ PROTAC-mediated MLKL depletion had a moderate impact on the synergistic activity of SM and HDACi. However, MLKL depletion in PDXs deficient for the executioner caspases 3 and 7 led to a complete attenuation of the synergistic score between SM and HDACi, further supporting necroptosis as a key downstream cell death execution mechanism (supplemental Figure 3A). Given that inhibition of HDACs results in rapid acetylation of several lysine residues of histone 3 (H3),²⁶ we investigated the acetylation levels of H3 at lysine 27 (H3K27ac) and lysine 36 (H3K36ac) upon combinatorial treatment. Although the acetylation patterns upon treatment with SM, Mo, or the combination were variable among different PDXs, H3K36ac was significantly higher upon combinatorial treatment than upon single-agent treatment. The increase in H3K27ac was less pronounced (supplemental Figure 3B). Thus, interference with transcriptional regulation by HDAC inhibition is a potential mechanism to amplify SM antitumor activity. Similar drug combinations have been described to be effective in myeloid leukemia²⁷ and ovarian cancer²⁸ cells where they promote TNF-α expression and its autocrine signaling. However, our previous studies failed to correlate TNF-α levels with SM sensitivity, whereas the ligand could only partially modulate SM response in ALL PDXs.^9,10 We could not detect a significant increase in the secretion of TNF-α upon SM-HDACi combination, with overall very low TNF-α levels (<5 pg/mL) (supplemental Figure 3C), whereas the synergistic potential of the cotreatment was maintained in TNF-α-depleted PID0117 cells¹⁰ (supplemental Figure 3D). Therefore, our data rule out the contribution of TNF-α to the synergistic effect of SM and Mo.

Acetylation processes at the histone level are known to positively regulate transcriptional activity.²⁹ In addition, sustained protein translation marks an important feature of necroptosis, which supports the proinflammatory nature of this pathway by guaranteeing cytokine and chemokine production.³⁰ Therefore, we sought to identify changes in the transcriptional landscape during early responses to SM in PID0117, a BCP-ALL PDX showing a necroptotic cell death phenotype upon SM treatment.⁹ To address the importance of RIPK1, we used a previously established isogenic RIPK1-deficient BCP-ALL PDX, generated by CRISPR-mediated editing,⁹ which shows resistance to SM (Figure 3A). RNA sequencing analysis was performed in RIPK1 proficient (RIPK1^+/+) and RIPK1-deficient cells (RIPK1^−/−) upon short SM treatment (2.5 and 5 hours; Figure 3B). To obtain a RIPK1-specific gene signature, we compared SM treated (5 hours) with untreated conditions, respectively, within RIPK1^+/+ and RIPK1^−/− cells, followed by a comparison between the 2 isogenic PDXs. Kyoto Encyclopedia of Genes and Genomes pathways analysis of upregulated genes revealed TNF and NF-κB signaling as the most significantly enriched pathways upon 5 hours, but not 2.5 hours, of SM treatment (Figure 3C; supplemental Figure 5A) as part of a RIPK1-specific signature (Figure 3D). Of note, the missing correlation between TNF-α levels and SM response already reported in our PDXs¹⁰ does not exclude involvement of TNF signaling pathway genes upon necroptosis induction, as evident from the transcriptomic data, supporting a TNF receptor–dependent but ligand-independent role for RIPK1-mediated death.¹⁰ In addition, SM treatment increased the expression of genes involved in DNA-binding transcription activity (GO:MF), whereas genes with DNA modification functions (GO:MF) were downregulated (Figure 3E). This suggests a potential role for RIPK1 during transcriptional activation, which dictates necroptosis initiation. When focusing on RIPK1-regulated genes, we found, among others, TNFAIP3 as significantly upregulated upon SM treatment in RIPK1^+/+ cells (Figure 3F). TNFAIP3, together with its binding partner TNIP1, acts as a negative modulator of NF-κB, described to counteract TNF-mediated cell death by promoting RIPK1 ubiquitylation.^31,32 Analysis of TRRUST (transcriptional regulatory relationships unraveled by sentence-based text mining) identified the TFs that may modulate the expression of the differentially regulated genes activated upon SM treatment. Among them, we found NF-κB subunits (RELA, NFKB1, and REL), HDACs (HDAC1 and HDAC2), and SP1 as master regulators of gene expression upon SM treatment (Figure 3G-H). These data support the key role of NF-κB–driven gene expression^33,34 and extend into broader relevance of transcriptional modulation for RIPK1-driven cell death. Remarkably, the master regulator SP1 seemed to modulate both upregulated and downregulated genes upon treatment (Figure 3G; supplemental Figure 5B), emphasizing its importance in necroptosis modulation. SP1 activity is negatively regulated by HDAC1 via direct binding of the protein to the E2F1-binding site of SP1.³⁵ Moreover, the SP1-HDAC1/2 interaction has been described to form a corepressor complex, which maintains the hypoacetylated status of H3 and H4. This can be reversed by treatment with HDACis, which allows SP1 association with the HAT p300, leading to transcriptional activation.^36,37 These findings further highlight the connection between SM-induced cell death, regulation of acetylation processes, and TF activity. Thus, gene modulation is key to necroptosis induction in ALL, which could be further enhanced by pharmacological inhibition of HDACs-operated deacetylation processes.

