TNF-induced necroptosis and PARP-1-mediated necrosis represent distinct routes to programmed necrotic cell death
- 9.1k Downloads
Programmed necrosis is important in many (patho)physiological settings. For specific therapeutic intervention, however, a better knowledge is required whether necrosis occurs through one single “core program” or through several independent pathways. Previously, the poly(ADP-ribose) polymerase (PARP) pathway has been suggested as a crucial element of tumor necrosis factor (TNF)-mediated necroptosis. Here, we show that TNF-induced necroptosis and the PARP pathway represent distinct and independent routes to programmed necrosis. First, DNA-alkylating agents such as 1-methyl-3-nitro-1-nitrosoguanidine (MNNG) or methyl methanesulfonate rapidly activate the PARP pathway, whereas this is a late and secondary event in TNF-induced necroptosis. Second, inhibition of the PARP pathway does not protect against TNF-induced necroptosis, e.g., the PARP-1 inhibitor 3-AB prevented MNNG- but not TNF-induced adenosine-5′-triposphate depletion, translocation of apoptosis-inducing factor, and necrosis. Likewise, olaparib, a more potent and selective PARP-1 inhibitor failed to block TNF-induced necroptosis, identical to knockdown/knockout of PARP-1, pharmacologic and genetic interference with c-Jun N-terminal kinases and calpain/cathepsin proteases as further components of the PARP pathway. Third, interruption of TNF-induced necroptosis by interference with ceramide generation, RIP1 or RIP3 function or by the radical scavenger butylated hydroxyanisole did not prevent programmed necrosis through the PARP pathway. In summary, our results suggest that the currently established role of the PARP pathway in TNF-induced necroptosis needs to be revised, with consequences for the design of future therapeutic strategies.
KeywordsNecroptosis Programmed necrosis TNF PARP-1
Programmed cell death (PCD) is essential for survival and homeostasis of multicellular organisms. Whereas induction of PCD through activation of caspases (subsequently termed “apoptosis”) represents the primary route to PCD in most physiological settings, alternative forms of PCD have been discovered that operate without detectable caspase activity. Of those, programmed necrosis represents the main mode of caspase-independent PCD . Programmed necrosis can participate in vital processes when caspase-dependent pathways are abrogated, for example in the removal of interdigital webs during embryogenesis in Apaf-1-deficient embryos . Furthermore, programmed necrosis contributes to the negative selection of lymphocytes, termination of immune responses, regulation of bone growth, ovulation, cellular turnover in the intestine or post-lactational regression of the mammary gland [2, 3, 4]. In addition, programmed necrosis has been implicated in the pathology of hyperacute shock , pancreatitis [6, 7], renal , cerebral and myocardial ischemia–reperfusion injury, epilepsy, muscular dystrophy, Alzheimer’s, Huntington’s and Parkinson’s disease, as well as in the destruction of cells by pathogens such as HIV, vaccinia virus, Shigella and Salmonella [1, 2, 9, 10]. As a backup mechanism, programmed necrosis can permit a cell to commit suicide if the caspase machinery fails or is inactivated (e.g., after virus infection or in apoptosis-resistant tumor cells) [2, 11].
Necroptosis represents a subset of programmed necrosis, which is elicited through death receptors and which depends on the activity of the protein kinases RIP1 and RIP3 [6, 7, 10, 11]. Downstream of RIP3, the very recently discovered proteins MLKL and PGAM5 further convey the necroptotic signal by inducing mitochondrial fragmentation . Enzymes of the energy metabolism and production of reactive oxygen species (ROS), e.g., by the NADPH oxidase Nox1 or by mitochondria, as well the deubiquitinase CYLD and the Bcl-2 family member Bmf represent further candidate mediators of necroptosis . Moreover, FADD and caspase-8 constitute essential negative regulators of RIP1/RIP3-mediated necroptosis during normal embryonic development .
