The TAT-RasGAP317-326 anti-cancer peptide can kill in a caspase-, apoptosis-, and necroptosis-independent manner

Tumor cell resistance to apoptosis, which is triggered by many anti-tumor therapies, remains a major clinical problem. Therefore, development of more efficient therapies is a priority to improve cancer prognosis. We have previously shown that a cell-permeable peptide derived from the p120 Ras GTPase-activating protein (RasGAP), called TAT-RasGAP317-326, bears anti-malignant activities in vitro and in vivo, such as inhibition of metastatic progression and tumor cell sensitization to cell death induced by various anti-cancer treatments. Recently, we discovered that this RasGAP-derived peptide possesses the ability to directly kill some cancer cells. TAT-RasGAP317-326 can cause cell death in a manner that can be either partially caspase-dependent or fully caspase-independent. Indeed, TAT-RasGAP317-326-induced toxicity was not or only partially prevented when apoptosis was inhibited. Moreover, blocking other forms of cell death, such as necroptosis, parthanatos, pyroptosis and autophagy did not hamper the killing activity of the peptide. The death induced by TAT-RasGAP317-326 can therefore proceed independently from these modes of death. Our finding has potentially interesting clinical relevance because activation of a death pathway that is distinct from apoptosis and necroptosis in tumor cells could lead to the generation of anti-cancer drugs that target pathways not yet considered for cancer treatment.


INTRODUCTION
Cancer ranks among the leading causes of death worldwide [1]; this makes the development of novel and improved anti-cancer treatment a priority. Current oncological therapeutics aim to activate apoptosis to achieve disease control. However, cancer cells can adapt and become refractory to therapy by mutating and acquiring the ability to resist apoptotic stimuli [2]. The ability to evade apoptosis is one of the hallmarks of cancer [3]. Resistance leads to cancer cell survival and relapse. Consequently, development of treatments able to trigger non-apoptotic forms of death in cancer cells is of prime interest. There are many ways for mammalian cells to die, which could be divided into two main categories: accidental cell death (ACD) or regulated cell death (RCD) [4][5][6]. ACD is a form of death that is not induced by physiological or pathological insults and that does not involve the signaling machinery of cells. For example, ACD can result from exposure to extreme mechanical or chemical stimuli. In contrast, RCD requires genetically encoded molecular signaling

Research Paper
pathways and consequently can be modulated by genetic or pharmacologic interventions. Regulated forms of cell death include the intensively studied apoptosis, necroptosis and autophagy but also less known forms of death such as pyroptosis and parthanatos. These different types of cell death diverge at the level of the morphological changes and biochemical features triggered by death stimuli [4][5][6].
TAT-RasGAP 317-326 , a cell-permeable peptide derived from the p120 GTPase-activating protein (RasGAP), bears anti-malignant activities, including inhibition of metastatic progression and tumor cell sensitization to cell death induced by anti-cancer therapies [7][8][9][10][11][12]. This compound, in the initial tumor cell lines tested and at the given doses assessed was not found to affect their viability [7]. By screening additional tumor cells, however, we discovered that some cancer cell lines are directly killed by the RasGAP-derived peptide. The aim of the present study was to investigate which type of death was activated by TAT-RasGAP 317-326 in the cells that are directly eliminated by the peptide.

