MOLECULAR MECHANISMS OF CELL DEATH
Lloyd Greene 4/3/00 and 4/5/00
I. The phenomenon of cell death
A. Development: Cell death is an important element in sculpting body plan and design during
C.Elegans – stereotypic death (same 131 of 1090 cells)
Metamorphosis: Insects, amphibia
Formation of digits
Neuronal cell death
B. Elimination of cells with DNA damage
C. Defense from viral infection & viral counter measures
D. Maintenance of appropriate cell number
Regulation of cell proliferation in immune system
E. Disease: Cause and prevention
Oxidative stress - Infarction/stroke
II. Two types death distinguished classically: Apoptotic and Necrotic
Nuclear and cellular pyknosis Usually cell swelling
Chromatin condensation No chromatin condensation
Formation of apoptotic bodies Release cytoplasm
Engulfed, cleared by phagocytes Spew contents
No inflammation Inflammation
Nuclear endonuleases-ladder Random DNA degradation
Follows stereotyped pathways
Activation of caspases No caspase involved
May be mixed types of death-will cover later
III. Models of mammalian apoptotic death
Most mechanistic studies presently performed on cultured cells exposed to some sort of stimulus that
evoked apoptotic death. Examples of popular models include:
Removal of trophic support (withdrawal of specific growth factors, serum, hormones)
DNA damage: UV, x-rays, topoisomerase inhibitors
Free radicals: oxygen free radicals, peroxynitrite
Forced expression of oncogenes in quiescent cells: eg. c-myc
Exposure to death promoting ligands: Fas ligand, TNFalpha
Hormonal exposure: corticosteroids in T cells
Animal disease models, knockouts, transgenics
IV. C. Elegans and the basal elements of eukaryotic apoptosis
A. Death in C. elegans. In this worm, there is sterotypic death of approximately 10% of the
1,000 cells present. Horvitz and colleagues used mutagenic analysis to identify genes that regulate
death in this organism. Four major genes have been found that affect death. These have been termed
ced3, ced4, ced9 and egl-1. It turns out that homologues of these genes play major roles in apoptotic
death throughout the eukaryotic tree.
B. Ced3. Animals null for this gene showed no cell death. Therefore, it was judged to be
pro-apoptotic. Ectopic over-expression drives death.
C. Ced4. Nulls shows less death, but ectopic expression does not drive death. Thus, ced4 is
necessary, but not sufficient to cause death. Ced4 drives death when expressed with levels of ced3
insufficient to cause death in the absence of ced4. Thus, appears to facilitate ced3-dependent death.
D. Ced9. Null mutants show excess cell death. Ectopic over-expression blocks death even
when expressd in cells with ced3 and ced4. Thus, is anti-apoptotic.
E. Egl-1 Nulls show less death, thus pro-apoptotic. Overexpression causes excess death,even
in presence of Ced9.
F. Several genes have been identified that regulate expression of the above genes. If a cell is
destined to die , it expresses ced3 and ced4, but not ced9. Cells not destined to die either express ced9
or lack ced3/ced4.
V. CED3 AND CASPASES.
A. Ced3 is homologous to ICE, a protease. Sequencing of ced3 revealed that it is homologous
to a previously described mammalian enzyme known as ICE (Interleukin 1beta Converting Enzyme)
which is a protease that cleaves pro-interleukin 1b to form the mature form of the protein. Ced3 retains
the catalytic site (QACRG) and is an active protease.
B. ICE/Caspase family. Subsequent homology cloning revealed ICE to be a member of a
family of related mammalian enzmes. There are now 14 family members, many of which have multiple
splice forms. The family has been called the caspase family.
C. Properties of mammalian caspases.
1. Cysteine proteases (catalytic site contains active cysteine residue and generally in a
2. Are aspartases in that they cleave after aspartate
3. Different family members tend to recognize different motifs n-terminal to asp. For
instance ICE (called now caspase 1, cleaves at YVAD preferentially, whereas caspase 3 prefers
sequence DEVD and caspase 8, ITED.
