Leading Edge
Previews
Necroptosis: A Specialized Pathway
of Programmed Necrosis
Lorenzo Galluzzi1,2,3 and Guido Kroemer1,2,3,*
1
INSERM, U848, F-94805 Villejuif, France
2
Institut Gustave Roussy, F-94805 Villejuif, France
3
Universit Paris-Sud/Paris XI, F-94270 Le Kremlin Bictre, France
*Correspondence: kroemer@igr.fr
DOI 10.1016/j.cell.2008.12.004
Necrosis is often viewed as an accidental and unregulated cellular event. However, accumulating
evidence suggests that necrosis, like apoptosis, can be executed by regulated mechanisms. Hitomi
et al. (2008) now describe an extensive network of genes that mediate a form of programmed
necrosis called necroptosis.
On the basis of genetic analysis in the
nematode Caenorhabditis elegans, it has
been postulated that developmental and
homeostatic programmed cell death,
as it occurs in higher animals, is usu-
ally mediated by apoptosis. Apoptosis is
defined by an ensemble of morphologi-
cal features, including chromatin con-
densation and nuclear fragmentation,
cell shrinkage, plasma membrane bleb-
bing, and formation of apoptotic bod-
ies. In contrast, necrosis fails to display
a stereotyped morphology (except for
the early rupture of plasma membranes)
and has historically been regarded as an
unregulated means of cell death that is
induced by nonspecific and nonphysi-
ological stress (Kroemer et al., 2008).
Defying this view, Hitomi et al. (2008) now
report the existence of a complex molec-
ular pathway mediating programmed
necrosis that they have unraveled using
a systems biology approach.
Multiple lines of evidence indicate
that necrosis can be a programmed
event, both in its occurrence and in its
mechanism: (1) cell death with a necrotic
appearance can contribute to embryonic
development and adult tissue homeosta-
sis, (2) necrotic cell death can be induced
by ligands that bind to specific plasma
membrane receptors, and (3) necrosis
can be regulated by genetic, epigenetic,
and pharmacological factors (Golstein
and Kroemer, 2007). Moreover, the inac-
tivation of caspases causes a shift from
apoptosis either to cell death morpholo-
gies with mixed necrotic and apoptotic
features or to full-blown necrosis (Kroe-
mer et al., 2008). This proves that differ-
ent cell death modalities crossregulate
each other and substantiates the notion
that necrosis may constitute a default
cell death pathway that is unmasked
when essential effectors of apoptosis are
inhibited (Golstein and Kroemer, 2007).
Recently, the term necroptosis has
been used to designate one particu-
lar type of programmed necrosis that
depends on the serine/threonine kinase
activity of RIP1 (Kroemer et al., 2008).
Indeed, RIP1 represents the molecular
target of a new class of cytoprotective
agents, the necrostatins (Degterev et al.,
2005). In their current work, Hitomi et al.
(2008) apply systems biology to study two
specific cases of necroptosisnamely,
the necrotic demise of L929 mouse fib-
rosarcoma cells induced either by liga-
tion of the tumor necrosis factor (TNF)
receptor or by caspase inhibition with
Z-VAD.fmk [Z-Val-Ala-Asp(OMe).fluorom-
ethylketone]. These responses were then
compared with classical apoptosis in
NIH 3T3 murine fibroblasts, as triggered
by TNF receptor ligation. Genome-wide
screens using small interfering RNAs
(siRNAs) combined with multiple in silico
analyses delineated a cellular signal-