TAT-RasGAP 317-326 directly kills a subset of cancer cells
By screening a variety of cancer cell lines for their ability to become more sensitive to genotoxins in the presence of TAT-RasGAP 317-326 , we found 13 tumor cell lines that were directly killed by the peptide (Table 1). Among these, five are B-cell-derived cell lines (Daudi, Namalwa, Raji, Ramos and SKW6.4). To determine if non-transformed B cells were affected by the peptide, purified B cells (CD19 + cells) from healthy donors were tested. Supplementary Figure S1 shows that normal B cells were killed by TAT-RasGAP 317-326 while peripheral blood lymphocytes (PBL) were barely affected. This indicates that the RasGAP-derived peptide can be detrimental to a fraction of immune circulating cells and this will have to be taken into account in case the peptide is therapeutically used.
Two of the tumor cells that were efficiently killed by the peptide, the Raji Burkitt lymphoma and the NB1 neuroblastoma cell lines, were used to investigate the manner by which the peptide induced death. Supplementary Figure S2 shows that the peptide did not alter the proportion of cells in a given cell cycle stage, suggesting that cell cycle regulators are not targeted by the peptide. When treated with an inactive point mutant of the peptide [TAT-RasGAP 317-326 (W317A)] [13], or with the TAT cell-penetrating peptide alone, viability of Raji and NB1 cells was not affected ( Figure 1A). This demonstrates that the direct killing ability of TAT-RasGAP 317-326 is not carried solely by the TAT moiety but depends on specific RasGAP sequences. TAT-RasGAP 317-326 induced membrane permeabilization in a dose-dependent manner in both Raji and NB1 cells (Supplementary Figure S3A). The kinetic of death, induced by the peptide was assessed using Annexin-V and 7AAD staining. Annexin-V binds to phosphatidylserine (PS) exposed at the cell surface, a phenomenon that is generally characteristic of apoptosis. The 7AAD dye is plasma membrane impermeable and thus only labels cells with compromised cell permeability, which is typically seen in necrotic cells. Apoptotic cells bind Annexin-V but remain initially 7AAD negative [14]. Figure 1B-1C shows that Annexin-V and 7AAD positivity occurred concomitantly in Raji and NB1 cells when incubated with TAT-RasGAP 317-326 . In contrast and as expected, apoptotic inducers in these cells (e.g. Fas ligand (FasL) or etoposide) induced PS exposure before membrane impermeability was compromised ( Figure 1D). Of note, in both Raji and NB1 cells, TAT-RasGAP 317-326 induced a drop in mitochondrial membrane potential (Supplementary Figure S3B).
Reactive oxygen species (ROS) production upon peptide treatment was investigated. Intracellular hydrogen peroxide (H 2 O 2 ) and cytosolic superoxide (CO 2 .-) were not augmented by TAT-RasGAP 317-326 (Supplementary Figure  S4A). The sensors for these ROS were functional as shown in Jurkat cells treated with APO866, a nicotinamide phosphoribosyltransferase inhibitor (Supplementary Figure S4B). In contrast to hydrogen peroxide and cytosolic superoxide, mitochondrial superoxide (mO 2 .-) was increased in NB1 and, less markedly, in Raji cells in response to TAT-RasGAP 317-326 treatment (Supplementary Figure S4A). To assess the impact of mitochondrial ROS production on the viability of these cells, they were incubated with MitoTEMPO, a mitochondria-specific ROS scavenger. Supplementary Figure S4C shows that MitoTEMPO efficiently reduced TAT-RasGAP 317-326induced mO 2 .production in NB1 cells and this lowered death induced by the peptide. However, in Raji cells, MitoTEMPO did neither reduce the level of mO 2