The two most extensively studied models of programmed necrosis are necroptosis initiated by the 55-kDa tumor necrosis factor (TNF) receptor (TNF-R1) and necrosis mediated via the poly(ADP-ribose) polymerase (PARP) pathway. The latter is initiated by overactivation of the DNA repair enzyme PARP-1, e.g., in response to DNA-alkylating agents such as 1-methyl-3-nitro-1-nitrosoguanidine (MNNG), resulting in the massive synthesis of poly(ADP-ribose) (PAR) from nicotinamide adenine dinucleotide (NAD+) and, in consequence, to the rapid depletion of intracellular NAD+ and adenosine-5′-triposphate (ATP) pools. Subsequently, and involving c-Jun N-terminal kinases (JNK), the mitochondrial protein Bax and calpain/cathepsin proteases, cleavage of apoptosis-inducing factor (AIF) occurs, which then translocates from mitochondria to the cytosol and further to the nucleus where it forms an active DNA-degrading complex with histone H2AX and cyclophilin A. This complex then induces caspase-independent, large-scale DNA fragmentation and finally, necrosis [1, 13, 14].
A previous study has implicated the PARP pathway as an integral part of TNF-induced necroptosis in L929 fibrosarcoma cells . Since the existence of one core program of necroptosis would allow a more efficient therapeutic interference than the existence of several distinct pathways that would have to be targeted simultaneously, this would also have implications for the future development of antinecrotic cytoprotective drugs .
We have previously demonstrated that the sphingolipid ceramide represents a key mediator of TNF-induced necroptosis in L929 cells as well as in other cell systems [16, 17]. In our efforts to more precisely define the targets of ceramide in TNF-induced necroptosis, we extended our investigations to components of the PARP pathway. Unexpectedly, our results indicate that the PARP pathway is not integral to TNF-induced necroptosis, but rather that both pathways represent distinct and independent routes to programmed necrosis.
Materials and methods
Reagents. Highly purified human recombinant TNF was provided by BASF Bioresearch (Ludwigshafen, Germany). Methyl methanesulfonate (MMS), cycloheximide (CHX), benzyloxycarbonyl-Phe-Ala-fluoromethylketone (zFA-fmk), trans-epoxysuccinyl-l-leucylamido(4-guanidino)butane (E-64), pepstatin A, imipramine and butylated hydroxyanisole (BHA) were purchased from Sigma (Deisenhofen, Germany), MNNG either from Sigma or from ABCR (Karlsruhe, Germany). Carbobenzoyl-l-leucyl-l-leucyl-l-leucinal (MG-132) and poly(ADP-ribose) glycohydrolase (PARG) purified from bovine thymus were from Enzo (Lausen, Switzerland). 3-aminobenzamide (3-AB) and N-(5,6-dihydro-6-oxo-2-phenanthridinyl)-2-acetamide hydrochloride (PJ-34) were obtained from TCI (Eschborn, Germany) and from Tocris (Ellisville, MO, USA), olaparib was purchased from Axon (Groningen, The Netherlands). Benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethylketone (zVAD) was from Bachem (Bubendorf, Switzerland), JNK-inhibitor III and necrostatin-1 were obtained from Merck (Darmstadt, Germany) and N-[l-3-trans-(propylcarbamoyl)-oxirane-2-carbonyl]-l-Ile-l-Pro methyl ester (CA-074 Me) from Biomol (Hamburg, Germany).
Cell culture. L929ATCC, HeLa, wild-type Jurkat, HT-29, 293T, and MCF-7 cells were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). L929sA cells  were kindly provided by Peter Vandenabeele (Ghent, Belgium). L929Ts is a tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-sensitive L929 subline derived in our laboratory . Wild-type and PARP-1-deficient mouse embryonic fibroblasts (MEF) were kindly provided by Françoise Dantzer (Illkirch, France). RIP1-deficient Jurkat cells were a gift from Brian Seed (Boston, MA, USA). Wild-type and RIP1-deficient MEF have been previously described , as have NIH3T3 cells naturally expressing or deficient for RIP3 . NIH3T3 transfectants stably expressing murine green-fluorescent-protein (GFP) fusions to wild-type or kinase-defective murine RIP3 in the vector pEGFP-N1 (Takara, Mountain View, CA, USA) were generated by selection in DMEM medium containing 1 mg/ml G418 (PAA, Cölbe, Germany). Positive clones were identified by flow cytometry. Primary lung fibroblasts from JNK1-, JNK2-, and RIP3-deficient mice and their wild-type littermate controls were prepared and cultured as described . Immortalized wild-type and JNK1/JNK2 double-deficient MEF were a kind gift from Roger Davis (Worcester, MA, USA). Wild-type, cathepsin D-deficient, and cathepsin D-deficient MEF stably transfected with the pro-cathepsin D cDNA have been previously described , as have TNF-R1/R2 double-deficient mouse fibroblasts and their transfectants stably re-expressing TNF-R1 . Wild-type and RIP3-deficient MEF were generated from d14.5 embryos according to standard procedures. Cells were cultivated in DMEM (NIH3T3, HeLa, MEF), McCoy’s (HT-29), RPMI 1640 (293T, MCF-7, TNF-R1/2 double-deficient mouse fibroblasts) or a mixture of Click’s/RPMI 1640 medium (all other cell lines) supplemented with 10 % v/v calf serum and 10 mM glutamine in a humidified incubator containing 5 % w/v CO2. Media were additionally supplemented with 1 mM sodium pyruvate (HT-29, 293T, MCF7) and 50 μg/ml each of streptomycin and penicillin (except for 293T and MCF-7 cells).
