A Perspective on Mammalian Caspases
as Positive and Negative Regulators of Inflammation
Seamus J. Martin,1,2,* Conor M. Henry,1and Sean P. Cullen1,2
1Molecular Cell Biology Laboratory, Department of Genetics, The Smurfit Institute
2Immunology Research Centre
Trinity College, Dublin 2, Ireland
Members of the caspase family of cysteine proteases coordinate the morphological and biochemical events
thattypify apoptosis. However,neutralization ofcaspase activityin mammalsfailsto block deathin response
to most proapoptotic stimuli. This is because many cell death triggers provoke mitochondrial dysfunction
upstream of caspase activation as a consequence of BAX/BAK channel opening. Although genetic or phar-
macological inactivation of caspases fails to block cell death in most instances, it does convert the pheno-
type from apoptosis to necrosis. This has important implications for how the immune system responds to
such cells, as necroticcells provoke inflammation whereas apoptotic cells typicallydo not. Here, we propose
an alternativeperspective on apoptosis-associated caspase function by suggesting that theseproteases are
activated, not to kill, but to extinguish the proinflammatory properties of dying cells. This perspective unifies
the mammalian caspase family as either positive or negative regulators of inflammation.
Apoptosis is a mode of programmed cell death that is used to
of disturbance to neighboring cells (Green, 2010). Apoptosis
complements mitosis as a means of regulating cell numbers in
multicellular organisms and for this reason is under molecular
control by a dedicated set of enzymes and their regulators—
the ‘‘cell death machinery.’’ Members of a family of cysteine
proteases, the caspases, become activated during apoptosis
and coordinate the events that take place to ensure swift recog-
nition and removal of apoptotic cells (Riedl and Salvesen, 2007;
Taylor et al., 2008). However, although it is well established that
caspases are required for the appearance of the major biochem-
ical and morphological ‘‘hallmarks’’ of apoptosis, it is now clear
that in many situations caspases are not required for terminating
cell viability in mammals (Chipuk and Green, 2005; Kroemer and
Martin, 2005). Because the great majority of injurious stimuli that
promote caspase activation do so by promoting permeabiliza-
tion of the mitochondrial outer membrane (Green and Kroemer,
2004), the latter event is usually sufficient to ensure cell death
irrespective of whether caspases are activated downstream or
not. Despite this, the caspases activated during apoptosis are
still typically viewed as ‘‘death effectors,’’ but much evidence
now points toward a more complex role for these proteases as
regulators of the inflammatory potential of apoptotic cells rather
than as arbiters of cell fate.
An Alternative Perspective on Apoptosis-Associated
Here we propose that the primary function of apoptosis-associ-
ated caspase activation in mammals is the avoidance of proin-
flammatory engagement ofthe immunesystem and its attendant
detrimental consequences—primarily autoimmunity—by cells
undergoing programmed elimination. There is a growing body
of evidence to argue that the caspase-dependent alterations to
the cell that typically occur during apoptosis not only ensure
recognition and uptake of apoptotic cells by phagocytes, but
also switch on an anti-inflammatory program in the engulfing
cell. Furthermore, there is also much evidence that apoptotic
cells, even those undergoing secondary necrosis, can actively
antagonize and override concurrent proinflammatory signals
delivered to phagocytes that have engulfed an apoptotic cell
(reviewed in Birge and Ucker, 2008).
Inthisreview, wewilldiscussevidenceto supportthe ideathat
caspases activated during apoptosis function to extinguish or
dampen the proinflammatory properties of dying cells through
directly inactivating and coordinating the sequestration of
numerous potentially proinflammatory molecules, collectively
called danger-associated molecular patterns (DAMPS) or
alarmins, that reside within (Matzinger, 1994; Kono and Rock,
2008). Where death occurs without caspase activation, we
propose that this results in a failure to sequester and inactivate
endogenous DAMPs and may lead to local activation of macro-
phages and dendritic cells (DCs), the key antigen-presenting cell
of the immune system. The latter event is particularly dangerous
because, apart from unnecessary engagement of the immune
system where a pathogenic threat is not present, DC activation
can ‘‘license’’ these cells to present self-antigens to T cells
(reviewed in Kono and Rock, 2008). This scenario is particularly
undesirable from an immunological standpoint, as this can lead
to a loss of tolerance toward ‘‘self’’ that is essential to safeguard
against potentially catastrophic autoimmune responses. Cas-
pase activation during cell death can thus be interpreted as
a means of conveying information to the immune system con-
planned or pathogen induced), rather than simply a means to
terminate cell viability. We suggest that apoptosis-associated
caspase activation serves primarily to coordinate the silent
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
removal of dying cells by actively repressing the mechanisms
normally used by the immune system to recognize and respond
to inappropriate, nonprogrammed, cell deaths. From this
perspective, apoptosis-associated caspase activation serves
a predominantly anti-inflammatory role, as opposed to a cell
The Classical View: Caspases as Regulators of Death
Mammals possess multiple caspases and these have tradition-
ally been split into two major subgroups based upon sequence
homology and function. The ‘‘inflammatory’’ caspases, which
belong to the caspase-1-related subset, are activated in cells
of the innate immune system, such as macrophages and DCs,
in response to infection as well as noxious agents that trigger
necrosis (Creagh et al., 2003; Schroder and Tschopp, 2010).
Caspase-1 subfamily members (i.e., caspase-1, caspase-4,
and caspase-5) have been implicated as regulators of inflamma-
tion through processing and activating two related cytokines,
IL-1b and IL-18, which have diverse effects and act to initiate
and amplify immune responses to infectious agents (Creagh
et al., 2003). In addition, the murine caspase-4 ortholog,
caspase-11, has recently been found to be important for
caspase-1-dependent IL-1b and IL-18 production in response
to a subset of inflammatory triggers (Kayagaki et al., 2011).
However, the precise role of caspase-4 and caspase-5 in inflam-
mation remains to be further clarified.
The ‘‘apoptotic’’ caspases, which belong to the caspase-3-
related subgroup, are activated during apoptosis and are widely
2008). To date, over 600 substrates for the cell death-related
caspases have been identified (Lu ¨thi and Martin, 2007).
However, the vast majority of these substrates, with some
notable exceptions, have not been linked with any specific
feature of apoptosis (Taylor et al., 2008). Furthermore, the failure
to cleave particular caspase substrates rarely, if ever, permits
a cell to survive the events that precede caspase activation.
One possibility is that these enzymes are engaging in a policy
of redundancy; targeting numerous vital molecules to ensure
that cell death is guaranteed. If this is so, it is impressive for
the scale of overkill employed. Alternatively, it is also plausible
that many of the proteins that are cleaved during apoptosis are
either postmortem events or ‘‘innocent bystander’’ cleavage
events that have no significance for the process and take place
when cell viability has already effectively been terminated. Either
way, given the sheer number of proteins that are cleaved during
apoptosis, it seems highly unlikely that the majority of these
proteins are targeted solely for the purpose of terminating cell
viability. Moreover, because the events that lead to activation
of the major ‘‘executioner’’ caspases irreversibly disrupt mito-
chondrial function (Green and Kroemer, 2004), an event that is
sufficient to ensure that most cells will die, it seems implausible
that caspases become activated merely to compound an
already fatal blow. Before we outline why caspases are dispens-
able for cell death, it is necessary to take a brief look at how
caspases become activated during apoptosis.