.-
nor TAT-RasGAP 317-326 -induced death. While it is not clear why MitoTEMPO failed to inhibit mitochondrial superoxide production in Raji cells, the data obtained in NB1 cells indicate that TAT-RasGAP 317-326 might kill some cells via production of mitochondrial ROS.
As the RasGAP-derived peptide impacted the functionality of mitochondria, we determined if it modulated cellular ATP levels. As shown in Supplementary Figure S4D, the peptide induced a drop in ATP levels in Raji and NB1 cells. Intriguingly, this drop started to occur before any detectable cell death, suggesting that the peptide impacts cellular metabolism before inducing membrane permeabilization.
To evaluate the morphological changes involved in TAT-RasGAP 317-326 -induced cell death, we performed ultrastructural analyses in Raji and NB1 cells using electron microcopy (Figures 2 and 3). To have a comparison with apoptosis, we also treated Raji cells with FasL, a well-known apoptotic inducer ( Figure 2). FasL treatment induced apoptotic cell death with classical morphological criteria including nuclear and cytoplasmic condensation, chromatin condensation throughout the nucleus (pyknosis), nuclear fragmentation, and minimal alterations of organelles (including mitochondria). TAT-RasGAP 317-326 -treated Raji cells displayed different morphological features than those seen in FasL-treated cells (Figure 2). At first, slight and heterogeneous chromatin condensation without nuclear fragmentation, shrinkage of the cytoplasm and accumulation of cytoplasmic materials around the nucleus were observed. Organelles aggregated in distinct portions of the cytoplasm whereas some cytoplasmic areas seemed to be devoid of organelles. Organelles displayed a relatively good preservation but some mitochondria showed condensed features. Then, at later stages, Raji cells displayed a necrotic-like phenotype such as huge organelle swelling and loss of plasma membrane integrity. The inactive TAT-RasGAP 317-326 (W317A) mutant did not alter the ultrastructural morphology of Raji cells. Whereas TAT-RasGAP 317-326 -induced Raji cell death displayed a non-canonical morphological phenotype, TAT-RasGAP 317-326 -induced NB1 cell death showed some morphological features of apoptosis including chromatin condensation, nuclear fragmentation and condensed mitochondria at early stage and then secondary necrosis morphological features at later stage ( Figure 3). Morphological features in TAT-RasGAP 317-326induced NB1 cell death are therefore relatively similar to those observed when NB1 cells were exposed to the pro-apoptotic drug etoposide. However, clear differences were detected at the ultrastructural level between TAT-RasGAP 317-326 -and etoposide-treated NB1 cells. For example, the peptide never induced pyknosis in NB1 cells while etoposide almost always did so. Moreover, cytoplasmic materials often accumulated close to the nucleus in response to TAT-RasGAP 317-326 , while this was rarely the case when cells were incubated with etoposide.
As seen in Raji cells, the inactive TAT-RasGAP 317-326 (W317A) mutant had no effect on the ultrastructure of NB1 cells.
The data shown in Figures 2 and 3 indicate that TAT-RasGAP 317-326 is inducing a form of death that is sharing some characteristics with apoptosis, at least in some cell lines, but which does not seem to be strictly equivalent to this mode of death. We therefore aimed to determine if other types of cell death were triggered by the RasGAPderived peptide.

Apoptosis inhibition does not protect against TAT-RasGAP 317-326 -induced cell death in Raji cells and only partially in NB1 cells
TAT-RasGAP 317-326 induced caspase-3 activation and PARP1 cleavage in both Raji and NB1 cells ( Figure  4A), indicating that the peptide can trigger an apoptotic program. However, inhibition of caspase activity with the pan-caspase Z-VD-fmk inhibitor [15,16] ( Figure  4A), while efficiently blocking apoptosis triggered by FasL, had no effect on death induced by TAT-RasGAP 317-326 in Raji cells ( Figure 4B). In NB1 cells, however, a partial protection was observed ( Figure 4B). Intrinsic mitochondrial apoptosis is characterized by the release of cytochrome c into the cytosol, which depends on the pore forming proteins Bax and Bak. Unlike tBid, TAT-RasGAP 317-326 did not induce cytochrome c release from isolated mitochondria (Supplementary Figure S5), suggesting no direct action at the mitochondria level. The activity of Bax and Bak can be inhibited by overexpression of anti-apoptotic Bcl-2 family members, such as Bcl-X L [17,18]. Bcl-X L over expression ( Figure  4C) partially prevented TAT-RasGAP 317-326 -induced apoptosis in NB1 cells ( Figure 4D). In Raji cells, etoposide, even at high doses, only induced a slight increase in apoptosis, most probably because Raji cells express high basal level of Bcl-2 [19]. However, Bcl-X L overexpression did not protect Raji cells against TAT-RasGAP 317-326 -induced death ( Figure 4D). Similar results were obtained when Bax and Bak expression were removed by gene disruption using the CRISPR/ Cas9 technology ( Figure 4E-4F). Combining Bcl-X L overexpression and Z-VD-fmk treatment did not induce a stronger inhibition of TAT-RasGAP 317-326 -induced death in NB1 cells as compared to individual inhibitor applications ( Figure 4G).

