Cell Death & Differentiation

Mouse models have historically been central to studies of TNF-induced cell death and guided pharmaceutical translation into clinic, based on the assumption that TNF signaling is conserved between human and mouse. Here, our work uncovers critical species-specific differences between the two. By systematically dissecting the roles of RIPK1, TRADD, and sensitivity to RIPK1 inhibitors in TNF signaling--including RIPK1 kinase-dependent and-independent apoptosis--we found that both apoptosis modalities diverge between human and mouse cells. In mouse cells, RIPK1 suppresses TRADD-mediated kinase-independent apoptosis, whereas in human cells, RIPK1 and TRADD act redundantly. Moreover, RIPK1 inhibitors block kinase-dependent apoptosis in mouse but not human cells, despite effectively inhibiting RIPK1 S166 phosphorylation. Cross-species complementation revealed that these discrepancies stem not from RIPK1 itself but from cell-context differences. These findings echo the clinical failures of RIPK1 inhibitors despite efficacy in mouse models and underscore the need for humanized models and therapeutics that more faithfully predict clinical outcomes.

Plasma membrane accumulation of phosphorylated mixed lineage kinase domain-like (MLKL) is a hallmark of necroptosis, leading to membrane rupture and inflammatory cell death. Pro-death functions of MLKL are tightly controlled by several checkpoints, including phosphorylation. Endocytosis and exocytosis limit MLKL membrane accumulation and counteract necroptosis, but the exact mechanisms remain poorly understood. Here, we identify linear ubiquitin chain assembly complex (LUBAC)-mediated M1 poly-ubiquitination (poly-Ub) as novel checkpoint for necroptosis regulation downstream of activated MLKL in human cells. Loss of LUBAC activity inhibits necroptosis, without affecting necroptotic signaling, but by preventing membrane accumulation of activated MLKL. Flotillin-1/2 act as putative necroptotic M1 poly-Ub targets that inhibit necroptosis suppression induced by LUBAC inhibition. Finally, we confirm LUBAC-dependent activation of necroptosis in primary human pancreatic organoids. Our findings identify LUBAC as species-specific regulator of necroptosis which promotes MLKL membrane accumulation and pioneer primary human organoids to model necroptosis in near-physiological settings.

Cytotoxic lymphocytes induce cancer cell death through death receptor-ligand interactions and the perforin-granzyme pathway. These pathways are generally thought to converge on the activation of executioner caspases to drive apoptosis. Here, we employed a reductionist approach to systematically disrupt key cell death mediators in a cytotoxic lymphocyte killing system to define their roles in determining cancer cell fate. We found that loss of executioner caspases conferred only limited resistance to cytotoxic lymphocyte-mediated killing. To identify cancer cell-intrinsic regulators that function beyond executioner caspases, we performed unbiased genome-wide CRISPR screens in executioner caspase-deficient cells. Unexpectedly, disruption of Fas or FADD--core components of the death receptor pathway--conferred substantial resistance to cytotoxic lymphocyte-mediated killing even in the absence of executioner caspases. This resistance persisted following additional disruption of known downstream mediators of Fas-FADD-caspase-8 (CASP8) signaling. Together, these findings identify the Fas-FADD-CASP8 axis as a central cancer cell-intrinsic determinant of susceptibility to cytotoxic lymphocyte-mediated killing whose function is not fully explained by canonical apoptotic or non-apoptotic effector pathways. Our results further suggest that CASP8 engages additional downstream substrates or mechanisms to promote cytotoxic lymphocyte-induced cancer cell death.