Cell Death within the Intrinsic Pathway to Apoptosis
A major route to caspase activation and apoptosis results from
cellular stresses—such as cytokine deprivation, heat shock,
and DNA damage, all of which provoke mitochondrial outer
membrane permeabilization (MOMP) and the release of cyto-
chrome c and other mitochondrial constituents into the cytosol
(Figure 1). This has been dubbed the intrinsic or mitochondrial
pathway to caspase activation and is employed by numerous
proapoptotic stimuli (Green and Kroemer, 2004; Youle and
Strasser, 2008). For simplicity, we will focus predominantly on
the intrinsic pathway to caspase activation, but we will also refer
to the second major route to caspase activation, the extrinsic or
death receptor pathway, toward the end of the review.
MOMP is achieved through opening of a mitochondrial
membrane channel comprised of the related proteins, BAX and
BAK, as a consequence of transcriptional upregulation or post-
translational modification of one or more members of the
‘‘BH3-only’’ protein family (Figure 1). BH3-only proteins are
activated in response to diverse triggers of apoptosis, the details
of which are outside of the scope of this review but have been
well documented elsewhere (Youle and Strasser, 2008). Acti-
vated BH3-only proteins either directly or indirectly promote
conformational changes within BAX and BAK that provoke their
Figure 1. BAX/BAK-Induced MOMP Is Sufficient to Kill
BH3-only proteins act as specific sensors for various apoptotic stimuli and
promote assembly of BAX/BAK oligomomers within the mitochondrial outer
membrane, leading to cytochrome c release, apoptosome formation, caspase
activation, and apoptosis. Importantly, caspase inhibition downstream of
cytochrome c release does not protect against cell death as ATP depletion,
reactive oxygen species generation, and subsequent mitochondrial dysfunc-
tion, brought about by release of mitochondrial intermembrane space
proteins, leads to death by necrosis.
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
oligomerization within the outer mitochondrial membrane. The
latter event results in the formation of a pore or channel that
permits the escape of numerous mitochondrial intermembrane
space proteins into the cytosol (Green and Kroemer, 2004).
The most notable of these is cytochrome c, as this protein acts
as a cofactor for the assembly of a caspase-activating complex
within the cytosol that has been dubbed ‘‘the apoptosome.’’
Binding of cytochrome c to Apaf-1 triggers its oligomerization
into a wheel-like structure and permits recruitment, homodime-
rization, and activation of caspase-9 within the Apaf-1 apopto-
some (Riedl and Salvesen, 2007). In turn, the apoptosome
activates caspase-3 and caspase-7, setting off a chain of
caspase activation events downstream and unleashing a torrent
Notwithstanding the dramatic activation of caspases upon
efflux of cytochrome c into the cytosol, MOMP itself heralds
the swift demise of the majority of cells in which this occurs.
This is because, in addition to cytochrome c, numerous mito-
chondrial intermembrane space proteins exit mitochondria
in ATP synthesis as well as the generation of reactive oxygen, all
of which swiftly compromise numerous cellular functions irre-
spective of the activation of caspases downstream (Green and
Kroemer, 2004; Kroemer and Martin, 2005). Because of the
essentially irreversible nature of MOMP, this is a key checkpoint
in apoptosis and is heavily policed by a complex web of proteins
that belong to the Bcl-2 family (Figure 1). Antiapoptotic Bcl-2
(Figure 1), thereby preventing assembly of the BAX/BAK channel
and consequently blocking MOMP and cytochrome c release
(Taylor et al., 2008; Youle and Strasser, 2008). Consequently, it
is the opening of the BAX/BAK channel, rather than caspase
activation, which represents the point ofno return for a celldeath
stimulus that engages the intrinsic pathway.
Caspase Activity Dictates the Switch between
Apoptosis and Necrosis
Although caspases are often thought of as the direct effectors of
cell death during apoptosis, this view is inconsistent with
numerous observations where caspase activity has been phar-
macologically inhibited or blocked through genetic inactivation.
In mammals, caspase inhibition does not prevent cell death in
response to stimuli that engage the intrinsic pathway to
apoptosis (Marsden et al., 2002; Ekert et al., 2004; Chipuk and
Green, 2005; Kroemer and Martin, 2005). Inhibition of caspase
activation downstream of MOMP merely delays, but does not
block, cell death. This explains why prosurvival members of
activation to block apoptosis (Figure 1).
While caspase activation is dispensible for cell death once
MOMP has occurred, there is a crucial difference in the outcome
if apoptosis-associated caspase activity is inhibited. The failure
to activate caspases dramatically changes the phenotype of
cell death, converting it from an apoptotic to a necrotic one (re-
viewed in Chipuk and Green, 2005; Kroemer and Martin, 2005).
This difference in outcome has very significant implications for
how the immune system responds to such cells, as we shall
Because MOMP is sufficient to ensure cell death for the
majority of cells, this leads us to ask what purpose caspase
activation serves during apoptosis? We suggest that the primary
role of caspase activation is the sequestration and inactivation
of cellular constituents, which—if permitted to leak out of the
cell—could activate the immune system and promote potentially
of this view, there is much evidence that phagocytes recognize
and respond to necrotic and apoptotic cells in fundamentally
different ways, even when apoptotic cells have entered
secondary necrosis and are leaking their cellular contents (Birge
and Ucker, 2008).
Sensing Danger: The Immune System Can Be Activated
by Cell Death
The ability of the sentinel cells of the immune system (e.g.,
from dead cells makes a great deal of sense in biological terms
(Matzinger, 1994). Our immune systems have evolved to protect
us from infectious agents by attacking and killing these upon
entry into the body. The simplest way to do this is to equip
sentinel cells of the innate immune system with a battery of
receptors that can detect molecules that are unique to foreign
organisms. Indeed, this is an important feature of innate immu-
nity, and macrophages, DCs, and other cells of the innate
immune system bristle with an array of Toll-like receptors
(TLRs) that are capable of detecting a wide variety of path-
ogen-associated molecular patterns (PAMPs). The conse-
quences of a PAMP binding to its corresponding receptor are
swift and lead to aggressive macrophage or DC activation fol-
lowed by the triggering of immune functions that are directed
atkilling theinfectious agent andpresenting associated antigens
to cells of the adaptive immune system (Iwasaki and Medzhitov,
Because it is not practical to have an endless array of TLRs
capable of recognizing all possible PAMPs, infectious agents
the immune system has also evolved its own array of endoge-
nous danger-associated molecular patterns (DAMPs) that are
released in response to sterile injury or infection that is associ-
ated with necrotic cell death (Matzinger, 1994; Kono and Rock,
2008). DAMPs can thus be viewed as surrogate markers for
infection that enable a host to mount an effective immune
response even in the absence of direct detection of a PAMP
(Figure 2). DAMPs bind to receptors, such as members of the
IL-1 receptor family, which have very similar intracellular
signaling domains as TLRs and instigate responses from macro-
phages or DCs almost identical to TLR engagement. Indeed,
some DAMPs have been reported to directly bind to TLRs,
although this is somewhat controversial (reviewed in Kono and
Rock, 2008). Thus, cells of the innate immune system can be
activated through encounter with either PAMPs or DAMPs,
with similar consequences. But how does infection or injury
trigger DAMP release?
