|
|
||||||||
1 Immunology and Disease Resistance Laboratory, United States Department of Agriculture-Agricultural Research Service, Beltsville 20705; and 2 Department of Medicine, University of Maryland School of Medicine, Baltimore, Maryland 21201
| |
ABSTRACT |
|---|
B ACTIVATION...
|
|---|
Gram-negative bacterial sepsis remains a common, life-threatening event. The prognosis for patients who develop sepsis-related complications, including the development of acute respiratory distress syndrome (ARDS), remains poor. A common finding among patients and experimental animals with sepsis and ARDS is endothelial injury and/or dysfunction. A component of the outer membrane of gram-negative bacteria, lipopolysaccharide (LPS) or endotoxin, has been implicated in the pathogenesis of much of the endothelial cell injury and/or dysfunction associated with these disease states. LPS is a highly proinflammatory molecule that elicits a wide array of endothelial responses, including the upregulation of cytokines, adhesion molecules, and tissue factor. In addition to activation, LPS induces endothelial cell death that is apoptotic in nature. This review summarizes the evidence for LPS-induced vascular endothelial injury and examines the molecular signaling pathways that activate and inhibit LPS-induced endothelial apoptosis. Furthermore, the role of apoptotic signaling molecules in mediating LPS-induced activation of endothelial cells will be considered.
endotoxin; inflammation; nuclear factor-
B; sepsis; vascular
injury
| |
INTRODUCTION |
|---|
B ACTIVATION...
|
|---|
DESPITE ADVANCES IN ANTIMICROBIAL therapy and overall medical care, gram-negative bacterial sepsis remains a common, life-threatening event (163). Complications arising from sepsis include disseminated intravascular coagulation (133, 197, 201), systemic vascular collapse (88, 201), multiorgan failure (22, 23, 65), and the development of vascular leak syndromes, including acute respiratory distress syndrome (ARDS) (26, 27, 142, 144, 161). One common denominator to all of these complications is endothelial cell (EC) injury and/or dysfunction (46, 47, 80, 132, 148, 153).
The vascular endothelium serves as the key barrier between the
intravascular compartment and extravascular tissues and plays a
critical role in a large number of physiological and pathological processes (38). ECs are integrally involved in regulating
blood flow, coagulation, leukocyte trafficking, edema formation, wound healing, and angiogenesis. Because of their location at the
blood-extravascular tissue interface, ECs are constantly exposed to
circulating mediators that may perturb the above-mentioned endothelial
barrier functions. One such mediator that targets the endothelium is
bacterial lipopolysaccharide (LPS) or endotoxin, a highly
proinflammatory molecule that is a component of the outer envelope of
all gram-negative bacteria (17, 159, 176). LPS is released
from the surface of replicating and dying gram-negative bacteria into
the circulation, where it interacts with the endothelial lining of the
vessel wall (173, 176). Evidence exists that LPS, alone or
in concert with other endogenous factors, is responsible for much of
the EC injury and/or dysfunction associated with gram-negative sepsis.
First, LPS bioactivity has been detected in the bloodstream of
gram-negative septicemic patients, and in selected studies, levels of
circulating LPS predict the development of multiorgan failure,
including ARDS (26). Second, administration of LPS alone
to experimental animals reconstitutes the EC injury seen after
gram-negative bacterial challenge (27, 201). Third,
immunological and pharmacological interventions that specifically
target the LPS molecule protect against these same vascular
complications (6, 210, 211, 225). Finally, LPS directly
elicits several of the EC responses in vitro that are similarly evoked
during sepsis, including: 1) the production of the
proinflammatory cytokines IL-6 (77, 109), IL-8 (7, 77, 223), and IL-1
(57, 138); 2) the
increased surface expression of the adhesion molecules E-selectin,
intercellular adhesion molecule (ICAM)-1, and vascular cell adhesion
molecule (VCAM)-1 (34, 59, 77, 108); and 3)
increased expression of tissue factor (42, 75). In
addition to activation, LPS induces EC-programmed cell death or
apoptosis (18, 21, 64, 67, 82, 95, 116, 143, 221),
an event that is believed to contribute to the pathogenesis of sepsis
and its attendant complications (82, 189). Several of the
signaling molecules involved in EC activation leading to increased
cytokine and adhesion molecule expression are similarly involved in LPS
proapoptotic signaling. Furthermore, there is evidence that LPS
proapoptotic signaling molecules have an additional role in
regulating LPS-induced NF-
B activation and nonapoptotic EC
responses to LPS. This review focuses on recent advances in the
elucidation of the mechanisms by which LPS elicits EC apoptosis
and the cross talk between signaling pathways leading to EC activation
and death.