RIPK1 activation, regulation, and downstream signaling during necroptosis have been broadly studied.³⁸ However, knowledge of additional genetic drivers that functionally regulate RIPK1-dependent cell death outcome is still scarce. With a targeted in vivo sgRNA CRISPR-based genome editing screen, we aimed to uncover thus far unrecognized genes that contribute to the SM response in ALL and to functionally determine the transcriptional regulation as a key mechanism for RIPK1-dependent cell death. We designed a targeted sgRNA library comprising the SM-induced genes detected by the transcriptome analysis for functional validation. In addition, genes encoding for HDACs/HATs and TFs of relevance were included (supplemental Material; supplemetal Data Set 2). The stably Cas9-expressing BCP-ALL PID0117 PDX⁹ was transduced with the sgRNA library and transplanted into NSG mice, which were subsequently treated with either vehicle or birinapant (Figure 4A). Although vehicle-treated cells expanded rapidly, SM treatment induced a delay in progression (supplemental Figure 6A). Upon targeted sequencing of the guides in harvested PDX cells, we identified a significant enrichment of guides including sgRNAs against SP1 and EP300. sgRNAs against RIPK1 served as positive controls and resulted to be enriched under both treatment conditions. In addition, we observed a consistent dropout of guides against HDAC2, TNIP1, and TNFAIP3 (Figure 4B-C). EP300 encodes for the acetyltransferase p300, which not only acetylates histones and TFs but also binds TFs, resulting in transcriptional activation of target genes in a feed-forward fashion, which is counteracted by HDACs.^39,40 Of note, the TF SP1 can be recruited by p300 through direct interaction with the N-terminal domain of SP1.⁴¹ The observed enrichment of cell populations with knockout of either SP1 or p300 upon SM treatment (Figure 4B) confirms their critical role in mediating sensitivity of ALL toward SM, in line with transcriptional activation being key for SM response (Figure 3E). Furthermore, depletion of HDAC2 (Figure 4B), an epigenetic transcriptional repressor, holds potential in sensitizing to SM treatment, which is supported by observed potentiated antitumor activity of SM by pharmacological HDAC inhibition (Figure 2A; supplemental Figure 3A). Interestingly, HDAC2 is specifically targeted by Mo,⁴² which we showed to have the highest synergistic potential when combined with SM (Figure 2A), but not by entinostat, another HDACi targeting mainly HDAC1 and HDAC3.²² Therefore, this functional screen highlights the pivotal role of HDAC2 and its inhibition by Mo in the observed synergy with SM.