TAT-RasGAP 317-326 does not trigger necroptosis
As apoptosis was not, or only partially, involved in the death induced by TAT-RasGAP 317-326 , we investigated whether other forms of death could be involved. Necroptosis, also called programmed necrosis, is a form of cell death that differs from apoptosis at morphological and signaling levels [20,21]. It is characterized by cell rounding, gain in cell volume, organelle swelling and plasma membrane rupture. Necroptosis requires receptorinteracting protein (RIP) 1 and 3. The downstream target of the complex formed by RIP1/RIP3 was identified as mixed lineage kinase domain-like protein (MLKL) [22,23]. Activation of MLKL leads to its translocation from the cytosol to plasma and intracellular membranes, and subsequent loss of membrane integrity [24]. In cells such as the HT29 colorectal adenocarcinoma, necroptosis can be triggered by tumor necrosis factor alpha (TNF-α) stimulation when caspases and translation are inhibited [25]. We were however unable to induce Raji and NB1 necroptosis using this protocol. This could be the consequence of a low MLKL expression ( Figure 5A). To assess the involvement of necroptosis in TAT-RasGAP 317-326 -induced death, Raji cells were treated with necrosulfonamide (NSA), an MLKL inhibitor [23]. NSA efficiently prevented necroptosis in HT29 cells ( Figure 5B) but had no effect on the death provoked by the RasGAP-derived peptide ( Figure 5C). One could argue that NSA might not be efficient in Raji cells, even at concentrations shown to be efficient in sensitive cell lines such as HT29. We therefore knocked out MLKL in Raji and NB1 cells as another approach to prevent necroptosis. Because endogenous levels of MLKL in these cells were low and could not be detected in Raji and NB1 cells ( Figure 5A), the targeted DNA region by the Cas9 endonuclease was sequenced. Figure 5D shows that both alleles of Raji clones #2 and #6 and NB1 clones #B3 and #A6 were disrupted, engendering frameshift mutations. These MLKL disrupted clones were then treated with the TAT-RasGAP 317-326 peptide but this did not prevent cell death ( Figure 5E). Collectively, these data demonstrate that TAT-RasGAP 317-326 does not require the molecular machinery of necroptosis to kill Raji and NB1 cells.

TAT-RasGAP 317-326 does not trigger pyroptosis
We next examined the implication of pyroptosis. This form of programmed cell death is stimulated by microbial and viral infections but also by stroke and cancer [26]. Morphological features displayed by pyroptotic cells are common with apoptosis and/or necrosis [26,27]. Defined as a caspase-1-dependent cell death, pyroptosis results in the production of inflammatory cytokines such as interleukin-1β (IL-1β) and IL-18 and ends up in cell lysis. To determine whether pyroptosis is a type of cell death induced by TAT-RasGAP 317-326 , Raji and NB1 cells lacking caspase-1 were generated. Loss of caspase-1 expression was confirmed in different clones by Western blotting ( Figure 6A). The ability of the peptide to cause cell death was not abrogated in caspase-1 knock-out Raji and NB1 cells ( Figure 6B). This is in line with the caspase inhibition results shown above ( Figure 4B) as Z-VDfmk is also expected to prevent caspase-1 activity [16]. Altogether, this indicates that the RasGAP-derived peptide does not elicit pyroptosis.

TAT-RasGAP 317-326 does not trigger parthanatos
Parthanatos is a cell death mode that is initiated by over-activation of poly (ADP-ribose)-polymerase 1 (PARP1). Under physiological conditions, PARP1 is involved in DNA repair. To maintain genomic homeostasis, PARP1 detects single strand DNA breaks, uses NAD + to synthetize poly (ADP-ribose) (PAR) and attaches PAR on itself and other target proteins [28,29]. This leads to the recruitment of critical proteins for DNA repair [30,31]. Hyperactivation of PARP1 contributes to NAD + and ATP depletion and translocation of apoptosis-inducing factor (AIF) from the mitochondria to the nucleus [32][33][34]. To investigate if TAT-RasGAP 317-326 triggers parthanatos, PARP1 knock-out cells were generated. Loss of PARP1 expression was controlled by Western blotting ( Figure  6C). Figure 6D shows that in the absence of PARP1, Raji and NB1 cells are still killed by the peptide. Hence, the peptide does not trigger parthanatos.