4. Multiple casepases are present in living (non-dying) mammalian cells, but in
pro-forms that are nearly inactive. During apoptotic death, the proforms are activated. One major
mechanism is that the inactive proforms become cleaved, leading them to become. Cleavage occurs
after asp residues, indicating that activation is by either autocleavage or by another member of the
eg. inactive proN____D______D________C -->
proN___D + _____D + _____C -->
D______ C (Active) (Actually a tetramer)
5. Activation: Some caspases have long prodomains that interact with other proteins.
Examples are caspases 2,8 and 9. Caspase 2 binds to another molecule RAIDD that may play a role in
its activation. Caspase 8 binds to FADD which is involved in its activation via FAS receptors.
Caaspase 9 binds to APAF1 (see below) which is associated with its activation. It appears that
long-pro-domain caspases can be auto-activated when brought to high local concentration by binding to
suitable partners in complexes. Once activated, the long-prodomain caspases may then cleave caspases
with short prodomains (eg 3.7) to achieve their activation, leading to death. Thus, at least one means
that pro-caspases remain inactive in cells is that they are at rather low concentrations; death stimuli
appear to bring at least some members of the family into high local concentration, leading to their
activation and subsequent cleavage and activation of downstream caspases.
D. Evidence that caspases play an active role in apoptotic death
1. See activation in cells subjected to apoptotic stimuli. Can see the processed form
with antibodies, or can measure activity using peptide substrates that change fluorescence when cleaved
2. Over-expression causes apoptosis
3. Inhibition of activity blocks apoptotic death.
a. Inhibition can be achieved with peptide inhibitors which have been made that
block the activity by mimicking cleavage sites. For instance, zYVADfmk and zDEVDfmk (z=
blocking group and fmk=fluoromethylketone) are cell-permeant inhibitors that block caspases 1 and 3,
b. Viral inhibitors. One host defense mechanism against viruses is for the
infected cell to activate an apoptosis pathway,thus preventing viral replication. Viruses have therefore
developed anti-apoptotic proteins that prevent host cell death, at least until replication. One set of these
are inhibitors of caspases; examples include the p35 baculovirus protein and the varicella crmA protein.
Expression of these blocks several caspases, inhibiting death.
4. Knockouts lead to defects in cell death. For instance, caspase 3 and 9 knockouts
lead to a vast over-production of neurons in the brain and to resistance of thymocytes to apoptosis
evoked by a variety of means including serum withdrawal and exposure to glucocorticoids.
E. How do caspases evoke apoptosis?
One major target of caspases is other caspases. For instance, caspase 9 is in at least some
instances, upstream of caspase 3 and appears necessary for its activation. Cells null for caspase 9 do
not show caspase 3 activation in response to apoptotic stimuli, whereas caspase 3 null cells show
unimpaired caspase 9 activation. Thus, there appears to be a caspase cascade.
Clearly many potential targets in cells with many proteins having motifs for cleavage by
caspases. Examples of proteins known to be cleaved include actin (disrupts cytoskeleton), nuclear
lamins (disrupts nuclear membrane), Rb (disrupts cell cycle machinery), and PARP (disrupts energy
metabolism). An endonuclease known as CAD is also activated by cleavage of its inhibitory partner,
ICAD. These may all contribute to the final apoptotic death. Caspases may also work upstream in a
loop that leads to final death (see below).
VI. CED4 AND APAF1-Facilitators of caspase activation.
A. Ced4 role: As noted above, pro-caspases have a low, basal activity, but at physiological
levels of expression are not active enough or at high enough local concentrations to autoactivate to cause
cell death. but do not autocleave and activate because they are normally at low concentration in the
cytoplasm. Ced4 facilitates ced3 cleavage by forming oligomers and by binding ced3. This complex
brings ced3 (the inactive proform) into high local concentration which results in its autoactivation.
B. APAF-1. A long search led to discovery of a mammalian ced4 that has been designated as
apaf-1. Apaf-1 plays a key role in activation of caspase 9, a major element in certain forms of
mammalian apoptotic death. APAF participates in formation of an oligomer with caspase 9 and other
elements to be discussed below. Binding of caspase 9 to APAF1 occurs through the former's
N-terminal CARD domain. There are two theories about what happens to activate caspase 9 in the
complex with APAF1.. One is that in the complex, caspase 9 changes conformation and becomes
activated without autocleavage. The other is that the high local concentration permits autocleavage and
autoactivation of caspase 9. Both mechanisms probably play roles in caspase 9 activation.