How p53 differentially activates cell cycle arrest versus cell death remains poorly understood. Here, we demonstrate that upregulation of canonical pro-apoptotic p53 target genes in colon cancer cells imposes a critical dependence on the long splice form of the caspase-8 regulator FLIP (FLIP(L)), which we identify as a direct p53 transcriptional target. Inhibiting FLIP(L) expression with siRNA or Class-I HDAC inhibitors promotes apoptosis in response to p53 activation by the MDM2 inhibitor Nutlin-3A, which otherwise predominantly induces cell-cycle arrest. When FLIP(L) upregulation is inhibited, apoptosis is induced in response to p53 activation via a novel ligand-independent TRAIL-R2/caspase-8 complex, which, by activating BID, induces mitochondrial-mediated apoptosis. Notably, FLIP(L) depletion inhibits p53-induced expression of the cell cycle regulator p21 and enhances p53-mediated upregulation of PUMA, with the latter activating mitochondrial-mediated apoptosis in FLIP(L)-depleted, Nutlin-3A-treated cells lacking TRAIL-R2/caspase-8. Thus, we report two previously undescribed, novel FLIP(L)-dependent mechanisms that determine cell fate following p53 activation.

Epithelial inflammasomes induce pyroptosis and release cytokines to defend against cytosolic pathogens. However, pyroptosis in epithelial barriers must be carefully regulated to facilitate elimination of infected cells while limiting widespread pyroptosis to preserve the single cell barrier. How epithelial cells achieve this is unknown. In this study, we describe a novel epithelial caspase regulation mechanism. By examining caspase-4 activation in human epithelial cells, we discovered that GSDMD pore formation serves as a signal to terminate caspase-4 activity thus facilitating epithelial cell expulsion while controlling cytokine secretion. Inhibition of epithelial pyroptosis led to IL-18 hyperproduction, likely as a mechanism to combat increased pathogen burden and initiate a wider immune response. Moreover, we demonstrate that full-length, rather than cleaved caspase-4 is active against IL-18 and propose that GSDMD pore formation facilitates cleavage of caspase-4 to terminate its catalytic activity. By comparing human cells and murine epithelial organoids to immune cells, we show that GSDMD pore mediated inhibition of caspase activity is largely specific to epithelial cells. Overall, these studies characterise a novel, epithelial-specific negative feedback loop that modulates inflammasome activity and challenge the dogma that autocatalytic caspase cleavage is required for caspase activity against substrates. Graphical AbstractIn intestinal epithelial cells, caspase activation simultaneously leads to GSDMD pore formation and IL-18 release. GSDMD pore formation provides a signal to terminate caspase activity and limit cytokine production. In GSDMD deficient cells, lack of an inhibition signal leads to caspase mediated IL-18 hyperproduction. Upon cell death this leads to release of massive amounts of IL-18. Created with BioRender.com O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=103 SRC="FIGDIR/small/578487v1_ufig1.gif" ALT="Figure 1"> View larger version (24K): org.highwire.dtl.DTLVardef@46f768org.highwire.dtl.DTLVardef@11cf918org.highwire.dtl.DTLVardef@125be35org.highwire.dtl.DTLVardef@ea9ead_HPS_FORMAT_FIGEXP M_FIG C_FIG

Activation of the NLRP3 inflammasome by infectious or sterile insults culminates in pyroptosis, a lytic and highly inflammatory form of programmed cell death. A safeguarded two-step process tightly regulates pyroptosis: priming, which drives NF-{kappa}B signaling, followed by execution, ultimately leading to plasma membrane rupture. Linear (Met1-linked) ubiquitination, catalyzed by the E3 ligase complex LUBAC, was previously shown to participate in pyroptosis, but the underlying mechanisms are not fully understood. In this study, we show that Met1-linked ubiquitin chains can assemble during both priming and execution phases, independently of the inflammasome sensor NLRP3. Genetic deletion of the LUBAC enzymes or pharmacological inhibition impaired pyroptosis. Conversely, cell death was enhanced without the deubiquitinase OTULIN, which selectively removes linear ubiquitination. Finally, using an optogenetic model to bypass priming, we demonstrate that Met-1-linked ubiquitination is required for the execution phase of pyroptosis. These findings offer insights into the regulation of pyroptotic cell death by linear ubiquitination.