DAMPs are normally sequestered within healthy cells and only
become released upon rupture of cells (i.e., necrosis), where-
upon such molecules spill out into the extracellular space and
trigger immune activation (Matzinger, 1994; Kono and Rock,
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
2008). Therefore, if an infectious agent provokes necrosis, this
will trigger the release of DAMPs, which will then activate the
immune system. It is highly appropriate that the presence of
necrotic cells can instigate immune responses, as cell rupture
is typically only caused by severe departures from normal
physiology. Thus, necrotic cell death can betray the activities
of viral infection (many viruses cause cell lysis during production
of new virions), the activities of bacterial toxins and other
due to compression injuries or burns, for example—is likely to
lead to infection. Thus, our immune systems are ‘‘hardwired’’
to become activated in response to the detection of PAMPs or
Whereas necrotic cells typically provoke inflammation,
apoptotic cells generally do not (Voll et al., 1997; Fadok et al.,
1998; Lucas et al., 2003). Furthermore, several laboratories
have independently reported that apoptotic cells are also
capable of profoundly attenuating responses to PAMPs deliv-
ered in parallel (Voll et al., 1997; Lucas et al., 2003; Serhan and
Savill, 2005). Indeed, apoptotic cells that have entered
secondary necrosis and are leaking their cellular contents also
retain this anti-inflammatory state, in contrast with cells that
have entered necrosis directly (Cocco and Ucker, 2001; Birge
and Ucker, 2008). This is curious and suggests that profound
alterations to cellular composition occur during apoptosis to
quell the activity of DAMPs within such that, even if these are
inadvertently released, their proinflammatory activity is blunted.
So, how do DAMPs become inactivated during apoptosis?
Because caspase activation is a fundamental difference
between apoptosis and necrosis, caspases are prime suspects
asthemajor effectorsoftheconversion ofcellsfromaproinflam-
matory to a noninflammatory or actively anti-inflammatory state.
We suggest that a major role of apoptosis-associated caspase
activity is to quell the proinflammatory properties of apoptotic
cells through inactivation (both directly and indirectly) and
sequestrationofpotentially proinflammatory molecules(DAMPs)
residing within (Figure 3). Before we examine the evidence in
support of this idea, we will first consider other apoptosis-asso-
ciated events that may limit the exposure of DAMPs during this
Apoptosis: A Calming Influence on the Immune System
Billions of cells die naturally on a daily basis as a consequence of
homeostatic tissue turnover and the vast majority of these cell
deaths occur via apoptosis (Green, 2010). This poses a signifi-
cant challenge for the immune system in terms of discriminating
between natural or programmed, as opposed to nonprog-
rammed or necrotic, cell deaths. Because of the capacity of
intracellular DAMPs to provoke immune responses (Figure 2), it
seems obvious that cells of the immune system need to be
able to discriminate between cell deaths that are programmed
Figure 3. Apoptotic and Secondarily Necrotic Cells Are
Caspase activation inactivates endogenous DAMPs and induces the
production of ‘‘find-me’’ and ‘‘eat-me’’ signals to coordinate the swift removal
of apoptotic cells by phagocytes and simultaneously suppress immune
responses. Unlike cells that haveentered necrosis directly, apoptotic cells that
have entered secondary necrosis carry inactivated DAMPs and are also anti-
inflammatory. In contrast, necrotic cells exhibit potent proinflammatory
properties due to release of active DAMPs.
Figure 2. Necrotic Cell Death Promotes Inflammation
The release of endogenous danger-associated molecular patterns (DAMPs)
from necrotic cells can activate various immune cell subsets. DAMPs activate
macrophages to secrete inflammatory cytokines that have broad stimulatory
effects on the immune system. Activation of DCs by DAMPs promotes DC
maturation and subsequent migration to lymph nodes where mature DCs can
initiate adaptive immunity by presenting antigens to antigen-specific T cells.
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
from those that are not. Because of the scale of ongoing
apoptosis in vivo, a failure to discriminate effectively between
apoptotic and necrotic cells could result in inappropriate
Because of the role of DAMPs as instigators of immune
responses, a strong case can be made for the idea that when
cells die by apoptosis, an important overarching objective is to
prevent the release of cellular constituents that could provoke
unnecessary immune responses. The simplest way of achieving
this objective is to ensure that cells undergoing apoptosis are
swiftly recognized and removed from a tissue before such cells
have an opportunity to leak their contents. Indeed, this is prob-
ably what happens to the majority of apoptotic cells and this is
ensured through caspase-dependent alterations to the plasma
membrane (called ‘‘eat-me’’ signals) that trigger recognition of
such cells by resident tissue macrophages, as well as nonpro-
fessional phagocytes (Savill and Fadok, 2000). The molecules
that have been implicated as triggers for phagocyte recognition
and removal of apoptotic cells include phosphatidylserine (PS),
which becomes externalized on the surface of apoptotic cells
in a caspase-dependent manner (Martin et al., 1995). However,
it is clear that PS exposure alone is insufficient to trigger uptake
of apoptotic cells by phagocytes and that additional molecules
are likely to appear in association with PS to qualify a cell as
apoptotic (Segawa et al., 2011). Although these other ligands
have yet to be identified, it is clear that caspase-dependent
exposure of ‘‘eat-me’’ signals undoubtedly plays a key role in
helping to ensure that apoptotic cells, in most cases, do not
linger in tissues for long enough to permit escape of DAMPs
into the extracellular space (Savill and Fadok, 2000). Moreover,
it is becoming increasingly clear that this clearance process is
not left up to chance encounters with neighboring phagocytes,
as recent studies suggest that factors which actively elicit the
attentions of phagocytes are released from apoptotic cells,
once again in a caspase-dependent manner (Gregory and
Pound, 2011; Ravichandran, 2011). Such factors have been
dubbed ‘‘find-me’’ signals.
‘‘Find-Me’’ Signals: Caspases Elicit Macrophage
Recruitment to Apoptotic Cells
While nonprofessional phagocytes may be capable of engulfing
neighboring cells that have undergone apoptosis, there is much
evidence that this is also carried out by resident tissue macro-
phages and DCs. Because these cells typically represent a rela-
tively small fraction of the cellular composition of most tissues,
this poses the question of how apoptotic cells are discovered
before they undergo secondary necrosis and awaken the full
force of the immune system. Evidence is now accumulating to
suggest that cells undergoing apoptosis signal their impending
demise through the release of one or more soluble factors that
act as chemoattractants for phagocytes (Gregory and Pound,
2011; Ravichandran, 2011). As mentioned above, such factors
have been dubbed ‘‘find-me’’ signals and there is increasing
evidence that apoptotic cells use such factors to guide phago-
cytes to their location to ensure a swift burial (Figure 4A).
Caspase-dependent release of lysophatidylcholine (LPC) was
one of the first molecules to be implicated as a ‘‘find-me’’ signal
that is released during apoptosis (Lauber et al., 2003). LPC
production by apoptotic cells appears to be instigated as a
consequence of caspase-3-dependent cleavage of calcium-
independent phospholipase A2, resulting in the hydrolysis of
membrane phosphatidylcholine to produce LPC (Lauber et al.,
2003). LPC can act as a chemoattractant for monocytic cells
and macrophages and may help to guide phagocytes to the
dying cell. However, evidence that LPC is an important ‘‘find
polypeptide II (EMAPII) is another molecule that has been impli-
cated as a chemotactic factor released from apoptotic cells
(Knies et al., 1998). EMAPII undergoes caspase-dependent
proteolysis during apoptosis and the C-terminal fragment of
this molecule can act as a trigger for monocyte attraction (Knies
et al., 1998).