Immunoblots. Unless otherwise indicated, cells were harvested after treatment and lysed at 4 °C in TNE buffer (50 mM Tris pH 8.0, 1 % v/v NP40, 2 mM EDTA) containing 10 μg/ml pepstatin/aprotinin/leupeptin, 1 mM sodium orthovanadate and 5 mM NaF. AIF was detected in purified cytosolic fractions generated with the ProteoJET (Fermentas, St. Leon-Rot, Germany) nuclear and cytoplasmic extraction kit. Identical amounts of cell protein per lane were resolved by electrophoresis on SDS polyacrylamide gels. After electrophoretic transfer to nitrocellulose, reactive proteins were detected using antisera specific for actin (sc-1615, Santa Cruz, Heidelberg, Germany; A1978, Sigma), PARP-1 (556362, Becton–Dickinson, Heidelberg, Germany; 9542, Cell Signaling, Danvers, MA, USA), TRADD (3684, Cell Signaling), FADD (7B5, Enzo), Caspase-8 (4927, Cell Signaling), RIP1 (610458, Becton–Dickinson), PAR (551813, Becton–Dickinson), AIF (sc-13116, Santa Cruz), COXIV (sc-58348, Santa Cruz), histone (ab1791, Abcam, Cambridge, UK), JNK1 (554286, Becton–Dickinson), JNK2 (sc-271133, Santa Cruz), phospho-JNK (AF1205, R&D Systems, Wiesbaden, Germany), TNF-R1 (sc-8436, Santa Cruz), human (PAB0287, Abnova, Heidelberg, Germany) or murine RIP3 (ADI-905-242, Enzo), GFP (8369-1, Takara), and the ECL detection kit (GE Healthcare, Munich, Germany). Equal loading as well as efficiency of transfer was routinely verified for all Western blots by Ponceau S staining, and by reprobing the membranes for actin.
Quantitative real-time polymerase chain reaction (qRT-PCR). Total RNA was isolated from L929Ts cells after induction of necroptosis with 100 ng/ml TNF for 0, 10, 15 and 20 h using the High Pure RNA isolation kit (Roche, Mannheim, Germany) and transcribed into single-stranded cDNA using the 1st Strand cDNA Synthesis Kit for RT-PCR (AMV, Roche) following the recommendations of the manufacturer. Quantification of PARP-1 expression levels was carried out with primers 5′-ATGCTACCACGCACAAC-3′ and 5′-CCAATCGGGTCTCCCT-3′ for murine PARP-1 with murine glyceraldehyde 3-phosphate dehydrogenase (primers 5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′) as internal standard, using a LightCycler System, the LightCycler–FastStart DNA Master SYBR Green I Kit and LightCycler software version 3.5 (Roche).
Measurement of intracellular NAD + and ATP levels. Cytosolic extracts were prepared in a buffer containing 0.5 % v/v Triton X-100, 10 mM Tris pH 7.5, and 1 mM EDTA. For measurement of NAD+, identical amounts of protein were incubated for 30 min at 37 °C with 100 μl assay buffer (8 % v/v ethanol, 1 % w/v polyvinylpyrrolidone, 0.5 mM EGTA, 50 mM Tris/HCl pH 8.5, 1 mM phenazine ethosulfate, 0.44 mM 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, 2 U alcohol dehydrogenase) before absorption was determined at 570 nm in a microplate reader (Tecan, Crailsheim, Germany). The cellular ATP content of the lysates was determined using a bioluminescence assay according to the manufacturer’s instructions (FL-AA, Sigma) in a FB-12 luminometer (Berthold, Bad Wildbad, Germany). Alternatively, intracellular ATP levels were determined with the Cell Titer Glo Assay Kit (Promega, Mannheim, Germany).