Cellular material in apoptotic cells must be efficiently cleared by phagocytes to maintain tissue homeostasis. Defects in this process can lead to the onset of secondary necrosis and the release of intracellular contents such as damage associated molecular patterns (DAMPs) and autoantigens that are often derived from the nucleus. Therefore, appropriate handling and clearance of apoptotic material is vital to prevent unwanted inflammatory response and the onset of autoimmune disorders. However, how nuclear material is packaged by apoptotic cells for effective clearance by phagocytes is not well understood. By utilising murine models of apoptosis, we observed that a distinct subset of large extracellular vesicles generated from apoptotic thymocytes, known as apoptotic bodies (ApoBDs), can harbour the majority of nuclear contents. Mechanistically, we discovered that apoptotic cells can asymmetrically partition the nucleus into a single large membrane bleb located at one side of the cell, with other cellular contents such as mitochondria and acid organelles distributed to the opposite side. Whilst this newly observed apoptotic morphology, coined as asymmetric cell death morphology (AsyCDM), is morphologically similar to the process of erythroblast enucleation, pharmacological compounds that could interfere with erythroblast enucleation did not block the establishment of AsyCDM during apoptosis. Notably, AsyCDM was reliant on the contractile forces generated by ROCK1-dependent plasma membrane blebbing. Taken together, this study suggests that intracellular contents are partitioned into different ApoBD subsets during apoptosis through a regulated process driven by ROCK1-dependent actomyosin contraction.

Inflammasome assembly rapidly triggers caspase-1 activation to initiate pyroptosis, an inflammatory cell death characterized by the release of cytosolic contents, including interleukin (IL)-1{beta}/. Here, we report that inflammasome activation drives necrotic cell death independent of gasdermin D (GSDMD) and GSDME, which are essential executors of pyroptosis by forming a pore on the plasma membrane and increasing membrane permeability. NLRP3 inflammasome activation induced necrotic cell death, coupled with IL-1{beta}/ release in Gsdmd-/-Gsdme-/- macrophages. Mechanistically, the oligomerization of ninjurin-1 (NINJ1) was caused by inflammasome activation even in the absence of GSDMD and GSDME. Concordantly, glycine, an inhibitor of NINJ1, blocked plasma membrane permeabilization triggered by inflammasome activation in Gsdmd-/-Gsdme-/- macrophages, but not in WT macrophages. The dimerizer-mediated ASC oligomerization promoted NINJ1-mNeonGreen cluster formation in the absence of GSDMD and GSDME. Moreover, NINJ1 deficiency prevented membrane permeabilization initiated by ASC oligomerization in Gsdmd-/-Gsdme-/- immortalized bone marrow-derived macrophages (iBMDM). Blocking of phosphatidylserine (PtdSer) exposure, a feature of inflammasome-driven necrotic cell death, by Xkr8 deficiency inhibited plasma membrane permeabilization in Gsdmd-/-Gsdme-/- iBMDM. These results suggest that inflammasome-triggered activation of caspase-1 itself drives inflammatory necrotic cell death independent of gasdermins.

Understanding how the malignant cells respond to chemotherapy is essential to prevent the development of resistance and to improve the efficiency of anti-cancer drugs. Recently, we established that, by intrinsic and paracrine mechanisms, taxol treatment in breast tumor cells increases NOXA a pro-apoptotic protein functioning as an endogenous inhibitor of survival protein MCL-1, thereby enhancing cytotoxic load on the compensatory survival protein BCL-xL. We herein sought to define the contribution of NOXA/MCL-1 to the modality of cell death secretome composition upon anti-mitotic treatment associated with a BCL-xL antagonist. We observed that genetic inactivation of NOXA (enforcing MCL-1 pro-survival activity) in cancer cells not only delays their death when exposed to taxol in combination with the BCL-xL antagonist A1331852, but also alters its morphological characteristics with the apparition of features evoking pyroptosis. We identified the Caspase3-GSDME axis as regulating pyroptotic-like features suggesting that NOXA may act as a negative regulator of this cell death process (and MCL-1 as a positive regulator for it). Furthermore, comparative analysis of secretomes from the NOXA proficient or deficient cancer cells treated by taxol reveals variations in inflammatory cytokine production including those of IL-1{beta} and IL-18. Thus, our results show that anti-mitotic treatments are able to induce death by apoptosis and/or pyroptosis depending on BCL-2 family balance in breast cancer cells. Furthermore, NOXA/MCL-1 ratio appears to control the communication between these two types of cell death and their associated extracellular inflammatory signals in coordination with the pore-forming gasdermin GSDME.