ing network that regulates necroptosis
and the molecular bifurcation between
necroptosis and apoptosis.
The morphological uniformity of
apoptosis hides a heterogeneous array
of lethal processes that differ in their
molecular determinants and functional
impact, leading for instance to the dis-
tinction between intrinsic (mitochondrial)
and extrinsic (death receptor-mediated)
apoptotic pathways (Kroemer et al.,
Cell 135, December 26, 2008 2008 Elsevier Inc. 1161
2008). Hence, an important problem is
whether necrosis is the manifestation
of a uniform process or of a collection
of cell death subroutines. The pharma-
cological or genetic inhibition of several
key enzymes has been shown to deeply
affect the execution of programmed
necrosis (Figure 1). These include RIP1,
cyclophilin D, poly(ADP-ribose) poly-
merase 1 (PARP-1), and apoptosis-
inducing factor (AIF). RIP1 kinase activ-
ity is required for death receptor- and
Z-VAD.fmk-mediated necroptosis in
both murine and human cells (Holler et
al., 2000), and necrostatins confer in
vivo neuroprotection against ischemia,
a condition that elevates necrotic stimuli
such as reactive oxygen species and
Ca2+ overload (Degterev et al., 2005).
Mice lacking cyclophilin D (a mitochon-
drial peptidylprolyl cis-trans isomerase
that contributes to the so-called perme-
ability transition pore complex [PTPC])
are substantially more resistant to car-
diac and cerebral ischemic injury than
their wild-type littermates (Nakagawa
et al., 2005). Excessive activation of the
DNA repair protein PARP-1 has been
implicated in necrosis triggered by alky-
lating DNA damage (Zong et al., 2004).
In rodent models of ischemia, neuropro-
tective effects are exerted by genetic
inhibition of calpains (which are Ca2+-
activated proteases) and of the caspase-
independent death effector AIF (Golstein
and Kroemer, 2007).
Some evidence suggests that these
elements are deeply interconnected.
RIP1 can promote the PTPC-dependent
mitochondrial permeability transition
(which is also triggered by reactive oxy-
gen species and Ca2+) by inducing the
accumulation of pronecrotic ceramides
or by binding to another PTPC compo-
nent, the adenine nucleotide translocase
(Golstein and Kroemer, 2007). Moreover,
Ca 2+-activated calpains favor the release
of AIF from mitochondria, as well as the
spillage of lysosomal cathepsins into the
cytosol (Golstein and Kroemer, 2007).
PARP-1 activation stimulates the mito-
chondrial release of AIF, and AIF-medi-
ated DNA degradation further activates
PARP-1 (Yu et al., 2002), thereby engag-
ing a vicious cycle. The implication of
PARP proteins in programmed necrosis
is further substantiated by the findings of
Hitomi et al. (2008), which reveal PARP-2
as one of the core regulators of necrop-
tosis.
Mitochondrial membrane permeabili-
zation contributes to cell death and can
either derive from the mitochondrial per-
meability transition (which first affects
the mitochondrial inner membrane) or
from mitochondrial outer membrane per-
meabilization, which is mediated by Bax
or Bak, two proapoptotic members of
the Bcl-2 protein family (Kroemer et al.,
2007). Reportedly, L929 cells undergo-
ing necroptosis do not exhibit mitochon-
drial outer membrane permeabilization
(Golstein and Kroemer, 2007). Moreover,
the knockout of ppif (the gene coding
Figure 1. The Interface between Apoptosis and Programmed Necrosis