Flow cytometric analysis of membrane integrity. Cells were seeded in 12-well plates at 5 × 104 cells/well. Following treatment, both detached and adherent cells were collected by centrifugation. The cells were resuspended in PBS/5 mM EDTA containing 2 μg/ml propidium iodide (PI), and the red fluorescence was measured on a FACSCalibur flow cytometer (Becton–Dickinson).
RNA interference. The validated siRNA specific for human PARP1 (ID # s1097), the predesigned siRNA specific for murine PARP-1 (ID # s62053), the custom-made siRNA specific for murine RIP1 (target sequence 5′-CCACUAGUCUGACUGAUG-3′), the predesigned siRNA specific for murine RIP3 (ID # s80755) as well as the negative control siRNA were obtained from Life Technologies, Darmstadt, Germany. L929Ts and Jurkat cells were transfected with 150 pmol siRNA by Amaxa nucleofection (Lonza, Cologne, Germany), using solution V and program X-001, and incubated for 72 h at 37 °C before further analysis. NIH3T3 cells naturally expressing RIP3 were transfected as above, except for using solution R and program A-24, and incubation for 48 h.
TNF and MNNG activate PARP-1, but with distinct kinetics
In L929 cells, TNF elicits PCD exclusively by necroptosis but not by apoptosis [16, 17, 23]. Likewise, MNNG kills cells by necrosis rather than by apoptosis [14, 24]. Accordingly, three L929 cell lines available in our laboratory (TRAIL-sensitive L929Ts cells, “genuine” L929ATCC cells freshly obtained from ATCC and TNF-hypersensitive L929sA cells) did not display caspase activation or apoptotic membrane blebbing after treatment with TNF or MNNG (unlike TNF/CHX-treated positive controls, Supplementary Figure S1), confirming that in the L929 cell lines utilized for this study, both MNNG and TNF induce cell death through necrotic pathways rather than through caspase-dependent apoptosis.
We confirmed the TNF-dependent activation/disappearance of PARP-1 in time-course analyses not only in L929Ts but also in L929ATCC and L929sA cells (Fig. 1g). In the same three cell lines, the potent PARP1 activator MNNG  likewise induced the disappearance of PARP-1, however more rapidly, with changes already detectable within 4–8 h after treatment (Fig. 1g). For TNF, this occurred at 10 h of treatment or later, indicating that the initial steps of TNF-induced necroptosis do not require activation of PARP-1. Furthermore, in vitro addition of PARG reverted disappearance, size shift, and poly(ADP-ribosyl)ation of PARP-1 also in lysates from MNNG-treated L929Ts cells (Fig. 1h), confirming that the disappearance of PARP-1 in response to MNNG is likewise due to activation and auto-poly(ADP-ribosyl)ation of PARP-1 itself (as independently reported previously [14, 27, 28]).
DNA-alkylating agents but not TNF cause rapid activation of the PARP pathway
As a direct result of excessive NAD+ consumption by overactivated PARP-1, intracellular ATP is depleted by the cell’s efforts to resynthesize NAD+ . In line, both MNNG and MMS caused the depletion of the intracellular ATP stores of L929Ts, L929ATCC, and L929sA cells within 4 h, whereas TNF again did not induce a notable reduction (Fig. 2b). This discrepancy between TNF and DNA-alkylating agents was further corroborated when we compared ATP levels with the progression of cell death. Both MNNG and TNF led to a pronounced reduction of membrane integrity in all three L929 cell lines. Whereas the rapid loss of intracellular ATP induced by MNNG clearly preceded loss of membrane integrity, this occurred at later time points and concurrent with cell death for TNF (Fig. 2c).