| |
EVIDENCE FOR LPS-INDUCED EC APOPTOSIS |
|---|
B ACTIVATION...
|
|---|
Apoptosis is an ATP-dependent form of cell death, morphologically characterized by chromatin condensation, nuclear fragmentation, cell shrinkage, and blebbing of the plasma membrane (76, 171). The end result of apoptosis is fragmentation of the cell into small membrane-bound bodies that are quickly cleared by phagocytotic cells (45, 181, 187). Biochemically, apoptosis is characterized by the activation of caspases, highly specific proteases that cleave a wide array of intracellular substrates (198). Activation of upstream caspases initiates a proteolytic cascade leading to DNA fragmentation and the cleavage of key regulatory proteins resulting in cell death. In contrast to apoptosis, necrotic cell death is an energy-independent process characterized by cell swelling and lysis (76, 171). Unlike apoptotic cells, necrotic cells release cellular constituents that elicit an inflammatory reaction in surrounding viable tissue.
LPS induces apoptosis in bovine and ovine ECs in vitro (18, 64, 94, 95) and elicits human EC apoptosis in the absence of new gene expression (164). That LPS-induced EC death is apoptotic in nature has been confirmed by several criteria, including morphological changes (64), DNA laddering (18, 99), TdT-mediated dUTP nick end labeling (18, 94, 95), nuclear histone release (15, 21), caspase activation (15, 19), and poly(ADP-ribose)polymerase cleavage (21, 36). Several studies have demonstrated that purified LPS itself, in the absence of host-derived mediators, evokes EC injury and/or apoptosis (18, 36, 64, 67, 82, 95, 116, 221). The ability of LPS to induce EC apoptosis in the absence of non-EC-derived host mediators is compatible with a direct effect. Furthermore, neutralizing antibodies to other known death receptors expressed on EC, including TNF type 1 receptor (TNFR1), Fas, and death receptor 3 (DR3), fails to inhibit LPS-induced apoptosis (36). Finally, LPS-induced EC apoptosis occurs independently of new protein synthesis, thereby precluding the involvement of upregulated gene products (84).
| |
CLINICAL RELEVANCE AND IN VIVO EVIDENCE |
|---|
B ACTIVATION...
|
|---|
Endothelial apoptosis has been implicated in the pathogenesis of several disease states, including atherosclerosis (30), hypertension (73), congestive heart failure (204), and systemic capillary leak syndrome (12). Several studies have reported that EC injury and/or death is a key pathological finding during bacterial sepsis. First, injection of Escherichia coli into rabbits (47) or baboons (40, 41) induces severe microvascular injury and EC detachment. Second, increased numbers of apoptotic ECs are detected in the pulmonary capillaries of a murine model of sepsis (93, 217). Third, evidence of EC injury is observed in postmortem biopsies obtained from patients who have died of sepsis-related ARDS (148). Fourth, an increase in circulating EC is observed in septic patients, and the magnitude of this increase correlates negatively with survival (153).
There are several reports that purified LPS elicits EC injury and
apoptosis in vivo. EC injury and/or detachment from the vascular wall has been reported after LPS injection into mice (124), rats (172, 192), rabbits (47,
69, 145), dogs (72, 127), sheep (27,
149), and primates (14). Liver sinusoidal ECs
obtained from LPS-treated rats display enhanced activation of
caspase-3, a central apoptotic effector protease (49).
Disseminated EC apoptosis has been reported in the lung, liver,
thymus, and intestine of mice challenged with LPS (67, 82, 116,
122). Finally, injection of a broad spectrum caspase inhibitor
decreases EC apoptosis in the lung after LPS administration and
improves survival in a murine model of acute lung injury
(116). The combined in vitro and in vivo data implicate
apoptosis as a key component of the EC response to LPS.
However, the extent to which LPS directly induces EC apoptosis
in human systems either alone or in concert with other known
apoptosis-inducing proinflammatory cytokines, such as TNF-
and/or IL-1
, remains unclear.
| |
CELL SURFACE RECOGNITION OF LPS |
|---|
B ACTIVATION...
|
|---|
As the name implies, LPS is composed of both lipid and polysaccharide components. The lipid portion is composed of a unique lipid, lipid A, which is the most widely conserved region of the LPS molecule (174). It is well established that the lipid A moiety of LPS is responsible for its proinflammatory properties (174). Lipid A alone is capable of eliciting EC responses identical to those induced by LPS (16, 50, 193). Agents that specifically target the lipid A moiety of LPS inhibit EC activation (16, 52) and protect against the development of vascular complications in endotoxin shock models (6, 210). Similar to activation, the lipid A moiety is also responsible for the proapoptotic properties of the LPS molecule (84). Neutralization of lipid A with polymyxin B completely abrogates LPS-induced EC apoptosis (18, 19, 64).