VII. CED 9 AND BCL2 FAMILY MEMBERS.
A. Bcl2 as Ced9 homolog. In studying mechanisms of B-cell tumorigenesis, Korsemeyer and
associates isolated an oncogene named BCL2. It was found that BCL-2 is over-expressed in certain
lymphatic tumors cells, preventing naturally occuring death of B cells, leading to their formation of
lymphomas. When ced9 was sequenced, it was found to bear homology to bcl2. Thus was recognized
a set of anti-apoptotic proteins.
B. BCL2 is a member of a family of mammalian proteins. Homology cloning revealed in
mammals a family of additional bcl2-like proteins (numbering roughly 20 at present). All members have
at least several shared domains called BH domains. Four such domains have been defined and there are
involved in protein-protein interaction. Some members of the family, when expressed in cells were
anti-apoptotic. A major example is BCL-X. This is gene expressed in several forms including a full
length form called bclx-long which is anti-apoptotic. Many family members also have transmembrane
domains and are membrane associated.
C. Evidence that Bcl2 and certain other members of the family are anti-apoptotic and important
regulators of apoptotic cell death.
1. Ced9 ectopic expression in C.elegans blocks normally occuring death. Bcl2 has the
same effect in C.elegans.
2. Over-expression of Bcl2 and other anti-apoptotic members of this family block
apoptosis evoked by a wide variety of means. Exceptions exist, particularly for death evoked via
receptors such as Fas.
3. Animals null for anti-apoptotic members of the family show excess developmental
cell death. For instance KO for Bclx leads to massive neuronal death during development. Cells from
Bcl2 mice are more sensitive to apoptotic stimuli.
VIII. EGL-1 AND PROAPOPTOTIC BCL2 FAMILY MEMBERS.
A. EGL-1 Binds Ced9 and blocks its anti-apoptotic activity. Egl-1 contains a BH3 domain
which permits it to bind to ced9. By binding to ced9, EGL blocks its anti-apoptotic activity, thus
rendering EGL-1 as pro-apoptotic. In C. Elegans cells fated to die, EGL-1 binds ced9, thus premitting
activation of ced3 by ced 4.
B. Some members of the vertebrate bcl2 family are pro-apoptotic . Cloning of bcl2 family
members brought the surprising finding that over-expression of some family members brings about
death rather than protection. These were found to be pro-apoptotic. Such members include proteins
BAX, BAD, and BAK.
The presence of pro-apoptotic members is in many cases required for apoptotic death. For
instance, Bax KO mice are larger than controls and show excess cell numbers in many organs. Neurons
and thymocytes from such animals are resistant to apoptosis in culture.
C. Interaction between pro- and anti-apoptotic members of the Bcl2 family.
1. Importance of pro-anti ratio. Both pro- and anti-apoptotic members of the Bcl2
family are generally present in the same cell. This suggests that some interplay between the two
opposing actions leads to death or survival. Indeed, the death/survival can be regulated by the
stoichiometric ratio of pro- to anti-apoptotic Bcl2 family members in the cell.
2. Dimerization. Bcl2 family members can form homo- and hetero-dimers via their BH
domains. This includes dimers between pro- and anti-apoptotic members of the family. There is
evidence that at least in part, anti-apoptotic family members may protect when in abundance by bindng
up pro-apoptotic members and vice versa, that death can be triggered when excess pro-apoptotic forms
bind up anti-apoptotic forms. Thus, depending on which form is in abundance, death or survival
ensues. In such as scheme, any means that removes or inactivates one form, would favor the activity of
the other form. However, as raised below, there is evidence in vertebrates that this is not the only
means by which pro-apoptotic forms function.
D. The ced9/Bcl2 family works upstream of caspases. In cells rescued from apoptotic stimuli
by over-expression of anti-apoptotic Bcl2 family members, pro-caspases do not become activated.
Conversely, when pro-apoptotic Bcl2 members lead to caspases activation and caspase-dependent death.