Recent studies from Ravichandran and colleagues have also
implicated efflux of ATP and UTP as ‘‘find-me’’ signals for
apoptotic cells (Elliott etal., 2009; Chekeni et al.,2010). Although
ATP efflux is also associated with necrosis, the magnitude of
release during apoptosis (?2% of the total cellular pool of this
nucleotide) appears to be much lower than that seen during
necrosis (Elliott et al., 2009). ATP/UTP efflux during apoptosis
has been shown to be caspase-dependent as a consequence
of proteolysis of the membrane channel pannexin-1 (Chekeni
Figure 4. Apoptotic Cells Release ‘‘Find-Me’’ Signals to Attract
phosphate (S1P), and lysophosphatidylcholine (LPC) that attract the attention
of professional phagocytes.
(B) Apoptotic cells generate ‘‘find-me,’’ ‘‘eat-me,’’ and ‘‘tolerate-me’’ signals
that actively downregulate immune responses and qualify this type of cell
death as harmless.
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
et al., 2010). ATP efflux from apoptotic cells promotes selective
recruitment of monocytes, but apparently not neutrophils, which
could be due to two factors. On the one hand, low concentra-
tions of ATP have been suggested to be anti-inflammatory,
and data from Gregory and colleagues suggest that apoptotic
cells may also release anti-inflammatory factors, such as lacto-
ferrin, to suppress neutrophil chemotaxis and possibly their
activation (Bournazou et al., 2009). Thus, although ATP release
has been identified as a trigger of strong immune reactions, it
is possible that release of more modest amounts of this nucleo-
tide may have the opposite outcome during apoptosis.
Another possible ‘‘find-me’’ signal that been reported to be
released in a caspase-dependent fashion during apoptosis is
the sphingolipid, sphingosine-1-phosphate (S1P), which is
generated through hydrolysis of ceramide (Gude et al., 2008).
Two S1P kinases (SphK1 and SphK2) have been implicated in
the generation of S1P, one of which (SphK2) undergoes
caspase-dependent cleavage during apoptosis (Weigert et al.,
2010). Proteolysis of SphK2 has been suggested to facilitate its
release into the extracellular space where it can generate S1P
and trigger chemotaxis of phagocytes toward apoptotic cells.
However, an important caveat is that the concentrations of
S1P released by apoptotic cells appear to be significantly lower
than those required to trigger robust macrophage chemotaxis
(Gude et al., 2008). Therefore the relative importance of S1P as
a ‘‘find-me’’ signal for apoptotic cells awaits further clarification.
Fractalkine, a germinal center B cell associated chemokine,
has also been proposed to serve as a ‘‘find-me’’ signal and is
released from cells in a caspase-dependent fashion, although
this chemokine is probably not cleaved directly by caspases
(Truman et al., 2008). Furthermore, because fractalkine appears
to be expressed predominantly by B cells, it is unlikely to repre-
sent ageneral ‘‘find-me’’ signal.This raises thesomewhat unpal-
atable prospect that particular cell types may utilize specific
‘‘find-me’’ signals that are not shared by other cell types. What
is perhaps more likely is that there are ‘‘find-me’’ signals
common to all cells, as well as additional signals that may be
used by particular tissues.
As the preceding discussion illustrates, there is evidence
to argue that caspases are instrumental in the generation of
both ‘‘find-me’’ and ‘‘eat-me’’ signals that coordinate the
swift discovery, recognition and removal of apoptotic cells
‘‘Tolerate-Me’’ Signals: Apoptotic Cells Are Actively
Rapid clearance of apoptotic cells minimizes the probability that
such cells will persist in tissues for long enough to undergo
secondary necrosis and release DAMPs to awaken the immune
system. Notwithstanding this, some studies indicate that
apoptotic cells undergo a transformation that actively discour-
ages inflammatory responses upon encountering phagocytes.
This suggests that apoptotic cells are not merely passively
noninflammatory, but are capable of actively inducing an anti-
inflammatory state in macrophages and DCs that encounter
such cells (Voll et al., 1997; Fadok et al., 1998; Stuart et al.,
2002). How this is achieved is still a matter of debate, but a
consistent observation is that direct contact with apoptotic cells
as well as other anti-inflammatory mediators by the ingesting
phagocyte (Voll et al., 1997; Fadok et al., 1998; Lucas et al.,
2006). Ucker and colleagues have also reported that direct
cell-cell contact between the phagocyte and the apoptotic cell
is sufficient to confer this anti-inflammatory state, without
recourse to soluble factors (reviewed in Birge and Ucker,
2008). Moreover, late stage apoptotic cells that have begun to
leak their contents, and even cellular fractions derived from
apoptotic cells, also retain their anti-inflammatory properties
(Birge and Ucker, 2008), once again suggesting that alterations
to their composition have occurred that have inactivated endog-
enous DAMPs within.
Collectively, these studies suggest that factors associated
with, or released from, apoptotic cells may actively reprogram
macrophages to an anti-inflammatory or ‘‘wound healing’’
phenotype. Alterations to the composition of apoptotic cells
that renders their contents anti-inflammatory during apoptosis
could thus be viewed as ‘‘tolerate-me’’ signals (Figure 4B). The
induction of an anti-inflammatory program within a DC that has
recently ingested an apoptotic cell (and is therefore loaded
with self-antigens) would prevent simultaneous encounters
with PAMPs from activating the DC and instigating an immune
response against self-antigens. Therefore, the anti-inflammatory
properties of apoptotic cells may be crucial for maintaining self-
tolerance (Figure 3 and Figure 4B). In addition, the generation of
‘‘tolerate-me’’signalswould alsoactasa safeguard againstsitu-
ations where apoptotic cells may not always be removed prior to
cell leakage. However, the cell-associated factors that confer
anti-inflammatory properties upon apoptotic cells have yet to
be convincingly identified.
In addition to the anti-inflammatory properties of apoptotic
may also directly or indirectly inactivate DAMPs within apoptotic
cells. It is also possible that inactive forms of DAMPs could exert
anti-inflammatory roles as discussed above. Some of these
DAMP-inactivating mechanisms will now be considered.
Switching Off the Alarm System: Caspases Coordinate
the Inactivation of DAMPs
The full spectrum of DAMPs released by necrotic cells that are
capable of engaging the innate immune system still await clarifi-
cation, but current evidence suggests that these include
cytokines such as IL-1a and IL-33, as well as other nonclassical
cytokines and immunostimulatory molecules, such as HMGB1,
uric acid, ATP, certain heat-shock proteins, single-stranded
RNA, and genomic DNA (reviewed in Kono and Rock, 2008).
Additional DAMPs almost certainly await discovery, as this is
an active area of investigation at present. DAMPs all share in
common the property that they represent ‘‘hidden self,’’ as these
molecules are not normally found in the extracellular space, thus
their presence in this compartment is indicative of a severe
departure from normality (Matzinger, 1994; Kono and Rock,
2008). Apart from limiting damage to neighboring cells by avoid-
ing release of cellular contents, one of the major benefits of
apoptosis may be to prevent the unmasking of ‘‘hidden self’’ or
the conversion of DAMPs into harmless forms, thereby prevent-
ing unwanted immune responses (Figure 3 and Figure 5).
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
Recent studies suggest that genomic DNA is an important
DAMP, capable of initiating DC maturation and the initiation of
immune responses to coadministered antigens (Marichal et al.,
2011; Ishii et al., 2001). It has long been known that aluminum
salts (Alum) can provoke strong immune reactions to proteins,
or protein fragments, that on their own elicit little or no immunity.