Downstream of NAD+/ATP depletion, translocation of AIF from mitochondria to the cytosol and to the nucleus is essential in PARP-1-mediated necrosis [1, 28]. In concordance with our above results, MNNG elicited a pronounced translocation of AIF into the cytosol of L929Ts, L929ATCC and L929sA cells, whereas TNF caused only a marginal translocation of AIF in L929Ts and L929ATCC cells (more pronounced in TNF-hypersensitive L929sA cells), again occurring in the late to end rather than the early stages of cell death (24 h in L929Ts cells, 48 h in L929ATCC cells, 16 h in L929sA cells, Fig. 2d).
We additionally analyzed MNNG- and TNF-induced activation/disappearance of PARP-1, depletion of intracellular ATP and loss of membrane integrity in human leukemic Jurkat T cells and human HT-29 colorectal adenocarcinoma cells as two additional established cell systems for necroptosis [6, 17]. As shown in Fig. 2e, f, we obtained essentially the same results as with the three L929 sublines, confirming that the above findings are not limited to L929 cells. In summary, individual components of the PARP pathway are activated only late in TNF-induced necroptosis (in contrast to their rapid activation by DNA-alkylating agents), most likely as a secondary consequence, but not the cause of necroptotic cellular disintegration.
Inhibition of the PARP pathway does not protect against TNF-induced necroptosis
We extended our investigations to Jurkat and HT-29 cells, including murine NIH3T3 fibroblasts as a further cell systems for which TNF-induced necroptosis has been reported [7, 17]. In addition, we employed olaparib as a more potent and selective inhibitor of PARP-1 , and also the PARP-1 inhibitor PJ-34. As shown in Fig. 3d, all PARP-1 inhibitors (left panels) but not zVAD-fmk (right panel) effectively interfered with MNNG-induced necrosis. However, neither PARP-1 inhibitor reduced TNF(/zVAD/CHX)-induced necroptosis in any of the four cell lines (left panels). As the only exception, PJ-34 reduced TNF-mediated necroptosis in L929Ts cells. However, this was not observed in any of the other cell lines or with any of the other PARP-1 inhibitors and is therefore most likely due to unspecific activities that have been recently reported for PJ-34 [30, 31]. We obtained identical results after downregulation of PARP-1 in murine L929Ts and human Jurkat cells as well as in PARP-1-deficient MEF whereas MNNG-induced necrosis was clearly blocked (Fig. 3e).
JNK have been described as components of the PARP pathway required for mitochondrial dysfunction, AIF translocation, and necrosis . However, pharmacological inhibition of JNK did not affect TNF-induced necroptosis in L929Ts cells, even at concentrations at which the inhibitor itself induced toxicity (Fig. 3f). In a direct genetic approach, neither primary lung fibroblasts from JNK1- or JNK2-deficient nor immortalized MEF from JNK1/JNK2 double-deficient mice showed reduced sensitivities to TNF-induced necroptosis relative to their respective wild-type littermate controls (Fig. 3g), dissimilar to the previously reported increased survival of JNK1−/− and JNK2−/− MEF in response to MNNG . Furthermore, the RIP1 inhibitor necrostatin-1 failed to inhibit TNF-induced JNK activation in L929Ts cells, although necroptosis in response to TNF was completely blocked (Fig. 3h). Therefore, JNK do not participate in TNF-induced necroptosis, in contrast to their reported function in the PARP pathway .
As further constituents of the PARP pathway, calpain and cathepsin proteases cleave AIF to initiate its mitochondrio-nuclear translocation . In TNF-induced necroptosis, caspase inhibitors, such as the broad-spectrum caspase inhibitor zVAD, but not individual application of inhibitors for calpain, cathepsin B, L, S, papain and cystein proteases potentiate the death response [17, 19, 23]. Even when combined, the cathepsin B/L inhibitor zFA-fmk, the cathepsin B inhibitor CA-074 Me and the broad-spectrum calpain/cysteine protease inhibitor E-64 failed to enhance TNF-induced necroptosis in L929Ts cells (Fig. 3i). Moreover, they failed to protect from the enhanced necroptotic response elicited by TNF/zVAD (Fig. 3i), suggesting that calpains, cathepsins B and L and cysteine proteases neither positively nor negatively participate in TNF-induced necroptosis. In controls, addition of zFA-fmk, E-64 and CA-074Me reduced MNNG-induced necrosis whereas zVAD-fmk had no effect (Fig. 3i), confirming the reported role of calpains and cathepsins in the PARP pathway. Furthermore, pharmacological inhibition of cathepsin D by pepstatin A (Fig. 3j) or analysis of wild-type versus cathepsin D-deficient or cathepsin D-deficient MEF ectopically re-expressing cathepsin D (Fig. 3k) revealed no differences in the necroptotic response to TNF, in summary demonstrating that despite their reported role in the PARP pathway, calpain and cathepsin proteases are not required for TNF-induced necroptosis. In control reactions, although pharmacological inhibition of cathepsin D indeed reduced MNNG-elicited necrosis in L929Ts, L929ATCC and L929sA cells (Fig. 3j), cathepsin D-deficient MEF were not protected from MNNG-induced necrosis in comparison to their wild-type counterparts or to cathepsin D-deficient MEF ectopically re-expressing cathepsin D (Fig. 3k). These results suggest that pepstatin A is protective in the PARP pathway by targeting proteases other than cathepsin D and that cathepsin D plays no role in either TNF-induced necroptosis or the PARP pathway. Altogether, the above data suggest that TNF-induced necroptosis is not affected by inhibition of PARP-1 or subsequent components of the PARP pathway.