Normal and cancer cells accumulate lipid droplets (LDs) under stress to buffer lipotoxicity, but their role in regulated cell death (RCD) remains incompletely understood. Here, we explored LD accumulation across diverse apoptotic and non-apoptotic RCD modalities in human cancer cells and Drosophila germ cells. We found that LD accumulation arises from de novo LD biogenesis, whereas LD lipolysis remains active--or even enhanced--in dying germ cells and cancer cells, respectively. In Drosophila, LD accumulation in the Brummer lipase mutant inhibited germ cell death, indicating a protective function. Proteomic and imaging analyses revealed a broad redistribution of LD-associated proteins, encompassing lipid metabolism and stress response factors, as well as the pro-apoptotic effector Bax in human cancer cells. Enhanced LD-mitochondria contacts promoted active Bax translocation from mitochondria to LDs, thereby delaying apoptosis execution. Conversely, depletion of LDs sensitized cells to Bax- or truncated Bid-induced apoptosis. Collectively, these findings define LD accumulation during cell death as a delaying mechanism in which LDs sequester mitochondrial cell death regulators, attenuating their pro-death activity and revealing potential therapeutic implications for apoptosis-resistant cancers.

Mixed lineage kinase-like (MLKL) is activated by RHIM-domain containing kinase (RIPK)3 to permeabilize the plasma-membrane and execute necroptosis, a form of regulated necrosis. We found that MLKL is activated in an atypical, RIPK3- and necroptosis-independent manner downstream of Toll-like receptor 3, resulting in its translocation to lysosomes and lysosomal membrane permeabilization. Damaged lysosomes then undergo exocytosis, leading to the integration of lysosomal MLKL into the plasma-membrane to trigger cell death. The ESCRT-machinery can repair damaged lysosomes and counteract cell death by packing lysosomal MLKL into intraluminal vesicles, which are subsequently released as extracellular vesicles. In this way, ESCRT-machinery balances life and death decisions by preventing lysosomal MLKL to reach its killing destination, which is the plasma-membrane.

Dependence receptors (DRs) induce cell death by apoptosis when unbound by their cognate ligands. Among them, Kremen1 was first described to induce cancer cell death in the absence of its ligand, DKK1. However, the precise mechanism of Kremen1-induced cell death remains unclear. In this study, we demonstrate that Kremen1 induces cell death with autophagic features, contrasting with the apoptotic process typically associated with dependence receptors. Specifically, the pharmacological inhibition of autophagy, or genetic silencing of key autophagy effectors, efficiently suppresses this cell death process. A biotin proximity labeling for protein-protein interactions identified SEC24C, a component of the COP-II complex, as a critical effector in Kremen1-induced autophagy and cell death. Our findings further reveal that Kremen1 is in proximity with SEC24C and ATG9A after vesicular trafficking and fosters the interaction of SEC24C with ATG8, ERGIC and ATG9A. This potentially underlies the increased number of autophagosomes leading to cell death. The induction of aberrant autophagy by Kremen1 deserves particular attention, especially as the Kremen1/DKK1 pair is frequently altered in cancers. Thus, targeting this pathway may offer a potential strategy for treating cancers resistant to current therapies.