Indeed, this property is widely exploited in the process of vacci-
nation, to boost immune responses to molecules that would
otherwise be ignored by the immune system. However, the
molecular basis for the ability of Alum to trigger immune reac-
tions to coadministered antigens has been debated. Recent
studies suggest that Alum simply triggers necrosis of cells at
the site of administration, leading to release of genomic DNA
that acts as a DAMP to activate the immune system (Marichal
et al., 2011). Indeed, administration of genomic DNA alone is
sufficient to replicate many of the immunostimulatory effects of
Alum (Marichal et al., 2011). Strikingly, hydrolysis of DNA with
endonucleases strongly attenuated its immune activating prop-
erties, as well as those of Alum (Marichal et al., 2011). This
may have particular relevance for apoptosis, as a curious feature
of this mode of cell death is that genomic DNA undergoes exten-
sive hydrolysis to small ?200 base pair fragments due to the
actions of a caspase-activated DNAase (CAD/DFF) that
becomes activated during this mode of cell death (Enari et al.,
1998). Indeed, this was one of the earliest molecular character-
istics of apoptotic cells to be reported but necrotic cells do not
manifest any such chromatin hydrolysis (Kroemer and Martin,
2005). It has long been puzzling why extensive DNA fragmenta-
tion occurs during apoptosis. However, in light of the recent
discoveries suggesting that short DNA fragments are much
less effective immune activators than their high-molecular-
weight equivalents, it seems plausible to suggest that caspases
instigate DNA hydrolysis during apoptosis to dampen the
immune activating properties of genomic DNA. Direct evidence
in support of this proposal awaits further investigation; however,
it is noteworthy that CAD?/?mice develop an autoinflammatory
condition on a DNAaseII?/?background (Kawane et al., 2003).
This appears to be due to incompletely digested DNA within
apoptotic cells (due to CAD deficiency), persisting in macro-
phages (due to DNAase II deficiency) and leading to the activa-
tion of a pathway for sensing cytoplasmic DNA fragments, the
RIG-I/IRF-3 pathway (Okabe et al., 2009).
Additional evidence in support of the idea that caspases asso-
ciated with apoptosis suppress responses to endogenous
DAMPs, which again may take the form of cytosolic genomic
DNA, comes from a series of intriguing studies by Wallach and
colleagues (Kovalenko et al., 2009; Rajput et al., 2011). These
studies demonstrate that caspase-8 plays an important role in
suppressing activation of the RIG-I/IRF-3 pathway, which is
invoked in response to cytoplasmic DNA and RNA. During
cornification of the skin, the nuclei of terminally differentiating
keratinocytes break down and this might be sufficient to trigger
activation of the RIG-I/IRG-3 pathway. However, recruitment of
of RIPK1, a key signaling component of this complex, thereby
attenuating expression of IRF-3-inducible genes which include
the interferons and other inflammatory factors (Rajput et al.,
2011). Conditional deletion of CASP-8 in the skin leads to a
spontaneous inflammatory disease due to excessive RIG-I-
dependent IRF-3 activation, quite possibly in response to
endogenous cytoplasmic DNA that is produced during keratino-
cye cornification (Kovalenko et al., 2009).
IL-33, a recently discovered cytokine and member of the IL-1
family, appears to be a bona fide alarmin and is released as a
full-length protein during necrosis, but undergoes caspase-
dependent proteolysis during apoptosis (Lu ¨thi et al., 2009;
Cayrol and Girard, 2009). Although an initial report suggested
that IL-33 was activated through caspase-1-dependent proteol-
ysis, several laboratories have comprehensively demonstrated
that this cytokine is inactivated by caspases-3/-7-dependent
proteolysis during apoptosis (Lu ¨thi et al., 2009; Cayrol and
HMGB1 is another molecule that has been repeatedly
implicated as a DAMP and represents another proinflammatory
molecule whose activity, as well as availability for release, is
differentially regulated in a caspase-dependent manner. This
chromatin-binding molecule is loosely bound to DNA in healthy
cells and is readily released during necrosis (Scaffidi et al.,
2002). However, upon condensation of chromatin during apo-
ptosis—a process that is caspase-dependent through protelysis
of the Mst1 kinase (reviewed in Taylor et al., 2008)—HMGB1
becomes much more tightly associated to chromatin and its
release is dramatically attenuated (Scaffidi et al., 2002). Thus,
caspase-dependent chromatin condensation decreases the
mobility of HMGB1 during apoptosis, thereby minimizing the
potential for escape into the extracellular space to awaken
the innate immune system. Furthermore, caspases have also
Figure 5. Caspases Coordinate the Inactivation of DAMPs during
Caspase activity, downstream of MOMP, directly and indirectly leads to
the inactivation and/or sequestration of multiple endogenous DAMPs such as
IL-33, genomic DNA, ATP, and HMGB1.
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
been implicated in the direct inactivation of HMGB1 through
proteoysis of the mitochondrial caspase substrate p75NDUF
(Kazama et al., 2008). The later event triggers a burst of reactive
oxygen, leading to oxidation of a critical cysteine residue on
HMGB1 that abolishes its DAMP activity (Kazama et al., 2008).
Studies also suggest that extracellular ATP has DAMP activity
at high concentrations (Idzko et al., 2007). However, ATP
production during apoptosis is sharply compromised in a cas-
pase-dependent manner through proteolysis of the mitochon-
drial electron transport component p75NDUF (Ricci et al.,
2004). Cellular ATP concentrations decline precipituously during
apoptosis and this can beattenuated through blocking caspase-
p75NDUF (Ricci et al., 2004). Thus, caspases may directly
contribute to reducing the availability of this nucleotide for
release if cell clearance is delayed.
Thus there is mounting evidence that caspases directly and
indirectly coordinate the inactivation of several important
DAMPs during apoptosis (Figure 5), although much work
remains to be done to explore how other DAMPs may be
affected during this process.
Challenges to the Model
Several observations, at first sight, are difficult to reconcile with
the idea that apoptosis-associated caspase activation plays an
anti-inflammatory role in mammals, rather than a strictly death-
inducing one. Although the majority of Apaf-1-null mice die at
birth, a small percentage of these mice do survive to adulthood
on the C57BL/6 background without any apparent spontaneous
inflammatory phenotype, or indeed any obvious tissue abnor-
malities (Honarpour et al., 2000). Note that these observations
are problematic both for the idea that caspases are required
for cell death, as well as for the idea that cell death-associated
caspase activation suppresses inflammation. However, one
reason for the lack of any obvious inflammatory phenotype in
Apaf-1-null animals may be that, in the absence of Apaf-1, cas-
pase activation may still occur via an alternative pathway to
effector caspase activation, which remains to be defined.
Indeed, Nagata and colleagues have recently shown that death
induced by staurosporine treatment, a stimulus that would nor-
mally promote caspase activation via the Apaf-1/caspase-9
pathway, is associated with caspase-3 activation in APAF1?/?
cells (Nagasaka et al., 2010), suggesting the existence of a
compensatory route to caspase-3 activation in these animals.
Another possibility is that inflammation due to deficiencies in
effector caspase activation only occurs in situations where large
numbers of cells undergo apoptosis en masse—during infection
or due to pathological injury involving death of large numbers of
cells for example. Thus, stochastic rates of cell death under
quantities of DAMPs to overwhelm natural anti-inflammatory
defenses, with caspase-dependent anti-inflammatory mecha-
nisms becoming critical only when large numbers of cells die
simultaneously. Therefore, the failure of Apaf-1-null animals to
manifest spontaneous inflammation may relate to the threshold
of DAMP release that is required to initiate inflammation. To
address this possibility, it will be interesting to explore whether
Apaf-1-null animals exhibit inflammatory phenotypes upon
challenge with cytotoxic drugs or pathogens thatprovoke robust
amounts of cell death within a restricted time window.