To confirm a sensitizing role for p300 in SM-induced necroptosis, we compared SM sensitivities in different BCP-ALL PDXs upon degradation of p300 by the PROTAC dCBP1. Preincubation with 3 nM dCBP1 for 6 hours was sufficient for p300 degradation (Figure 5A, left). We detected a significant decrease in sensitivity to RIPK1-dependent cell death upon dCBP1 treatment (Figure 5A, right), confirming the sensitizing role of p300 suggested by the CRISPR screen. T-ALL and SM-insensitive B-ALL PDXs did not show a significant difference in sensitivity upon cotreatment (supplemental Figure 6B). To investigate which specific property of p300 rescues the response to SM in ALL, we combined SM with A-485, a p300 catalytic inhibitor,¹⁵ and INB (CCS1477), a p300 bromodomain inhibitor.⁴³ Cotreatment with A-485 caused a significant dose-dependent decrease in SM sensitivity, whereas for INB it plateaued at the second concentration used (Figure 5B). Thus, our data suggest that the catalytic activity of p300 might be predominant during SM response in ALL, whereas its ability to bind to acetylated proteins via the bromodomain could only have a limited effect. To confirm the functional relevance of SP1 and HDAC2 in RIPK1-dependent cell death, we performed an in vivo competition assay in which 2 stably Cas9-expressing BCP-ALL PDXs were used to first deplete SP1 or HDAC2 in ∼50% of the population (supplemental Figure 6C) with subsequent selective pressure by SM in vivo. Indeed, in the spleens and BM from both PDXs, the SP1-deleted populations were highly enriched in SM-treated mice compared with vehicle-treated mice. In contrast, we detected depletion of HDAC2-deficient populations in SM-treated vs vehicle-treated mice. Finally, we noticed a significantly lower human leukemic load upon HDAC2 depletion by SM treatment than the shuttle construct-containing counterpart, suggesting that even partial depletion of HDAC2 can sensitize leukemic cells to SM treatment (Figure 5C; supplemental Figure 6D). Taken together, these findings further support the functional role of SP1, p300, and HDAC2 in the regulation of necroptosis by SM, suggesting that transcriptional regulation is a potent driver of necroptosis.

The superior antitumor activity of the SM and Mo combination was assessed in vivo using 8 different BCP-ALL PDXs. ALL cells from each PDX were transplanted in 4 NSG mice, resulting in a 1-mouse-per-cohort study design.⁴⁴ ALL progression was monitored throughout the duration of the experiment, until the end point (ALL in peripheral blood of ≥50%) was met (Figure 6A). Within each PDX, the SM + Mo combination delayed ALL progression compared with the single agents, with ALL values remaining below 1% (in 6/8 PDXs) during the treatment window. The combination delayed progression even in cases where Mo treatment alone was not effective (Figure 6B-C; supplemental Figure 7A). Overall, the ALL-progression difference between control (vehicle) and treated mice, calculated at the control end point, was significantly greater for SM + Mo than the single agents (Figure 6D). There was no evidence of toxicity or body weight loss over the course of the treatment (supplemental Figure 7B). ALL-containing spleens from the mice were collected at individual end points, and ALL cells were screened ex vivo against SM and Mo to confirm that no resistant clones were selected during in vivo treatment (supplemental Figure 7C). These data confirm the combinatorial antileukemic action of SM and Mo in xenograft mouse models.

Insufficient activation of apoptosis represents a hallmark of cancer, with leukemia constituting a remarkable example.^3,45 Therefore, alternative strategies are needed to resensitize cancer cells to treatment, circumventing therapy resistance. Deregulation in the expression of IAPs is a strategy adopted by tumor cells, especially in hematologic malignancies, which contributes to disease progression and poor treatment response. Consequently, compounds including SM have been developed for targeting and inhibiting IAPs, along with RIPK1 deubiquitylation and necroptotic cell death engagement.⁴⁶ Efforts have been made to elucidate the molecular mechanisms behind RIPK1 switch from prosurvival to death signaling, describing distinct protein-protein interactions, alterations in NF-κB signaling,^38,47 and metabolic regulation.⁴⁸ Our data extend knowledge on the regulatory mechanisms of RIPK1-mediated cell death, revealing transcriptional aspects of necroptosis that are amenable to pharmacological interference to deepen or inhibit the response to SM treatment. Here, extending data on solid tumors,⁴⁹ we show that decreasing the threshold for mitochondrial membrane depolarization with BCL2 antagonists enabled sensitization to necroptosis by SM in our human leukemia model. Nevertheless, insufficient mitochondrial priming in drug-resistant leukemia may hamper a strategy of using BCL2 antagonists,³ and there is still a need for alternative approaches. Our data, combining DRP with transcriptome analysis and functional in vivo CRISPR screening, indicate that RIPK1-driven cell death is notably dependent on transcriptional regulation by the master regulators SP1 and HDAC2, where SP1 activity is integral to ALL response to SM, whereas HDAC2 antagonizes this process. In addition, epigenetic remodeling by the acetyltransferase p300 contributes to the necroptotic response, extending data from other model systems.⁵⁰ Consequently, pharmacological interference with acetylation/deacetylation programs alters the SM response and nominates the combination of SM and HDAC inhibition as an attractive, so far unexplored, starting point for future development of SM-potentiating strategies for ALL (Figure 6E). Accordingly, combinatorial treatment revealed a significant delay in leukemia progression in xenografted mouse models. Therefore, preclinical combinatorial DRP on patient samples represents a powerful tool to detect promising combinations and may inform individualized treatment options. This concept is being tested currently in an observational registry (ClinicalTrials.gov identifier: NCT06550102), in which this combination is a prime candidate for inclusion, also given the acceptable safety profiles of birinapant⁵¹ and Mo.⁵² The patient cohort tested here may be biased toward rare and high-risk subtypes. Although the combinatorial effect of SM and HDACi may be less efficient in subtypes not included in our cohort, the herein shown activity of this combination representing patient population at critical unmet medical need further supports efforts to translate this concept into clinical application. Confirming our previous findings on TNF-α–independent regulation of SM-induced cell death,¹⁰ the synergistic action between SM and HDACi in ALL is not driven by TNF-α secretion. Therefore, further investigation at the mechanistic level is needed to identify specific downstream players whose expression is influenced by the HDAC-dependent manipulation of transcription during SM-induced cell death. This will be beneficial for the development of more targeted HDACi treatments that, along with combinatorial partners like SM, could achieve higher therapeutic potential and lower toxicities in humans.⁵³ Although we highlight a contribution of the NF-κB signaling modulators TNFAIP3 and TNIP1 as negative regulators of necroptosis downstream of RIPK1, consistent with data from different models,^31,54,55 RIPK1 drives regulatory events independent of NF-κB, which may rely on protein-protein interactions.⁵⁶ These data allow for a reflection on strategies to manipulate necroptosis also in the context of other RIPK1-driven mechanisms such as antitumor immunity,⁵⁷ inflammatory responses,⁵⁸ or in degenerative diseases.⁵⁹ Necroptosis may have primarily evolved to cope with inflammatory signaling stress,⁶⁰ yet the underlying regulatory programs may be exploited to develop strategies to eradicate resistant leukemia cells, particularly under apoptosis-insufficient circumstances.