These observations indicate that Bcl2 family members work upstream of caspases.
E. Death in C.Elegans. There is evidence that ced9 binds to ced3 and ced4, keeping them from
activating ced3. For a cell to die, Egl-1 binds ced9, sequestering it, thereby permitting ced3 to be
activated by binding to ced4. Things, however must be more complex in mammalian cells because they
contain the equivalents of ced3,4,9 and Egl-1, but can be triggered to die by apoptotic signals described
above. What therefore is responsible for such activation?
F. Interaction of BCLX-L with caspases and ced4/apaf. Several studies have now shown that
the bcl2 family member bclxl (like ced9) can bind to ced4 and apaf as well as to caspases such as
caspase 9. This has suggested the model that one means by which bcl2 family members may prevent
caspase activation is to compete for ced4/apaf and thereby prevent it from binding to and facilitating
activation of caspases. In this model, as in the c Elegans model, proapoptotic members such as BAX
would bind Bclx, thereby releasing caspase 9 and APAF1 to interact. Beware, however, that this model
is currently controversial and the evidence for it is open to question.
IX.. DEATH AND THE MITOCHONDRION: ROLE OF BCL2 FAMILY MEMBERS IN
APOPTOTIC DEATH OF VERTEBRATE CELLS.
A. What happens when death is triggered by an apoptotic stimulus? What permits regulated
death when all death molecules are present in a cell? Are there alternatives to the model in VIII F?
Yes, and the key turns out to be mitrochondria.
B. Cytochrome C and death. In an in vitro reconstitution system, Wang found that APAF1 and
caspase 9 were not sufficicient for activation. In purifying the additional needed factors from
cytoplasm of dying cells, he found that the missing factors were dATP and cytochrome C!
C. Mitochondria respond to apoptotic stimuli. Once cytochrome C was found to be involved,
attention focused on mitochondria which turn out to respond to a wide variety of apoptotic stimuli.
Such stimuli lead to loss of mitochondrial membrane potential and release of mitochondrial proteins
such as cytochrome c and a flavoprotein known as aif (apoptosis inducing factor). These responses turn
out to be an important feature of the vertebrate apoptotic mechanism and in many instances is the
upstream element of caspases in the apoptotic pathway. Thus, death stimuli lead to release of
cytochrome C from mitochrondria into the cytoplasm which in turn interacts with APAf1 and caspase 9
to form “apoptosomes”. The apoptosome permits activation of cytochrome C, either by autocleavage
or by favoring formation of an active conformation. Thus, mitochondrial release of cytochrome C is in
many instances a required element for caspase activation and death. In some apoptotic paradigms,
however, such as those triggered by death receptors, caspases may be triggered independently of
mitochondrial material and may in turn act on mitochondria to trigger release of cytochrome C ad other
materials (see below).
X. BCL2 FAMILY MEMBERS AND MITOCHONDRIA-WHAT REGULATES RELEASE OF
CYTOCHROME C FROM MITOCHONDRIA?.
B. BCL2 family members are present in mitochrondria and can form ion channels. Bcl2 and
many (but not all) members of this family are present outer mitochondrial membrane and may in some
manner affect mitochondrial stability and function. It is intriguing that the structure of bcl2 family
members is similar to that of bacterial pore-forming proteins and that they can form ion channels in
artificial membranes. The anti-apoptotic family members appear to aid in maintenance of
mitochondrial function and mitochrondrial membrane potential In contrast, pro-apoptotic family
members appear to drive loss of mitochondrial membrane potential. One model is that bax causes
deleterious effects on mitochondria and that this is countered by bcl2 normally and that in response to
death signals (through increases in bax levels, increased bax activity on mitochondria, increased bax
movement to mitochondria and/or by decreased counter-effects of bcl2). Note that this direct
mitochondrial function is not inconsistent with the ability of bclxL to interact with and sequester
ced4/apaf. In at least several death models the presence of BAX is required for death.