Autophagy and TAT-RasGAP 317-326 -induced death
Autophagy is a process of self-degradation. During starvation, it allows cells to maintain energy levels via the degradation and recycling of cellular cytoplasmic constituents, allowing cell survival. Autophagosome formation involves the lipidation of the LC3 protein [35]. This pro-survival function of autophagy is well accepted [36,37]. Autophagy may in some conditions trigger cell death [38]. To test if the peptide modulates autophagy, the autophagic marker lipidatino of LC3 was examined by Western blotting. Figure 7A shows that conversion of the LC3-I unlipidated form to the LC3-II lipidated form is similar in untreated cells and in cells treated with TAT-RasGAP 317-326 . To rule out the involvement of autophagy in TAT-RasGAP 317-326 -triggered death, two autophagic genes, ATG5 and ATG6, were disrupted ( Figures 7B and  7D). Disruption of ATG5 fully prevented autophagy in Raji cells, as assessed by the absence of LC3 lipidation in serum deprivated conditions ( Figure 7F). Moreover, autophagy induced by serum starvation fully prevented in cells lacking ATG6 ( Figure 7F). However, the absence of ATG5 and ATG6 did not prevent TAT-RasGAP 317-326induced cell death in Raji and NB1 cells ( Figures 7C  and 7E), suggesting that autophagy plays no role in TAT-RasGAP 317-326 -mediated death. www.impactjournals.com/oncotarget It is possible that the peptide triggers cell death via the activation of several pathways in parallel. We therefore investigated the effect of multiple cell death inhibition on TAT-RasGAP 317-326 toxicity. Figure 7G shows that inhibition of apoptosis and necroptosis in autophagy-deficient Raji and NB1 cells did not protect against the cytotoxic effect of the peptide. Taken together, these results indicate that none of the "classical" death pathways mediate the killing activity of TAT-RasGAP 317-326 .  Wild-type and ATG5 knock-out NB1 cells were treated or not with 40 µM TAT-RasGAP 317-326 (P) for 24 hours. Cell death was then assessed by flow cytometry after PI staining. Results correspond to the mean +/-95% CI of 3 independent experiments. E. Wild-type and ATG6 knockout Raji and NB1 cells were treated as described in panel D. Cell death was then assessed by flow cytometry after PI staining. Results correspond to the mean +/-95% CI of 3 independent experiments. F. Wild-type, ATG5 knock-out and ATG6 knock-out Raji cells were cultured for 48 hours in presence or absence of serum. LC3 conversion was analyzed by Western blotting. G. Wild-type, ATG5 knock-out and ATG6 knock-out Raji cells were pretreated or not with 10 µM Z-VD-fmk and 10 μM necrosulfonamide (NSA) for 1 hour and then treated or not with 20 μM TAT-RasGAP 317-326 (P) for 16 hours. Wild-type, ATG5 knock-out and ATG6 knock-out NB1 cells were pretreated or not with 10 µM Z-VD-fmk and 10 μM necrosulfonamide (NSA) for 1 hour and then treated or not with 40 µM TAT-RasGAP 317-326 (P) for 24 hours. Cell death was then assessed by flow cytometry after PI staining. Results correspond to the mean +/-95% CI of 3 independent experiments. www.impactjournals.com/oncotarget DISCUSSION Apoptosis had been the most intensively studied mode of regulated cell death for years. Thus, apoptotic inducers were and are largely used in the clinic as cancer therapies. However, tumor cell resistance to apoptosis, leading to cancer progression, is currently a major clinical problem. Some strategies could overcome this problem. For example, restoring the sensitivity of cancer cells to apoptosis might improve the efficacy of anti-tumor drugs. Another possibility would be to trigger alternate modes of death to which cancer cells are not resistant. The results presented here indicate that the TAT-RasGAP 317-326 peptide has the capacity to kill some tumor cells in a manner distinct from the known characterized forms of death.
Our data show that necroptosis, autophagy, parthanatos and pyroptosis are neither activated nor involved in the toxicity induced by TAT-RasGAP 317-326 . In contrast, the apoptotic pathway is efficiently stimulated by the peptide. However, inhibiting apoptosis, either pharmacologically or genetically, did not (Raji cells), or only partially (NB1 cells), prevent TAT-RasGAP 317-326 cytotoxicity. Hence, TAT-RasGAP 317-326 has the potential to activate apoptosis and another form of cell death that is distinct from the above mentioned death pathways. The ability of triggering multiple forms of death has been reported for other compounds. Indeed, shikonin and cisplatin, depending on the concentrations used, stimulate either apoptosis or necroptosis/necrosis [39,40]. Another example is the lipophilic mitochondria-targeted F16 compound that was shown to elicit death via apoptosis [41]. However, Bcl-2 overexpression did not block the capacity of F16 to trigger a necrotic cell death, demonstrating that one compound can have a dual ability to kill through both apoptosis and necrosis.