ObjectiveDiffuse cutaneous systemic sclerosis (dcSSc) is a life-limiting fibrotic disease. We and others have shown that dcSSc fibroblasts accumulate numerous somatic mutations associated with senescence-like features; however, the mechanism(s) enabling their survival remain unclear. MethodsSkin biopsies were obtained from lesional tissues from dcSSc (n=10), dcSSc treated with autologous hematopoietic stem cell transplantation (ASCT, n=8) or 7 age/sex-matched healthy controls. Primary dermal fibroblasts were generated from biopsies. Spatial RNA sequencing, immunoblotting, confocal microscopy, and functional assays were used to mechanistically delineate signaling pathways linking DNA-damage with fibroblast survival. ResultsdcSSc fibroblasts demonstrated increased pH2AX DNA double-strand-break foci yet remained apoptosis resistant. These cells displayed features of metabolic-stress remodeling, including mitochondrial hyperpolarization, increased reactive oxygen species production, and enhanced mitochondrial biogenesis. Spatial transcriptomics and subsequent biochemical analyses identified activation of a PERK/ATF4/FOXO1 axis, characterized by PERK phosphorylation, selective ATF4 translation, FOXO1 nuclear translocation, and induction of downstream antioxidant and metabolic programs. In contrast, fibroblasts from post-ASCT patients exhibited normalization of DNA-damage markers and mitochondrial parameters without ATF4/FOXO1 activation. Pharmacologic inhibition of either PERK or FOXO1 selectively restored mitochondrial-dependent apoptosis in dcSSc fibroblasts, demonstrating that this axis is required for their survival following extensive genomic injury. ConclusiondcSSc fibroblasts persist despite substantial genomic injury by engaging a PERK/ATF4/FOXO1 metabolic-adaptation program that suppresses mitochondrial-dependent apoptosis. This survival axis is not present after ASCT. Targeting PERK or FOXO1 restores apoptosis selectively in dcSSc fibroblasts, highlighting its potential use as a therapeutic target for eliminating pathogenic senescence-like fibroblasts in dcSSc. HighlightsO_LIBoth ex-vivo skin and in-vitro primary dermal fibroblasts derived from dcSSc patients have a higher frequency of intrinsic DNA damage signals and senescence-associated features; yet they evade mitochondrial-dependent apoptosis. C_LIO_LIPathogenic dcSSc fibroblasts rewire their metabolism, characterized by mitochondrial hyperpolarization and elevated ROS. C_LIO_LISpatial transcriptomics and functional analyses reveal a PERK/ATF4/FOXO1 stress-adaptation axis that drives fibroblast survival in dcSSc. C_LIO_LIThis maladaptive survival program characterized by increased genotoxic stress, and mitochondrial remodelling is absent in post-ASCT fibroblasts. C_LIO_LITargeting PERK or FOXO1 selectively sensitizes dcSSc fibroblasts to apoptosis revealing a potential promising therapeutic strategy in dcSSc. C_LI Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=142 SRC="FIGDIR/small/706443v1_ufig1.gif" ALT="Figure 1"> View larger version (31K): org.highwire.dtl.DTLVardef@1b3f791org.highwire.dtl.DTLVardef@42548aorg.highwire.dtl.DTLVardef@bc4ce1org.highwire.dtl.DTLVardef@5b4d48_HPS_FORMAT_FIGEXP M_FIG C_FIG

Death receptor ligand TRAIL is a promising cancer therapy due to its ability to selectively trigger extrinsic apoptosis in cancer cells. However, TRAIL-based therapies in humans have shown limitations, mainly due inherent or acquired resistance of tumor cells. To address this issue, current efforts are focussed on dissecting the intracellular signaling pathways involved in resistance to TRAIL, to identify strategies that sensitize cancer cells to TRAIL-induced cytotoxicity. In this work, we describe the oncogenic MEK5-ERK5 pathway as a critical regulator of cancer cell resistance to the apoptosis induced by death receptor ligands. Using 2D and 3D cell cultures and transcriptomic analyses, we show that ERK5 controls the proteostasis of TP53INP2, a protein necessary for full activation of caspase-8 activation in response to TNF, FasL or TRAIL. Mechanistically, ERK5 phosphorylates and induces ubiquitylation and proteasomal degradation of TP53INP2, resulting in cancer cell resistance to TRAIL. Concordantly, ERK5 inhibition or genetic deletion, by stabilizing TP53INP2, sensitizes cancer cells to the apoptosis induced by recombinant TRAIL and TRAIL/FasL expressed by Natural Killer cells. The MEK5-ERK5 pathway regulates cancer cell proliferation and survival, and ERK5 inhibitors have shown anticancer activity in preclinical models of solid tumors. Using endometrial cancer patient-derived xenograft organoids, we propose ERK5 inhibition as an effective strategy to sensitize cancer cells to TRAIL-based therapies and Natural Killer cells.