The recently described phenomenon of ‘‘immunogenic cell
death’’ also appears to be at odds with the view that apoptosis
is an immunologically silent mode of cell death (Obeid et al.,
2007). Kroemer and colleagues have reported that certain drugs
with the potential to induce apoptosis, predominantly anthracy-
clins, promote immune reactions in vivo that can result in the
efficient clearance of tumors (Obeid et al., 2007). The latter
observation challenges the idea that a key outcome of apoptotic
cell death is the avoidance of immune activation. However, it is
important to note that proapoptotic drugs capable of triggering
immunogenic cell death were found to be in the minority when
compared with numerous other proapoptotic stimuli in this
model (Obeid et al., 2007). Second, it is possible that certain
proapoptotic stimuli may override ananti-inflammatory outcome
because these stimuli are inherently proinflammatory. For
example, TNF is a major proinflammatory stimulus that is also
capable of inducing apoptosis. Thus, cells dying in response to
TNF treatment may also simultaneously produce and secrete
proinflammatory cytokines, thereby negating the anti-inflamma-
tory effects of ‘‘apoptotic’’ caspase activation. A similar argu-
ment can be applied to Fas/CD95 and TRAIL, which are also
capable of triggering the production of proinflammatory cyto-
kines, as well as apoptosis (Leverkus et al., 2003; Farley et al.,
2006; Altemeier et al., 2007). Thus, it is plausible that triggers
of immunogenic cell death may also promote the production of
proinflammatory cytokines concurrently with apoptosis. Third,
although anthracyclins may induce apoptosis in vitro, it is also
possible that these agents induce a significant amount of
mation that acts as a driver for tumor clearance. Therefore, the
precise nature of the proapoptotic stimulus, in terms of its ability
to instigate theproduction of proinflammatory cytokines, may be
logically silent or not.
Caspase Activation Also Plays an Anti-inflammatory
Role within the Extrinsic Pathway
Much of the preceding discussion has focused upon caspase
activation downstream of MOMP within the intrinsic pathway.
However, a second major route to caspase activation and
apoptosis involves members of the TNF receptor family, which
includes TNF itself, Fas/CD95, and TRAIL, among others (Wal-
lach et al., 1999). In this pathway, engagement of the latter
receptors with their cognate ligands can recruit caspase-8 and
caspase-10 to the cytoplasmic tails of these receptors, via
adaptor proteins, and result in the activation of downstream
effector caspases. Propagation of caspase activation within
the extrinsic pathway occurs either through proteolysis and
activation of the BH3-only protein, Bid, which leads to MOMP
and/or through direct proteolytic processing and activation of
downstream effector caspase-3 and caspase-7 (reviewed in
Taylor et al., 2008). Unlike within the intrinsic pathway, neutrali-
zation of caspase activation within the extrinsic pathway does
block apoptosis, which appears to contradict the idea that
caspases play an anti-inflammatory role. However, what is
frequently overlooked is that TNF, Fas, and TRAIL are potent
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
proinflammatory molecules that are capable of triggering
cytokine and chemokine production from diverse cell types,
through recruitment of RIPK1 as well as other signaling kinases
to their receptor complexes (Leverkus et al., 2003; Farley et al.,
2006; Altemeier et al., 2007). Inhibition of caspase activation in
thelatter contexts can enhance deathreceptor-induced produc-
tion of proinflammatory cytokines through blocking cell death
(Farley et al., 2006; Altemeier et al., 2007), which argues that
caspase activation also plays an anti-inflammatory role in these
Furthermore, similar to the role of caspase-8 as a negative
regulator of RIG-I/IRF-3-dependent inflammation, as discussed
earlier, a raft of recent papers have implicated caspase-8, as
well as its regulators FADD and FLIP, as negative regulators of
TNF-driven necrosis and inflammation (reviewed by Green
et al., 2011). Collectively, these studies suggest that caspase-8
plays a role as an inhibitor of excessive RIPK1 and RIPK3 activa-
RIPK1 deubiquitinating enzyme, CYLD (Green et al., 2011). Nor-
mally, caspase-8-mediated proteolysis of CYLD and/or RIPK1
restrains the activity of this kinase, keeping its activity within
a desirable range. However, in the absence of caspase-8 (or
its regulators FADD/FLIP), RIPK1 becomes deubiquitinated,
due to CYLD stabilization, which leads to excessive RIPK1-
driven RIPK3 activation. Deregulated RIPK3 activation promotes
necrosis, also called necroptosis in this context, which is associ-
ated with considerable inflammation. This is another good
example of caspase-dependent inhibition of molecules, RIPK1
and RIPK3, which have the potential to promote a mode of cell
death that leads to robust immune activation. This pathway
may act as a failsafe for the detection of viruses and other infec-
tious agents that may inhibit caspase activity.
Evolutionary Conservation of an Immune-Regulating
Role for Caspases
We have confined the preceding discussion to mammalian
caspases, because it is in these organisms where MOMP rather
than caspase activation appears to be the commitment point for
cell death within the intrinsic pathway, although this may also
apply to other vertebrates. Thus, the anti-inflammatory role of
cell death-associated caspases proposed here may be a rela-
tively recent adaptation, with caspases playing a predominantly
cell killing-related role in more primitive organisms, such as
nematodes. This could be due to the lower risk of autoimmune
reactionsinlesscomplexand short-livedmulticellular organisms
that lack adaptive immune systems. However, it is worth noting
that cell death-related caspases provide protection against
bacterial and viral infections in the worm (Aballay and Ausubel,
2001; Liu et al., 2006) and the fly (Leulier et al., 2000). Therefore,
the role of caspases as positive and negative regulators of
immune reactions could well have an ancient origin.
As the preceding discussion illustrates, there is now much
evidence to argue that a major function of caspase activation
during apoptosis is to ensure swift recognition and engulfment
of dying cells by phagocytes to prevent cell rupture and avoid
release of proinflammatory DAMPs that could activate the
immune system. Moreover, caspases are also capable of
actively disabling DAMPs, such as IL-33, HMGB1, and genomic
tory potential of apoptotic cells. Viewed in this light, caspases
activated during apoptosis can be interpreted to function
primarily as anti-inflammatory enzymes, rather than cell killing
enzymes, serving to avoid the potentially proinflammatory con-
sequences of cell death. This leads to a new perspective con-
cerning the role of the mammalian caspase family, with certain
members of this family (the caspase-1-related branch) acting
as proinflammatory proteases, and others (the caspase-9/
caspase-3-related branch) acting as anti-inflammatory prote-
ases. Thus, the majority of members of the mammalian caspase
family can be construed to act as either positive or negative
regulators of inflammation. This interpretation is a significant
refinement of the existing model, which regards some caspases
as cell death effectors and others as regulators of inflammation.
Work in the Martin laboratory is supported by SRC (07/SRC/B1144) and PI
(08/IN.1/B2031) grants from Science Foundation Ireland. S.J.M. is a Science
Foundation Ireland Principal Investigator. We apologize for failing to cite
many primary papers due to space constraints.
Aballay, A., and Ausubel, F.M. (2001). Programmed cell death mediated by
ced-3 and ced-4 protects Caenorhabditis elegans from Salmonella typhimu-
rium-mediated killing. Proc. Natl. Acad. Sci. USA 98, 2735–2739.
Fas (CD95) induces macrophage proinflammatory chemokine production via
a MyD88-dependent, caspase-independent pathway. J. Leukoc. Biol. 82,
Birge, R.B., and Ucker, D.S. (2008). Innate apoptotic immunity: the calming
touch of death. Cell Death Differ. 15, 1096–1102.