Interference with TNF-induced necroptosis does not block the PARP pathway
Inhibition of RIP1 by necrostatin-1 clearly protected from TNF-induced necroptosis in all three L929 sublines, but had no effect on the course of MNNG-induced necrosis over 16 h (Fig. 4b). Likewise, TNF/zVAD/CHX-induced necroptosis was reduced by necrostatin-1 in both wild-type and PARP-1-deficient MEF, whereas their differential response towards MNNG-induced necrosis remained completely unaffected (Fig. 4c). RIP1-deficient MEF were fully resistant to TNF/zVAD/CHX-induced necroptosis, but not noticeably protected from MNNG (Fig. 4d). Downregulation of RIP1 in NIH3T3 cells (that naturally express RIP3) , protected from TNF/zVAD-induced necroptosis, but not from MNNG (Fig. 4f). RIP1-deficient Jurkat cells proved fully resistant against TNF/zVAD/CHX-induced necroptosis, but, other than MEF and NIH3T3 cells, displayed a transient, statistically significant protection from MNNG, however not preventing final death (Fig. 4h). Absence or downregulation of RIP1 did not prevent the MNNG-induced rapid loss of intracellular ATP as an early step of the PARP pathway (Figs. 4e, g, i), in line with previous studies suggesting that RIP1 may also function in the PARP pathway, however at later stages [14, 27] rather than in the central, initial steps (as in TNF-induced necroptosis).
In line with the crucial function of RIP3 in necroptosis [6, 7, 10], TNF(/zVAD/CHX) elicited necroptosis in L929Ts and NIH3T3 cells expressing RIP3, but not in 293T, MCF-7 and HeLa cells that lack endogenous RIP3 expression. In contrast, both MNNG and MMS uniformly induced necrosis in all cell lines (Fig. 4j). To exclude artifacts from these different constitutively RIP3-positive or -negative cell lines, we overexpressed wild-type or kinase-defective RIP3 in naturally RIP3-deficient NIH3T3 cells  and downregulated RIP3 in NIH3T3 cells that naturally express RIP3 . Similar to RIP1-deficient Jurkat cells (Fig. 4h), RIP3-deficient NIH3T3 cells displayed a statistically significant transient attenuation, but not protection from MNNG, while being fully resistant against TNF/zVAD-induced necroptosis (Fig. 4k). Stable re-expression of wild-type, but not kinase-defective RIP3 restored full sensitivity to TNF/zVAD [6, 10]. In contrast, sensitivity for MNNG was restored by re-expression of both wild-type and kinase-defective RIP3, indicating that the kinase activity of RIP3 is not required for its function in the PARP pathway. In reverse, downregulation of RIP3 protected naturally RIP3-expressing NIH3T3 cells from TNF/zVAD-induced necroptosis while only transiently (but statistically significant) attenuating MNNG-induced necrosis (Fig. 4m). In primary MEF or primary lung fibroblasts from RIP3-deficient mice, TNF/zVAD/CHX-induced necroptosis was completely blocked, whereas the course of MNNG-induced necrosis was not altered (Fig. 4o, q). As for RIP1, the rapid loss of ATP during MNNG-induced necrosis occurred identically in RIP3-deficient, -downregulated or -retransfected cells (Fig. 4l, n, p, r), implicating that RIP3 also does not participate in the initial but later steps of the PARP pathway.