B. Permeability transition pore (PTP). PTP is a large conductance pore in mitochondrial
membranes that may contain or is influenced by a variety of molecules present in the inner and outer
mitochondrial membranes. It has been proposed that BAX binds to this and activates it, permitting
mitorchondrial depolaraization and ultimately, cytochrome C release. BCl2, has been proposed in
contrast, to stabilize the pore. Although BAX might work in part by antagonizing BCL2, it also
appears to have direct effects on mitochondria. For instance, in yeast which have to BCL2-like
proteins, BAX affects mitrochondria and causes death.
C. Movement of BAX etc to mitochondria and death. At least in some cell models, BAX
appears to be cytoplasmic in living cells and then in response to death stimuli, to move to mitochondria.
Other proapoptotic members of the family also move to mitochondria after apoptotic signals including
BIM, BAK, BID, and BAD. (We shall return to BID later). Thus, the major signal for release of
cytochrome C and subsequent activation of casepase 9 is movement of a pro-apoptotic BCL2 family
member to mitochondria. These may either antagonixe anti-apoptotics (by bindng to them) or may
have direct effects on the mitochondria.
XI. MITOCHONDRIA AND NON-APOPTOTIC DEATH:
Loss of cytochrome C from mitochondria can trigger rapid apoptotic death. However, loss of
mitochondrial function is itself fatal to cells, so that if apoptotic death does not occur (for instance in
presence of caspase inhibitors) cells may still die by a non-apoptotic mechanism due to absence of ATP
XII. UPSTREAM OF BCL2 AND THE MITOCHONDRION: WHAT TRIGGERS
MITOCHONDRIAL PROBLEMS AND MOVEMENT OF PROAPOPTOTIC BCL2 FAMILY
A. Different death initiators use a variety of means to trigger mitochondrial dysfunction and
apoptotic death. There seem to be a variety of routes to mitochondrial release of cytochrome C and
activation of caspases, depending on the initiating means of death and the particular cell type involved.
B. Transcription. In many instances of apoptosis triggered for example by growth factor
deprivation, DNA damage or exposure to initiators such as corticosteroids, mitochrondrial responses,
caspase activation and apoptotic death are blocked by inhibitors of protein and RNA synthesis. This
suggests that apoptotic stimuli lead to synthesis of "death genes" responsible for death. The nature of
such genes is in general unclear. In a few cases, apoptotic stimuli enhance levels of pro-apoptotic
BCL2 family members such as BAX. In other cases, the ability of the cell to handle oxidative stress
appears to be decreased, perhaps contributing to mitochondrial dysfunction. In yet other cases, there
appears to be upregulation of the Fas/FasL (see below) system which in turn triggers death by activation
of FAS. However, the search is on for additional regulated proteins that will affect mitochondrial
function in the apoptotic pathway.
C. Upstream pathways that affect transcription and apoptosis.
1. p53. It is clear that p53 plays a critical role in death triggered by a variety of DNA
damaging agents. For instance, cells of p53 null mice are resistant to death caused by DNA damage.
DNA damage often increases p53 expression. p53 seems to be a guardian of the genome and if
excessive DNA damage occurs, p53, by unknown means (which appear to involve transcription) in turn
triggers apoptotic death. p53 may also play a role in death triggered by oxidative stress (which also
causes DNA damage) and perhaps even in death caused by other means such as trophic factor
deprivation. Bax is in some cases regulated by p53, but other genes must be involved, In many
cancers, p53 is absent or defective, thus permitting cells with DNA damage to continue to replicate,
even in the face of anti-tumor agents. This is one prime example of how the apoptotic pathways
interface with cancer.
2. c-Jun. In some dying cells, c-Jun is upregulated or becomes hyperphosphorylated
and plays a required role in death as shown by the protective action of dominant negative jun. The
downstream genes here are also unclear , but among them may be FAS/FASL
3. JNKs. Jun terminal kinases (JNKs) are activated by a variety of cellular stresses
including apoptotic stimuli and also play an active and required part in certain death paradigms. The
pathway includes cdc42/rac1 (small G proteins)/MAPKK4/7 and JNKs. KO of JNKs can block death in
vivo as well as in vitro. Blockade of this pathway blocks death in many, but not all paradigms. There
are 3 JNKs, double Kos show defects in developmental death. Among possible targets, JNKs in turn
phosphorylate c-Jun and regulate its activity.