Mitochondrial Encephalopathy, Lactic Acidosis, and Stroke-like episodes (MELAS) syndrome is primarily caused by the heteroplasmic m.3243A>G/MT-TL1 pathogenic variant. Patients exhibit elevated circulating cell-free mtDNA (cf-mtDNA) in plasma, which acts as a damage-associated molecular pattern. Using patient-derived fibroblasts and neuronal progenitors, as well as transmitochondrial cytoplasmic hybrids (cybrids), we show that mutant cells release higher levels of cf-mtDNA than wild-type controls, demonstrating that the m.3243A>G pathogenic variant drives mtDNA release. Mechanistically, increased mitochondrial oxidative stress promotes mtDNA oxidation and fragmentation, leading to Ca2+ overload and subsequent mtDNA extrusion. This, in turn, triggers inflammasome activation and pyroptosis, resulting in the secretion of pro-inflammatory cytokines and the activation of innate immune pathways. Pharmacological inhibition of the Mitochondrial Calcium Uniporter (MCU) or Voltage-Dependent Anion Channel (VDAC) reduced mtDNA release, confirming their involvement. Overall, our findings reveal a previously unrecognized mechanism in MELAS linking mitochondrial dysfunction to innate immune activation, with potential implications for therapeutic intervention. TeaserMELAS mutation drives mtDNA release, triggering inflammation via oxidative stress, calcium imbalance, and inflammasome activation

HSAN9 is a rare progressive neurodegenerative disease in children linked to bi-allelic loss-of-function mutations in the TECPR2 gene. TECPR2 is a multi-domain protein harboring N-terminal WD repeats and C-terminal TECPR repeats, followed by a functional LIR motif that serves in autolysosomal targeting. Here, we show that the lack of TECPR2 leads to impairment of mitophagy that can be recovered by the expression of its C-terminal domain. Accordingly, we uncover severe mitochondrial dysfunction and accumulation of mitochondrial content in primary fibroblasts derived from an HSAN9 patient, and in embryonic fibroblasts and dorsal root ganglia derived from an HSAN9 mouse model. Strikingly, these mitochondrial defects are mediated by a mitochondrial stress through activation of the integrated stress response (ISR), whereas mitochondrial function is recovered by pharmaceutical or genetic suppression of ISR. Our findings provide a new link between mitophagy and ISR in mitochondrial homeostasis during neurodegeneration.

Malignant pleural mesothelioma (MPM) is an aggressive malignancy linked to asbestos exposure and highly resistant to chemotherapy, potentially due to upregulated expression of the pro-survival proteins, BCL2/BCL-XL/MCL-1. Using clinically-relevant models of MPM we show that patient-derived primary MPM cell lines and ex-vivo 3D tumour explants are highly resistant to apoptosis induced by the BCL2/BCL-XL inhibitor, ABT-737. Importantly, we discover that 2-deoxyglucose (2DG), a glycolytic inhibitor, can sensitize MPM cells to ABT-737 and show this correlates with loss of the pro-survival protein, MCL-1. siRNA knockdown of MCL-1 (MCL-1 KD) combined with ABT-737 induced BAX/BAK-dependent, but BIM/PUMA-independent apoptosis, mimicking 2DG/ABT-737 treatment. MCL-1 KD/ABT-737 induced mitochondrial cytochrome c release and caspase-independent inhibition of mitochondrial respiration. Moreover, we observed a hitherto unreported caspase-dependent cleavage of glycolytic enzymes and subsequent inhibition of glycolysis. 2DG inhibited ERK/STAT3 activity, decreased MCL-1 mRNA and protein levels, with concurrent activation of AKT, which limited loss of MCL-1 protein. However, co-treatment with a specific AKT inhibitor, AZD5363, and 2DG/ABT-737 potently induced cell death and inhibited clonogenic cell survival, while in MPM 3D tumour explants MCL-1 protein expression decreased significantly following 2DG or 2DG/AZD5363 treatment. Notably, a similar synergy was observed in MPM cell lines and MPM 3D tumour explants using ABT-737 in combination with the recently developed MCL-1 inhibitor, S63845. Importantly, our study provides a mechanistic explanation for the chemoresistance of MPM and highlights how this can be overcome by a combination of metabolic reprogramming and/or simultaneous targeting of MCL-1 and BCL-2/BCL-XL using BH3-mimetics.