Our results clearly indicate that TAT-RasGAP 317-326 has the potential to stimulate both apoptosis and an alternative form of death in tumor cells of various origins. However, it is currently not known whether these two forms of death are activated independently or whether the triggering event leading to the alternative form of death has the capacity to also stimulate apoptosis.
In Raji cells, the sensitivity to the alternative form of death stimulated by TAT-RasGAP 317-326 is seemingly high and the apoptotic pathway can be blocked without affecting the overall death response. In NB1 cells, activation of the alternate form of death appears suboptimal and efficient killing requires concomitant activation of apoptosis. Consequently, it seems that there is a continuum of sensitivities among cancer cell lines to the direct killing action of TAT-RasGAP 317-326 .
Only a fraction of the tested cell lines were found to be sensitive to the killing activity of the peptide. To assess if those cells share common biological features that were not present in the resistant cell lines, we performed mutational and transcriptomics bioinformatical analyses on six sensitive and six resistant cell lines (subjected to 40 µM TAT-RasGAP 317-326 ). Supplementary Figure S6A lists the 6 genes with the strongest differential mutational status between resistant and sensitive cell lines. Even in these genes, there was no strict association between the mutation status and the sensitivity of the cell lines to the peptide. It is unlikely that these genes, individually at least, drive the peptide sensitivity of a cell line. In agreement with this is the absence of correlation between their expression and the sensitivity of a cell line to be killed by the peptide (Supplementary Figure  S6B). Moreover, there is no statistical support for a difference in the overall mutational rate of these cell lines (Supplementary Figures S6C-S6D). Gene profiling analysis revealed that cell lines cluster according to their origin rather than their sensitivity to TAT-RasGAP 317-326 (Supplementary Figure S6E). We finally performed a differential expression analysis to find whether the expression of certain genes was specifically associated with TAT-RasGAP 317-326 sensitivity but failed to detect any (Supplementary Figure S6F). These results suggest that neither mutation nor transcriptional regulation are involved in the regulation of TAT-RasGAP 317-326 sensitivity. However, we should take into account that the cell lines were not always tested in identical experimental conditions (e.g. different culture media used for the experimentation). Moreover, as the peptide requires entry into cells via the HIV-TAT 48-57 portion, it is also possible that variations in peptide intake could explain some of the differences in the sensitivity observed between the cell lines.
At present, we cannot rule out the possibility that the alternate form of death triggered by the peptide is a form of necrosis. One hypothesis that we cannot dismiss is that TAT-RasGAP 317-326 enters cells and then once in the cytoplasm alters, directly or indirectly, plasma membrane integrity by interacting with specific molecules that are enriched in the inner leaflet of the membrane. This mode of action has been reported for defensin NaD1, a host defense peptide [42]. NaD1 was shown to enter mammalian cells and bind to phosphatidylinositol 4,5-bisphosphate (PIP2), leading to rapid membrane destabilization and permeabilization. As highlighted in this example, cell death induced by a peptide may not, or not only, depend on protein binding but also on specific membrane lipid interaction.
To conclude, our finding could potentially have interesting clinical relevance since current anti-cancer therapies are mostly based on drugs that induce tumor cell apoptosis. Consequently, determining that a peptide is able to induce a distinct form of death in tumor cells could lead to the generation of innovative anti-cancer drugs that complement or be combined with existing ones. www.impactjournals.com/oncotarget