Consistent with the above studies [14, 27], our data suggest a function of RIP1 in both TNF-induced necroptosis and MNNG-mediated necrosis. However, while RIP1 is clearly crucial for TNF-induced necroptosis in a cell type-independent manner, its contribution to MNNG-induced necrosis appears to be of more limited importance and cell type-specific. Although previously not investigated, our data indicate that the same is also true for RIP3. Notably, Xu et al.  have previously ruled out that the function of RIP1 in MNNG-induced necrosis depends on the TNF signaling pathway by demonstrating that MEF deficient for TNF-R1 remain as sensitive to MNNG as wild-type MEF. We extended this experiment by showing that the course of MNNG-induced necrosis was identical in both TNF-R1/TNF-R2 double-deficient fibroblasts and TNF-R1/TNF-R2 double-deficient control fibroblasts that ectopically re-express TNF-R1, thus clearly demonstrating that any role of RIP1 or RIP3 in the PARP pathway cannot depend on signals mediated by TNF-R1 (Fig. 4s). In combination with the kinase-independent function of RIP3 in the PARP pathway (contrasting TNF-dependent necroptosis), our results demonstrate that although RIP1 and RIP3 participate in both TNF-induced necroptosis and the PARP pathway, their functions in each pathway are independent and do not connect the two pathways.
Finally, scavengers of ROS such as BHA can protect L929 cells from TNF-induced necroptosis [17, 23]. Correspondingly, BHA decreased necroptosis in TNF-treated L929Ts, L929ATCC and L929sA cells, but failed to protect against necrosis induced by MNNG and MMS in any of the three cell lines (Fig. 4t).
In this study, we present evidence that TNF-induced necroptosis and the PARP pathway represent two distinct and independent routes to programmed necrosis. The importance of programmed necrosis, especially in pathological situations, becomes increasingly obvious and the possibility for therapeutic suppression of necrosis has raised great expectations . A role in the pathology of diseases has been reported for both necroptosis and PARP-1-dependent necrosis. Inhibition of necroptosis by targeting RIP1 effectively reduced renal , cardiac and retinal ischemia/reperfusion injury , limited brain damage, and improved motor and cognitive performance in mouse models . Similarly, genetic deficiency for RIP3 limited tissue damage in a pancreatitis model . PARP-1-dependent necrosis plays a central role in inflammatory responses in cells of the central nervous, cardiovascular, and immune systems. PARP-1 inhibitors have been effective in the treatment of ischemia/reperfusion damage after stroke or brain, myocardial, hepatic, and renal ischemia/reperfusion injury, other acute forms of cardiomyopathies and heart failure, tissue injury in response to septic and hemorrhagic shock, as well as acute lung inflammation, peritonitis, and pancreatitis. Inhibition of the PARP pathway has also demonstrated benefit in chronic disease models, e.g., cardiovascular aging, atherosclerosis, and diabetic cardiovascular complications. Furthermore, the PARP pathway represents a promising target for treatment of autoimmune conditions, including type 1 diabetes, glomerulonephritis, multiple sclerosis, and rheumatoid arthritis [1, 25, 33].
Notably, Jouan-Lanhouet et al.  have recently reported that TRAIL induces necroptosis at acidic extracellular pH (pHe) involving RIP1/RIP3-dependent PARP-1 activation, seeming to be at odds with our results at first glance. However, the same group has previously reported that in spite of a necrosis-like cell death morphology, caspase activity is required for cell death in this system and that both zVAD-fmk and the caspase-8 inhibitor zIETD-fmk fully block TRAIL-induced cell death at acidic pHe . This dependence on caspases is however not consistent with the molecular mechanisms described for necroptosis  but rather points to a mode of cell death different from TNF-induced necroptosis studied here, where we have never observed an inhibitory effect of zVAD-fmk in the many experiments where it was added. Therefore, an involvement of PARP-1 in the TRAIL-induced mode of cell death described by Jouan-Lanhouet is feasible based on the data presented in their study, but does not implicate a similar role of PARP-1 in TNF-induced necroptosis as studied here. Also, such a role still would have to be validated for true necroptosis induced by TRAIL, e.g., in the presence of zVAD-fmk, and at physiologic pHe (7.4). Moreover, the effects described in the study of Jouan-Lanhouet et al. do not occur at physiologic pHe but are exclusively limited to acidic conditions, again different from our observations for TNF-induced necroptosis. Finally, Jouan-Lanhouet et al.  themselves demonstrate that their findings do not apply to necroptosis induced by either TNF or FasL, regardless of physiologic or acidic pHe, again suggesting that the conclusions from their study cannot be easily generalized for other death receptors or physiologic conditions.