The authors are grateful to all members of the Bourquin and Bornhauser group for support and discussions. They thank the staff at the animal facility, the flow cytometry, and the microscopy core units at University Children’s Hospital Zürich (UZH), and acknowledge members of the Functional Genomic Center Zürich and GalSeq srl for their support.

B.C.B. was supported by funding from the Swiss Cancer Research Foundation (KFS-4384-02-2018), the Swiss Cancer League (KLS-5396-08-2021), the Swiss National Science Foundation (215340), and the Stefanini-Foundation Switzerland. J.-P.B. and B.C.B were supported by funding from the Stiftung Kinderkrebsforschung Schweiz, the Swiss Biobank Society of Pediatric Oncology and Hematology, and the clinical research priority program “Precision Hematology/Oncology” of UZH. The research was supported, in part, by the National Institutes of Health (Baylor College of Medicine, Houston [R01-CA268518; J.W.]), the Cancer Prevention and Research institute of Texas (RP220480 [J.W.]), and the Michael E. DeBakey Professorship in Pharmacology (J.W.).

Contribution: A.S. and B.C.B. conceptualized the study; A.S. curated the data; A.S. and B.G. performed the formal analysis; J.-P.B. and B.C.B. acquired funding; A.S., A. Dehler, B.G., F.S., M.R., J.K., and A. Drakul performed the investigation; A.S., A. Dehler, B.G., F.S., X.Y., and S.K. performed the methodology; A.S. and B.C.B. performed the project administration; J.-P.B. and B.C.B. were responsible for the resources; B.G. was responsible for the software; J.-P.B. and B.C.B. performed the supervision; A.S. and B.C.B. performed the validation; A.S., B.G., and F.S. performed the visualization; A.S. and B.C.B. wrote the original draft; and A.S., A. Dehler, B.G., F.S., M.R., J.K., A. Drakul, S.K., J.W., J.P.B., and B.C.B. reviewed and edited the manuscript.

Conflict-of-interest disclosure: J.W. is the cofounder of Chemical Biology Probes LLC and Fortitude Biomedicines, Inc; holds equity interest in Fortitude Biomedicines, Inc; and has stock ownership in, and serves as a consultant for, CoRegen Inc. The remaining authors declare no competing financial interests.

Correspondence: Beat C. Bornhauser, Department of Oncology and Children’s Research Centre, University Children’s Hospital Zürich, Lenggstrasse 30, 8008, Zurich, Switzerland; email: beat.bornhauser@kispi.uzh.ch.