for cyclophilin D) in mice fails to prevent
The programmed necrosis pathway studied by Hitomi et al., termed necroptosis, centers on the activa-
apoptosis that is dependent on permea- tion of the serine/threonine kinase RIP1 (receptor-interacting protein kinase 1), which can be triggered in
bilization of the mitochondrial outer mem- L929 murine fibrosarcoma cells by ligation of the tumor necrosis factor receptor (TNFR) or inhibition of
caspases. In NIH 3T3 murine fibroblasts, TNFR activation ignites the extrinsic apoptotic pathway, which
brane (Nakagawa et al., 2005). None- depends on caspase-8. Caspase-8-mediated degradation of RIP1 may represent one of the major mo-
theless, necroptosis is linked to rapid lecular switches between apoptosis and necroptosis. Apoptosis and necroptosis may preferentially in-
mitochondrial dysfunction that leads volve mitochondrial outer membrane permeabilization (MOMP) and the mitochondrial permeability transi-
tion (MPT), respectively. The figure has been assembled from the data reported by Hitomi et al. (2008) in
to the excessive production of reactive this issue, as well as from previous publications on programmed necrosis. For the sake of clarity, multiple
oxygen species. Thus, it remains for- intermediate regulators of apoptosis have not been depicted. Abbreviations: AIF, apoptosis-inducing fac-
mally possible that a sudden breakdown tor; ANT, adenine nucleotide translocase; Cyt c, cytochrome c; GPI, glycosylphosphatidylinositol; IFN, in-
terferon; LMP, lysosomal membrane permeabilization; PTPC, permeability transition pore complex; ROS,
of mitochondrial function (which causes reactive oxygen species; TLR, Toll-like receptor; TNF, tumor necrosis factor ; TNFR, tumor necrosis
a drop in intracellular ATP), perhaps sus- factor receptor; Z-VAD.fmk, Z-Val-Ala-Asp(OMe).fluoromethylketone.
tained by excessive activation of PARP
proteins (which causes the depletion of its interaction with (and inhibition of) the (Golstein and Kroemer, 2007; Kroemer et
NAD+), might account for the bioener- antiapoptotic proteins Bcl-2 and Bcl- al., 2007). Thus, it is tempting to specu-
getic crisis that constitutes the final step XL (Puthalakath et al., 2001). The exact late that Bmf (and maybe other BH3-only
of programmed necrosis. molecular mechanisms by which Bmf proteins) might promote apoptosis by
Surprisingly, Hitomi et al. identified the promotes apoptosis versus necropto- relieving Bcl-2/Bcl-XL-mediated inhibition
BH3-only protein Bmf as a component sis remain to be clarified. Antiapoptotic of mitochondrial outer membrane per-
of the core machinery for necroptosis. Bcl-2 proteins can suppress both apop- meabilization, and necrosis by interfering
In healthy cells, Bmf is sequestered to tosis and necrosis, and inhibit both mito- with the Bcl-2/Bcl-XL-mediated blockage
cytoskeletal structures by virtue of its chondrial outer membrane permeabiliza- of the mitochondrial permeability transi-
interaction with dynein light chain 2. tion (by sequestering Bax and Bak) and tion (Figure 1). Hitomi et al. (2008) failed
Under conditions of cellular distress, the the mitochondrial permeability transition to identify cyclophilin D (or any other
release of Bmf from this dock facilitates (through their interaction with the PTPC) protein from the PTPC) in their screen for