Of particular importance, this study provides a first answer to the recently raised question whether there is one core program or several independent pathways of necrosis, with implications for the future development of necroptosis-inhibitory cytoprotective drugs as well as for the design of therapeutic strategies. The existence of several distinct pathways that trigger necrosis implies that an efficient protection requires their simultaneous interruption via combination therapies . As a consequence of the results obtained in this study, we suggest that the simultaneous rather than the individual application of highly specific necroptosis and PARP-inhibitors (e.g., necrostatin and olaparib) may be the approach of choice for a future, improved treatment of necrosis-induced damage in diseases such as shock, stroke, or myocardial infarction.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 415, Project A4 and Cluster of Excellence “Inflammation at Interfaces”, EXC306-PMTP1, both to D. A., and SFB 877, Project B2, to D. A. and S. S.) and by a fellowship from the German Academic Exchange Service (DAAD, A/08/79433, to J. S.).
Conflict of interest
The authors declare that they have no conflicts of interest.
- 11.Mocarski ES, Upton JW, Kaiser WJ (2012) Viral infection and the evolution of caspase 8-regulated apoptotic and necrotic death pathways. Nat Rev Immunol 12(2):79–88Google Scholar
- 15.Los M, Mozoluk M, Ferrari D, Stepczynska A, Stroh C, Renz A, Herceg Z, Wang ZQ, Schulze-Osthoff K (2002) Activation and caspase-mediated inhibition of PARP: a molecular switch between fibroblast necrosis and apoptosis in death receptor signaling. Mol Biol Cell 13:978–988PubMedCentralPubMedCrossRefGoogle Scholar
- 22.Schneider-Brachert W, Tchikov V, Neumeyer J, Jakob M, Winoto-Morbach S, Held-Feindt J, Heinrich M, Merkel O, Ehrenschwender M, Adam D, Mentlein R, Kabelitz D, Schütze S (2004) Compartmentalization of TNF receptor 1 signaling; internalized TNF receptosomes as death signaling vesicles. Immunity 21:415–428PubMedCrossRefGoogle Scholar
- 34.Seok JH, Park KA, Byun HS, Won M, Shin S, Choi BL, Lee H, Kim YR, Hong JH, Park J, Hur GM (2008) Long-term activation of c-Jun N-terminal kinase through receptor interacting protein is associated with DNA damage-induced cell death. Korean J Physiol Pharmacol 12(4):185–191PubMedCentralPubMedCrossRefGoogle Scholar
- 42.Slamenová D, Horváthova E, Robichová S, Hrusovská L, Gábelova A, Kleibl K, Jakubiková J, Sedlák J (2003) Molecular and cellular influences of butylated hydroxyanisole on Chinese hamster V79 cells treated with N-methyl-N’-nitro-N-nitrosoguanidine: antimutagenicity of butylated hydroxyanisole. Environ Mol Mutagen 41(1):28–36PubMedCrossRefGoogle Scholar
- 43.Jouan-Lanhouet S, Arshad MI, Piquet-Pellorce C, Martin-Chouly C, Le Moigne-Muller G, Van Herreweghe F, Takahashi N, Sergent O, Lagadic-Gossmann D, Vandenabeele P, Samson M, Dimanche-Boitrel MT (2012) TRAIL induces necroptosis involving RIPK1/RIPK3-dependent PARP-1 activation. Cell Death Differ 19(12):2003–2014PubMedCrossRefGoogle Scholar
- 44.Meurette O, Rebillard A, Huc L, Le Moigne G, Merino D, Micheau O, Lagadic-Gossmann D, Dimanche-Boitrel MT (2007) TRAIL induces receptor-interacting protein 1-dependent and caspase-dependent necrosis-like cell death under acidic extracellular conditions. Cancer Res 67(1):218–226PubMedCrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.