4. NFKappaB. Depending on the cell type and initiating cause of death, NFKB can be
either pro- or anti-apoptotic. For instance, in some cell systems NFKb appears to protect from death
evoked by TNFalpha and to play a required role in death evoked by oxidative stresses.
5. Cell cycle molecules and E2F. Early studies showed that if c-myc was expressed in
otherwise quiessecent cells, it would lead to their apoptotic death. This suggests that forcing cells
inappropriately into the cycle can lead to apoptosis. In neurons dying due to various causes, cell cycle
marker proteins (cyclins and cdks) are turned on, indicating that here too cell cycle may be involved. In
this case, inhibitors of cell cycle can rescue from death. One key player in this aberrant cell cycle death
appears to be the transcription factor E2F. Non-regulated activation and expression of E2F caused by
aberrant cell cycle activity participates in gene regulation that triggers apoptotic death via mitochondrial
dysfunction and casepase activation
6. Forkhead. This transcription factor is normally in a phosphorylated state and kept
out of the nucleus by binding to 14-3-3 protein. In some dying cells such as neurons, forkhead becomes
dephosphorylated and moves to the nucleus where is appears to particpate in the death mechanism.
D. Multiple upstream elements. It appears that a single upstream element as described above
may not be sufficient to trigger apoptotic death. For instance activation of Jun alone does not lead to
death. It seems more likely that death requires the simultaneous participation of several transcriptional
regulatory elements. This may serve as a fail-safe system to prevent accidental death.
XIII. RECEPTOR MEDIATED DEATH IN MAMMALIAN CELLS.
A. Suicide and murder: In many of the paradigms discussed above, cells commit suicide in
response to stress, infection, damage or loss of trophic support. A system of cell execution has also
evolved in mammalian cells. This involves a specific set of receptors and ligand which trigger death.
B. The TNF (tumor necrosis factor) alpha superfamily and ligands. This is a superfamily of
transmembrane receptors found in many tissues. Receptor (and ligands) include: FAS/CD95 (FAS
ligand or FASL); TNFR1 (TNF); TRAILR1 and R2 (TRAIL); and the p75 NGF (NGF) receptor. FAS
and TNRF1 have a so-called death domain in their intracellular domains and have been the best
C. Initiating apoptotic death via TNFa receptors. Binding of ligand causes receptor
trimerization. This in turn permits binding of adapter proteins with C-terminal death domains such as
TRADD, FADD and RAIDD. These in turn interact with caspases 8 or 10 via mutual DEDs (death
effector domain). This recruitment of caspases 8 or 10, by bringing them into high local concentrations,
causes their autocleavage and activation. Thus, ligand binding leads to direct and rapid activation of a
D. How does activation of caspase 8 cause death?
1. Mitochondrial independent death. In some cases, death caused by activation of TNF
superfamily receptors occurs rapidly and cannot be blocked by anti-apoptotic members of the BCL2
family. In such cases, it appears that caspase 8 leads to activation of caspases 3,6,7 and other effector
caspases, thus leading to death by a mitochondrial independent mechanism. This appears to be the
major pathway when caspase 8 levels are high in cells.
2. Mitochondrial/BCl2 family- dependent death. In most cases, death via TNF family
members is dependent on mitochondrial events and is sensitive to BCL2 family members. A key player
here appears to be the BH3-only BCL2 family member BID. BID is cleaved by activated caspase 8,
forming a 15 Kd C terminal fragment which translocates to mitochondria. The cleaved BID fragment
appears to act as a chaperone for BAX and has been proposed to bring it also to the mitochondrion
where both participate in release of cytochrome C and thereby, of death via caspases 9 and 3,6,7. This
pathway thus brings amplification of a weak caspase 8 signal. Caspase 3 can also cleave BID, perhaps
setting off a loop upstream of the mitochondrion to further amplify a weak death signal.
E. FAS, TRAILR and death evoked by transcription-dependent mechanisms. One target of the
transcription dependent mechanisms mentioned above is FAS. For instance, FAS can be upregulated
via JUN, forkhead and p53. There is evidence in some systems that various stresses can lead to
upregulation of FAS levels in cells and that this contributes, via endogenous FASL, to apoptotic death.