Bournazou, I., Pound, J.D., Duffin, R., Bournazos, S., Melville, L.A., Brown,
S.B., Rossi, A.G., and Gregory, C.D. (2009). Apoptotic human cells inhibit
migration of granulocytes via release of lactoferrin. J. Clin. Invest. 119, 20–32.
Cayrol, C., and Girard, J.P. (2009). The IL-1-like cytokine IL-33 is inactivated
after maturation by caspase-1. Proc. Natl. Acad. Sci. USA 106, 9021–9026.
Chekeni, F.B., Elliott, M.R., Sandilos, J.K., Walk,S.F., Kinchen,J.M.,Lazarow-
ski, E.R., Armstrong, A.J., Penuela, S., Laird, D.W., Salvesen, G.S., et al.
(2010). Pannexin 1 channels mediate ‘find-me’ signal release and membrane
permeability during apoptosis. Nature 467, 863–867.
Chipuk, J.E., and Green, D.R. (2005). Do inducers of apoptosis trigger
caspase-independent cell death? Nat. Rev. Mol. Cell Biol. 6, 268–275.
Cocco, R.E., and Ucker, D.S. (2001). Distinct modes of macrophage recogni-
tion for apoptotic and necrotic cells are not specified exclusively by phospha-
tidylserine exposure. Mol. Biol. Cell 12, 919–930.
Creagh, E.M., Conroy, H., and Martin, S.J. (2003). Caspase-activation
pathways in apoptosis and immunity. Immunol. Rev. 193, 10–21.
Ekert, P.G., Read, S.H., Silke, J., Marsden, V.S., Kaufmann, H., Hawkins, C.J.,
Gerl, R., Kumar, S., and Vaux, D.L. (2004). Apaf-1 and caspase-9 accelerate
apoptosis, but do not determine whether factor-deprived or drug-treated cells
die. J. Cell Biol. 165, 835–842.
Elliott, M.R., Chekeni, F.B., Trampont, P.C., Lazarowski, E.R., Kadl, A., Walk,
S.F., Park, D., Woodson, R.I., Ostankovich, M., Sharma, P., et al. (2009).
Nucleotides released by apoptotic cells act as a find-me signal to promote
phagocytic clearance. Nature 461, 282–286.
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
Enari, M., Sakahira, H., Yokoyama, H., Okawa, K., Iwamatsu, A., and Nagata,
S. (1998). A caspase-activated DNase that degrades DNA during apoptosis,
and its inhibitor ICAD. Nature 391, 43–50.
Fadok, V.A., Bratton, D.L., Konowal, A., Freed, P.W., Westcott, J.Y., and
Henson, P.M. (1998). Macrophages that have ingested apoptotic cells
in vitro inhibit proinflammatory cytokine production through autocrine/para-
crine mechanisms involving TGF-beta, PGE2, and PAF. J. Clin. Invest. 101,
Farley, S.M.,Dotson,A.D.,Purdy,D.E.,Sundholm,A.J.,Schneider, P.,Magun,
B.E., and Iordanov, M.S. (2006). Fas ligand elicits a caspase-independent
proinflammatory response in human keratinocytes: implications for dermatitis.
J. Invest. Dermatol. 126, 2438–2451.
Green, D.R. (2010). Means to and End: Apoptosis and Other Cell Death
Mechanisms (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press),
Green, D.R., and Kroemer, G. (2004). The pathophysiology of mitochondrial
cell death. Science 305, 626–629.
Green, D.R., Oberst, A., Dillon, C.P., Weinlich, R., and Salvesen, G.S. (2011).
RIPK-dependent necrosis and its regulation by caspases: a mystery in five
acts. Mol. Cell 44, 9–16.
Gregory, C.D., and Pound, J.D. (2011). Cell death in the neighbourhood: direct
microenvironmental effects of apoptosis in normal and neoplastic tissues.
J. Pathol. 223, 177–194.
Gude, D.R., Alvarez, S.E., Paugh, S.W., Mitra, P., Yu, J., Griffiths, R., Barbour,
S.E., Milstien, S., and Spiegel, S. (2008). Apoptosis induces expression of
sphingosine kinase 1 to release sphingosine-1-phosphate as a ‘‘come-and-
get-me’’ signal. FASEB J. 22, 2629–2638.
Honarpour, N., Du, C., Richardson, J.A., Hammer, R.E., Wang, X., and Herz, J.
(2000). Adult Apaf-1-deficient mice exhibit male infertility. Dev. Biol. 218,
Idzko, M., Hammad, H., van Nimwegen, M., Kool, M., Willart, M.A., Muskens,
F., Hoogsteden, H.C., Luttmann, W., Ferrari, D., Di Virgilio, F., et al. (2007).
Extracellular ATP triggers and maintains asthmatic airway inflammation by
activating dendritic cells. Nat. Med. 13, 913–919.
Ishii,K.J., Suzuki, K.,Coban, C.,Takeshita,F.,Itoh,Y.,Matoba, H.,Kohn, L.D.,
and Klinman, D.M. (2001). Genomic DNA released by dying cells induces the
maturation of APCs. J. Immunol. 167, 2602–2607.
Iwasaki, A., and Medzhitov, R. (2004). Toll-like receptor control of the adaptive
immune responses. Nat. Immunol. 5, 987–995.
Kawane, K., Fukuyama, H., Yoshida, H., Nagase, H., Ohsawa, Y., Uchiyama,
Y., Okada, K., Iida, T., and Nagata, S. (2003). Impaired thymic development
in mouse embryos deficient in apoptotic DNA degradation. Nat. Immunol. 4,
Kayagaki, N., Warming, S., Lamkanfi, M., Vande Walle, L., Louie, S., Dong, J.,
Newton, K., Qu, Y., Liu, J., Heldens, S., et al. (2011). Non-canonical inflamma-
some activation targets caspase-11. Nature 479, 117–121.
Kazama,H., Ricci,J.E.,Herndon,J.M., Hoppe,G.,Green, D.R., and Ferguson,
T.A. (2008). Induction of immunological tolerance by apoptotic cells requires
caspase-dependent oxidation of high-mobility group box-1 protein. Immunity
Knies, U.E., Behrensdorf, H.A., Mitchell, C.A., Deutsch, U., Risau, W., Drexler,
H.C., and Clauss, M. (1998). Regulation of endothelial monocyte-activating
polypeptide II release by apoptosis. Proc. Natl. Acad. Sci. USA 95, 12322–
Kono, H., and Rock, K.L. (2008). How dying cells alert the immune system to
danger. Nat. Rev. Immunol. 8, 279–289.
Kovalenko, A., Kim, J.C., Kang, T.B., Rajput, A., Bogdanov, K., Dittrich-
Breiholz, O., Kracht, M., Brenner, O., and Wallach, D. (2009). Caspase-8
deficiency in epidermal keratinocytes triggers an inflammatory skin disease.
J. Exp. Med. 206, 2161–2177.
Kroemer, G., and Martin, S.J. (2005). Caspase-independent cell death.
Nat. Med. 11, 725–730.
Lauber, K., Bohn, E., Kro ¨ber, S.M., Xiao, Y.J., Blumenthal, S.G., Lindemann,
R.K., Marini, P., Wiedig, C., Zobywalski, A., Baksh, S., et al. (2003). Apoptotic
cells inducemigrationof phagocytes viacaspase-3-mediated releaseof alipid
attraction signal. Cell 113, 717–730.