1162 Cell 135, December 26, 2008 2008 Elsevier Inc.

genes involved in necroptosis. Whether
this is due to methodological limitations
or rather hints at a hitherto unsuspected
mechanism of Bmf-dependent cell death
remains to be established.
future will tell which among the hits iden-
tified by Hitomi et al. will serve as new
targets for the therapeutic manipulation
of programmed necrosis. Indeed, sev-
eral of the necroptosis-relevant proteins
this issue.
Holler, N., Zaru, R., Micheau, O., Thome, M., At-
tinger, A., Valitutti, S., Bodmer, J.L., Schneider, P.,
Seed, B., and Tschopp, J. (2000). Nat. Immunol.
1, 489495.

The challenge of characterizing the
precise mechanisms of programmed
necrosis, as well as of the molecular
switches between apoptosis and necro-
sis, has major therapeutic implications.
In some instances, the selective inhibi-
tion of necrosis (and/or the facilitation of
apoptotic cell death) may limit inflamma-
tion, and hence reduce secondary tissue
damage. Conversely, it may be desirable
to trigger the necrotic death of cancer
cells that are resistant to apoptosis.
Necrostatins have proved their therapeu-
tic potential in a murine model of stroke
(Degterev et al., 2005). Similarly, phar-
macological and/or genetic inhibition
of cyclophilin D, AIF, PARP-1, calpains,
and cathepsins afford cytoprotection in
vivo in several models of acute cell loss
(Golstein and Kroemer, 2007). Only the
identified, including proteins belonging
to the interferon and Toll-like receptor
signaling systems, have already been
accused of mediating pathological cell
death in vivo.
Acknowledgments
We thank O. Kepp for expert help in preparation
of the figure.
References
Degterev, A., Huang, Z., Boyce, M., Li, Y., Jagtap,
P., Mizushima, N., Cuny, G.D., Mitchison, T.J.,
Moskowitz, M.A., and Yuan, J. (2005). Nat. Chem.
Biol. 1, 112119.
Golstein, P., and Kroemer, G. (2007). Trends Bio-
chem. Sci. 32, 3743.
Hitomi, J., Christofferson, D.E., Ng, A., Yao, J., De-
gterev, A., Xavier, R.J., and Yuan, J. (2008). Cell,
Kroemer, G., Galluzzi, L., and Brenner, C. (2007).
Physiol. Rev. 87, 99163.
Kroemer, G., Galluzzi, L., Vandenabeele, P.,
Abrams, J., Alnemri, E.S., Baehrecke, E.H.,
Blagosklonny, M.V., El-Deiry, W.S., Golstein,
P., Green, D.R., et al. (2008). Cell Death Differ.,
in press. Published online October 10, 2008.
10.1038/cdd.2008.150.
Nakagawa, T., Shimizu, S., Watanabe, T., Yamagu-
chi, O., Otsu, K., Yamagata, H., Inohara, H., Kubo,
T., and Tsujimoto, Y. (2005). Nature 434, 652658.
Puthalakath, H., Villunger, A., OReilly, L.A., Beau-
mont, J.G., Coultas, L., Cheney, R.E., Huang, D.C.,
and Strasser, A. (2001). Science 293, 18291832.
Yu, S.W., Wang, H., Poitras, M.F., Coombs, C.,
Bowers, W.J., Federoff, H.J., Poirier, G.G., Daw-
son, T.M., and Dawson, V.L. (2002). Science 297,
259263.
Zong, W.X., Ditsworth, D., Bauer, D.E., Wang,
Z.Q., and Thompson, C.B. (2004). Genes Dev. 18,
12721282.

The Long and Short of Membrane Fission

Aurlien Roux1,* and Bruno Antonny2,*
1
Laboratoire Physico-Chimie Curie, CNRS, Institut Curie, 75248 Paris Cedex 05, France
2
Institut de Pharmacologie Molculaire et Cellulaire, Universit de Nice Sophia Antipolis, CNRS, 06560 Valbonne, France
*Correspondence: aurelien.roux@curie.fr (A.R.), antonny@ipmc.cnrs.fr (B.A.)
DOI 10.1016/j.cell.2008.12.003

During membrane fission, the GTPase dynamin forms helical assemblies at the neck of membrane
buds. Although it has been proposed that fission results from the constriction of the dynamin helix,
new work by Bashkirov et al. (2008) and Pucadyil and Schmid (2008) unexpectedly shows that
helix disassembly is also necessary for membrane fission.

Membrane fission and membrane fusion
are the two processes fundamental to
membrane trafficking. Membrane fission
is the process by which a bud separates
from a lipid membrane. The best-studied
fission protein is the GTPase dynamin
(Praefcke and McMahon, 2004), which
forms collar structures at the neck of
nascent buds (Takei et al., 1995) and
can spontaneously polymerize into heli-
cal polymers in the presence or absence
of membranes (Sweitzer and Hinshaw,
1998). These findings immediately sug-
gest a mechanism for fission: the helix
of dynamin would undergo a confor-
mational change upon GTP hydrolysis
to break the bud neck. Three different
movements of the helix have been pro-
posed: constriction (Sweitzer and Hin-
shaw, 1998), expansion (Stowell et al.,
1999), or twisting (Roux et al., 2006).
These mechanisms were initially thought
to be incompatible, but theories unify-
ing them have since been proposed
(Roux et al., 2006; Lenz et al., 2008).
These theories suggest that because
dynamin is a right-handed helix, further
right-handed twisting would induce con-
striction. In this issue, Bashkirov et al.
(2008) and Pucadyil and Schmid (2008)
now propose a different mechanism for
fission whereby depolymerization of
the dynamin helix after GTP hydrolysis
enables bud separation. In this model,
the dynamin helix, having brought the
lipids in the neck region of the bud into

Cell 135, December 26, 2008 2008 Elsevier Inc. 1163