Cell lines
All cell lines were cultured in 5% CO 2

Peptides
TAT and TAT-RasGAP 317-326 are retro-inverso peptides (i.e. synthesized with D-amino acids in the opposite direction compared to the natural sequence). The TAT moiety corresponds to amino acids 48-57 of the HIV TAT protein (RRRQRRKKRG) and the RasGAP 317-326 moiety corresponds to amino acids 317-326 of the human RasGAP protein (DTRLNTVWMW). These two moieties are separated by two glycine linker residues in the TAT-Ras-GAP 317-326 peptide. TAT-RasGAP 317-326 (W317A) has the tryptophan at position 317 mutated into an alanine. The peptides were synthesized at the department of biochemistry, University of Lausanne, Switzerland, using FMOC technology, purified by HPLC and tested by mass spectrometry.

Cell cycle analysis
Cells were collected, washed once in PBS and fixed in 70 % ethanol at 4°C for 2 hours and then washed twice with PBS. DNA was stained with PI solution (10 μg/mL PI, 150 μg/mL RNase A in water) at 37°C for 30 minutes. Samples were analyzed by flow cytometry using a Beckman Coulter FC500 flow cytometer.

TMRM staining
Cells were collected, washed once in PBS and stained with 100 µL of 150 nM tetramethylrhodamine methyl ester (TMRM) solution by incubating during 20 min at 37°C. 500 µL of PBS were added and then transferred to a sample tube which was analyzed by flow cytometry using a Beckman Coulter FC500 flow cytometer.

Mitochondria isolation and cytochrome c release
Cells were harvested in PBS and centrifuged 10 min at 1,000xg. Cells were then resuspended in isotonic mitochondrial buffer (MB) (10 mM HEPES pH 7.4, 210 mM mannitol, 70 mM sucrose, 1 mM EDTA supplemented with one tablet of protease inhibitor cocktail per 50 mL), broken by five passages through a 25G1 0.5-by 2.5-mm needle fitted on a 2 mL syringe and centrifuged at 1,500xg for 5 min. This procedure was repeated twice and supernatants from each step were pooled and centrifuged 5 min at 1,500xg. Supernatant was collected, centrifuged 5 min at 2,000xg and further centrifuged 10 min at 9,000xg. Pellet was resuspended in MB (100 µL), centrifuged 10 min at 7,000xg and the pellet, representing the mitochondrial fraction, was finally resuspended in a volume of 100 µL of MB.

Antibodies
The rabbit anti-Bcl-X L , the rabbit anti-caspase-3, the rabbit anti-LC3, the mouse anti-PARP1, the rabbit anti-actin antibodies were from Cell Signaling (ref.

Determination of intracellular ATP content
Intracellular ATP content was measured with the ATP determination kit (Molecular Probes, ref. no A22066) according to manufacturer's instructions. Briefly, after drug treatment, cells were collected, washed twice in PBS, resuspended in 100 µL lysis buffer (NaHCO 3 20 mM + Na 2 CO 3 100 mM) and kept at -80°C for at least 4h. Cell lysates and ATP standards (10 µL) were mixed with standard reaction solution (90 µL) and luminescence was determined using an automatized bioluminometer (Promega Glomax 96 microplate luminometer). ATP level for each sample was normalized to protein content.

Genome editing by CRISPR method
Single guide RNAs targeting the early exon of the protein of interest were chosen in the sgRNA library [46] and are listed in Table 2. LentiCRISPR plasmids specific for a gene were created according to the provided instructions. Oligos were designed as follow: Forward 5'-CACCGnnnnnnnnnnnnnnnnnn nn-3'; Reverse-3'-CnnnnnnnnnnnnnnnnnnnnCAA-5', where nnnnnnnnnnnnnnnnnnnn in the forward oligo corresponds to the 20 bp sgRNA. Oligos were synthetized, then phosphorylated and annealed to form oligo complexes. LentiCRISPR vector was BsmBI digested and dephosphorylated. Linearized vector was purified and gel extracted and ligated to oligo complexes. The lentiCRISPR vector containing the sgRNA was then used for lentivirus production. Cells were infected and selected with the appropriate dose of puromycin (2 µg/ml for Raji, 1 µg/ml for NB1). Clone isolation was performed by limiting dilution in 96 well-plate.

TA cloning
TA cloning kit (Life technologies, ref. no. K202020) was used according to manufacturer's instructions to sequence DNA fragment containing the region where Cas9 was guided by a sgRNA. Ultrathin sections (with silver to gray interference) were cut with a diamond knife (Diatome), mounted on Formvar-coated single slot grids, and then counterstained with 3% uranyl acetate for 10 min and then with lead citrate (0.2%, Sigma-Aldrich, ref. no. 15326) for 10 min. Sections were visualized using a Philips CM100 transmission electron microscope.