Leulier, F., Rodriguez, A., Khush, R.S., Abrams, J.M., and Lemaitre, B. (2000).
The Drosophila caspase Dredd is required to resist gram-negative bacterial
infection. EMBO Rep. 1, 353–358.
Leverkus, M., Sprick, M.R., Wachter, T., Denk, A., Bro ¨cker, E.B., Walczak, H.,
and Neumann, M. (2003). TRAIL-induced apoptosis and gene induction in
HaCaT keratinocytes: differential contribution of TRAIL receptors 1 and 2.
J. Invest. Dermatol. 121, 149–155.
Liu, W.H., Lin, Y.L., Wang, J.P., Liou, W., Hou, R.F., Wu, Y.C., and Liao, C.L.
(2006). Restriction of vaccinia virus replication by a ced-3 and ced-4-depen-
dent pathway in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 103,
Lucas, M., Stuart, L.M., Savill, J., and Lacy-Hulbert, A. (2003). Apoptotic cells
and innate immune stimuli combine to regulate macrophage cytokine secre-
tion. J. Immunol. 171, 2610–2615.
Lucas, M., Stuart, L.M., Zhang, A., Hodivala-Dilke, K., Febbraio, M.,
Silverstein, R., Savill, J., and Lacy-Hulbert, A. (2006). Requirements for
apoptotic cell contact in regulation of macrophage responses. J. Immunol.
Lu ¨thi, A.U., and Martin, S.J. (2007). The CASBAH: a searchable database of
caspase substrates. Cell Death Differ. 14, 641–650.
Lu ¨thi, A.U., Cullen, S.P., McNeela, E.A., Duriez, P.J., Afonina, I.S., Sheridan,
C., Brumatti, G., Taylor, R.C., Kersse, K., Vandenabeele, P., et al. (2009).
Suppression of interleukin-33 bioactivity through proteolysis by apoptotic
caspases. Immunity 31, 84–98.
Marichal, T., Ohata, K., Bedoret, D., Mesnil, C., Sabatel, C., Kobiyama, K.,
Lekeux, P., Coban, C., Akira, S., Ishii, K.J., et al. (2011). DNA released from
dyinghostcells mediates aluminum adjuvantactivity.Nat.Med. 17,996–1002.
Marsden, V.S., O’Connor, L., O’Reilly, L.A., Silke, J., Metcalf, D., Ekert, P.G.,
Huang, D.C., Cecconi, F., Kuida, K., Tomaselli, K.J., et al. (2002). Apoptosis
initiated by Bcl-2-regulated caspase activation independently of the cyto-
chrome c/Apaf-1/caspase-9 apoptosome. Nature 419, 634–637.
Martin, S.J., Reutelingsperger, C.P., McGahon, A.J., Rader, J.A., van Schie,
R.C., LaFace, D.M., and Green, D.R. (1995). Early redistribution of plasma
membrane phosphatidylserine is a general feature of apoptosis regardless
of the initiating stimulus: inhibition by overexpression of Bcl-2 and Abl.
J. Exp. Med. 182, 1545–1556.
Matzinger, P. (1994). Tolerance, danger, and the extended family. Annu. Rev.
Immunol. 12, 991–1045.
Nagasaka, A., Kawane, K., Yoshida, H., and Nagata, S. (2010). Apaf-1-inde-
pendent programmed cell death in mouse development. Cell Death Differ.
Obeid, M.,Tesniere,A.,Ghiringhelli,F., Fimia, G.M.,Apetoh,L., Perfettini, J.L.,
Castedo, M., Mignot, G., Panaretakis, T., Casares, N., et al. (2007). Calreticulin
exposure dictates the immunogenicity of cancer cell death. Nat. Med. 13,
Okabe, Y., Sano, T., and Nagata, S. (2009). Regulation of the innate immune
response by threonine-phosphatase of Eyes absent. Nature 460, 520–524.
Rajput, A.,Kovalenko, A.,Bogdanov, K., Yang, S.H., Kang, T.B., Kim, J.C., Du,
J., and Wallach, D. (2011). RIG-I RNA helicase activation of IRF3 transcription
factor is negatively regulated by caspase-8-mediated cleavage of the RIP1
protein. Immunity 34, 340–351.
Ravichandran, K.S. (2011). Beginnings of a good apoptotic meal: the find-me
and eat-me signaling pathways. Immunity 35, 445–455.
Ricci, J.E., Mun ˜oz-Pinedo, C., Fitzgerald, P., Bailly-Maitre, B., Perkins, G.A.,
Yadava, N., Scheffler, I.E., Ellisman, M.H., and Green, D.R. (2004). Disruption
thep75subunit of complex Iof theelectron transport chain.Cell 117,773–786.
Riedl, S.J., and Salvesen, G.S. (2007). The apoptosome: signalling platform of
cell death. Nat. Rev. Mol. Cell Biol. 8, 405–413.
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.
Savill, J., and Fadok, V. (2000). Corpse clearance defines the meaning of cell
death. Nature 407, 784–788.
Scaffidi, P., Misteli, T., and Bianchi, M.E. (2002). Release of chromatin protein
HMGB1 by necrotic cells triggers inflammation. Nature 418, 191–195.
Schroder, K., and Tschopp, J. (2010). The inflammasomes. Cell 140, 821–832.
Segawa, K., Suzuki, J., and Nagata, S. (2011). Constitutive exposure of phos-
phatidylserine on viable cells. Proc. Natl. Acad. Sci. USA 108, 19246–19251.
Serhan, C.N., and Savill, J. (2005). Resolution of inflammation: the beginning
programs the end. Nat. Immunol. 6, 1191–1197.
Stuart, L.M., Lucas, M., Simpson, C., Lamb, J., Savill, J., and Lacy-Hulbert, A.
(2002). Inhibitory effects of apoptotic cell ingestion upon endotoxin-driven
myeloid dendritic cell maturation. J. Immunol. 168, 1627–1635.
Taylor, R.C., Cullen, S.P., and Martin, S.J. (2008). Apoptosis: controlled
demolition at the cellular level. Nat. Rev. Mol. Cell Biol. 9, 231–241.
Truman, L.A., Ford, C.A., Pasikowska, M., Pound, J.D., Wilkinson, S.J.,
Dumitriu, I.E., Melville, L., Melrose, L.A., Ogden, C.A., Nibbs, R., et al.
(2008). CX3CL1/fractalkine is released from apoptotic lymphocytes to
stimulate macrophage chemotaxis. Blood 112, 5026–5036.
Voll, R.E., Herrmann, M., Roth, E.A., Stach, C., Kalden, J.R., and Girkontaite, I.
(1997). Immunosuppressive effects of apoptotic cells. Nature 390, 350–351.
Wallach, D., Varfolomeev, E.E., Malinin, N.L., Goltsev, Y.V., Kovalenko, A.V.,
and Boldin, M.P. (1999). Tumor necrosis factor receptor and Fas signaling
mechanisms. Annu. Rev. Immunol. 17, 331–367.
Weigert, A., Cremer, S., Schmidt, M.V., von Knethen, A., Angioni, C.,
Geisslinger, G., and Bru ¨ne, B. (2010). Cleavage of sphingosine kinase 2 by
caspase-1 provokes its release from apoptotic cells. Blood 115, 3531–3540.
Youle, R.J., and Strasser, A. (2008). The BCL-2 protein family: opposing
activities that mediate cell death. Nat. Rev. Mol. Cell Biol. 9, 47–59.
Molecular Cell 46, May 25, 2012 ª2012 Elsevier Inc.