TP53 (also called TRP53 or p53) is a critical tumour suppressor that prevents cancer development by inducing a transcriptional program which can lead to diverse cellular responses, most prominently, cell proliferation arrest/senescence with survival of cells or cell death by apoptosis. Why distinct cell types undergo different outcomes after p53 activation remains unclear. Using integrated RNA-sequencing, proteomic and functional analyses across a diverse range of murine primary cell types, we demonstrate that cell fate is governed by the balance between pro-survival BCL-2 and pro-apoptotic BH3-only proteins. Cells resistant to apoptosis displays a higher starting ratio of pro-survival BCL-2 to pro-apoptotic BH3-only proteins, along with transcriptional upregulation of the pro-survival gene Bcl2l1, encoding BCL-XL. This control of cell fate is also seen in human wild-type p53 cancer cell lines. These findings reveal the mechanism for understanding p53-driven cell fate decisions, suggest therapeutic strategies to shift p53-induced cell proliferation arrest/senescence toward apoptotic cell death and allowed generation of an RNAseq data-based predictor of outcome for cancer cells after p53 activation.

Intrinsic apoptosis is principally governed by the BCL-2 family of proteins, but some non-BCL-2 proteins are also critical to control this process. To identify novel apoptosis regulators, we performed a genome-wide CRISPR-Cas9 library screen, and it identified the mitochondrial E3 ubiquitin ligase MARCHF5/MITOL/RNF153 as an important regulator of BAK apoptotic function. Deleting MARCHF5 in diverse cell lines dependent on BAK conferred profound resistance to BH3-mimetic drugs. The loss of MARCHF5 or its E3 ubiquitin ligase activity surprisingly drove BAK to adopt an activated conformation, with resistance to BH3-mimetics afforded by the formation of inhibitory complexes with pro-survival proteins MCL-1 and BCL-XL. Importantly, these changes to BAK conformation and pro-survival association occurred independently of BH3-only proteins and influence on pro-survival proteins. This study identifies a new mechanism by which MARCHF5 regulates apoptotic cell death and provides new insight into how cancer cells respond to BH3-mimetic drugs. These data also highlight the emerging role of ubiquitin signalling in apoptosis that may be exploited therapeutically.

Damaged or superfluous cells are often eliminated by apoptosis. Although a cell-autonomous process, apoptotic cells communicate with their environment in different ways. However, the extent to which apoptotic cells alerting their neighbours to potential danger is unclear. Addressing this question, here we describe a mechanism whereby dying cells can promote survival of neighbouring cells. We find that during apoptosis, cells release the growth factor FGF2, leading to MEK/ERK-dependent transcriptional upregulation of pro-survival BCL-2 proteins in a non-cell autonomous manner. This transient upregulation of prosurvival BCL-2 proteins in turn can protect neighbouring cells from apoptosis. Accordingly, we find in certain cancer types a correlation between FGF-signalling, BCL-2 expression and worse prognosis. Importantly, either co-treatment with FGF-receptor inhibitors or removal of apoptotic stress restores apoptotic sensitivity. These data reveal a pathway by which dying cells can increase resistance to cell death in surrounding cells. Beyond mediating cytotoxic drug resistance, this process may serve additional roles, for instance limiting tissue damage in response to stress.