Thus, it appears that many different apoptotic pathways can be activated in cells and that these can
variably contribute to the final event-cell death. Recent evidence indicates that TRAILRs may also be
downstream targets of p53 and are upregulated in some cells in a p53-dependent manner, leading to
XIV. ADDITIONAL PROTECTION MECHANISMS.
A. Why. Given that many molecules involved in the apoptotic machinery are constitutively
present in cells, additional means have evolved (in addtion to BCL2 family members) to assure that
death does not occur by accidental or inadvertent activation.
B. Decoy receptors/adapters/caspases. .
1. TRAILR decoys. TRAIL receptors that are truncated in their intracellular domains
have been found in many tissues. These bind TRAIL ligand and, in one case, even FASL, but lack a
death domain and cannot bind adapters or caspases. These compete with full length receptors for
ligand, thus conferring protection. These ,may be upregulated in certain tumors, thereby making them
more resistant to death.
2. c-FLIP (FADD-like ICE inhibitory protein): Decoy/dominant negative adapter. This
has DED domains and has a caspase 8-like domain that is inactivatable. It binds to FADD and
procaspase 8 and thereby block interactions with receptor. This provides protection from death caused
by death receptors, but not by other apoptotic stimuli. A viral form of FLIP has also been found which
presumably prevents death upon infection.
3. SODD-silencer of death domains. This binds to inactivated TNFR1 via a death
domain and, prevents ligand unoccupied receptors from binding adapters and caspases.
C. IAPs. Inhibitors of apoptosis proteins This family confers protection from a wide varieity
of death stimuli. Five forms have been described in mammalian cells (c-IAP1,2, XIAP, survivin and
NAIP) and forms are also present in viruses and flies. IAPs appear to inhibit a variety of caspases as
well as caspase processing, each with a different set of specificities. Several IAPs also have CARD
domains and thereby bind to caspase 9 and APAF1 and perhaps other molecules in the apoptotic
pathway. A second domain shared by all in the BIR domain which also is required for interaction with
XI. WHAT DO TROPHIC GROWTH FACTORS DO THAT BLOCKS DEATH AND WHY DO
CELLS DIE WHEN DEPRIVED OF GROWTH FACTORS?
A. PI3K. A major signaling pathway activated by growth factors and their receptor tyrosine
kinases is the phosphatidyl inositol 3' kinase pathway. In a number of cell systems in in which survival
is supported by growth factors, death can be triggered by blocking PI3K with inhibitors such as the drug
wortmannin. Thus PI3K appears to be at the top of a survival pathway.
B. AKT. This protein kinase is a downstream target activated by the phosphoinositides
generated by activation of PI3K. AKT is a major mediator of survival. Constitutively active forms
promote survival in absence of growth factors and dominant negative forms block growth factor
C. What is downstream of AKT and how does it block death? Thus far, this not entirely
worked out, but some clues exist. In general, it seems that there are many targets for AKT and that it
therefore blocks the apoptotic pathway at multiple sites..
Caspase 9 is substrate for AKT (at least in humans) and once phosphorylated, cannot be
activated. This would be a fairly downstream target.
BAD, a proapoptotic member of the bcl2 family, is also a target. When phosphorylated
it binds 14-3-3 and may be sequesered away from mitochondria. When growth factors are removed,
AKT becomes inactive and BAD becomes dephosphorylated and moves to mitochondria where it
participates in death.
Forkhead. As noted above, the transcription factor forkhead is also a target of AKT
which keeps it out of the nucleus.
GSK3beta This kinase is inhibited by AKT dependent phosphorylation. When AKT
turns off, GSK3b dephosphorylates and activates. This appears to be important in some death pathways
in that inhibition of GSK3b with Li or with D/N GSK3b blocks death. The relevant targets for GSK3b
in the apoptotic pathway are currently unknown.
D. Ras. In some cell systems, activation of ras also promotes survival in absence of growth
factors. Thus, the ras pathway may also promote growth factor mediated survival. In some cases, this
is achieved by activation of PI3K downstream of ras, and in others this is not the case, so there must be
additional downstream targets. One of these was recently identified as the kinase RSK (ribosomal S6