An early identified receptor implicated in cellular recognition of LPS
was membrane-associated CD14 (mCD14) (202). mCD14 is a
glycoprotein found on cells of monocytic origin and to a lesser extent
on neutrophils. Although LPS can directly bind mCD14, its affinity for
the receptor is greatly increased when LPS is complexed with the acute
phase protein LPS-binding protein (LBP) (159). ECs, which
lack mCD14, are activated by LPS in the presence of LBP and soluble
CD14 (sCD14), the latter of which is released from mCD14-bearing cells
(9, 74). It has been proposed that LBP facilitates the
transfer of LPS to sCD14 and that this LPS-sCD14 complex is recognized
by ECs (202, 205). Similar to EC activation, a requisite
role for sCD14 in mediating LPS-induced apoptosis has been
established. In the absence of sCD14, ECs are resistant to LPS-induced
apoptosis (9, 64, 219). In the presence of CD14-containing serum, anti-CD14 antibodies inhibit LPS-induced EC
cytotoxicity (9, 64, 219). Because CD14 is a glycosyl phosphatidylinositol-anchored protein and lacks an intracellular cytoplasmic domain, it was postulated that another transmembrane receptor must exist that can transduce LPS signaling across the plasma
membrane. This receptor, which has since been identified as Toll-like
receptor (Tlr)-4, is expressed in cells of monocytic lineage as well as
in non-mCD14-bearing ECs (37, 61). There has been some
controversy in the past regarding whether another member of the
Toll-like receptor family, Tlr-2, mediates LPS-induced signaling. Two
groups have reported that LPS-induced activation is mediated by Tlr-2
(121, 218). However, a subsequent study reported that
certain commercial preparations of LPS are contaminated with bacterial
lipoproteins and that these lipoproteins, not LPS, are responsible for
Tlr-2 activation (90). The genetic evidence clearly
establishes Tlr-4 as a true LPS receptor (166). C3H/HeJ and C57BL/10ScCr mice, which contain a missense mutation in or a null
mutation for the Tlr-4 gene, respectively, are resistant to
LPS (166). Furthermore, transfection of Tlr-4 into
LPS-insensitive HEK-293 cells confers sensitivity to LPS-induced
NF-
B activation (37). Finally, anti-Tlr-4 antibodies
abrogate LPS-induced EC NF-
B activation, whereas anti-Tlr-2
antibodies have no inhibitory effect (61).
| |
TLR-4 SIGNALING LEADING TO NF- B ACTIVATION |
|---|
B ACTIVATION...
|
|---|
Tlr-4 is integrally involved in LPS signaling and has a requisite
role in the activation of the transcription factor NF-
B. The
extracellular domain of Tlr-4 contains repeating leucine-rich motifs
characteristic of innate immune response pattern recognition receptors
(Fig. 1) (48). The
cytoplasmic domain contains regions that are homologous to the
intracellular signaling domain of the type 1 IL-1 receptor. Although
the exact mechanism by which LPS is recognized by Tlr-4 remains
unclear, cell activation is dependent on the cell surface assembly of a
multiprotein recognition complex consisting of CD14, MD-2, and Tlr-4
(3) (Fig. 2). After
activation of the Tlr-4 receptor complex, the adapter protein
myeloid differentiation factor 88 (MyD88) is recruited to the
cytoplasmic domain of Tlr-4 through homotypic binding of respective
Toll receptor-IL-1 receptor (TIR) domains (146, 154).
MyD88 contains an additional protein-binding domain, the death domain
(DD) (Fig. 1), which facilitates its association with another
DD-containing signaling molecule, IL-1 receptor-associated kinase-1
(IRAK-1) (48). After autophosphorylation, IRAK-1
dissociates from MyD88 and interacts with TNF receptor-associated factor-6 (TRAF-6) (137, 194), resulting in the activation
of a downstream kinase cascade involving NF-
B-inducing kinase
(NIK) and I
B kinase (IKK) (Fig. 2). The IKK-mediated phosphorylation of I
B, an inhibitor of NF-
B, leads to I
B degradation through the proteasome pathway and enables NF-
B to translocate to the nucleus where it promotes new gene expression (48).
|
|
| |
TLR-4 SIGNALING LEADING TO APOPTOSIS |
|---|
B ACTIVATION...
|
|---|
Several of the upstream signaling molecules involved in LPS
activation of NF-
B are similarly involved in promoting LPS-induced apoptosis. Macrophages derived from C3H/HeJ mice, which have a missense mutation in the third exon of Tlr-4 (166), are
resistant to LPS-induced apoptosis (110).
Furthermore, neutralization of the lipid A moiety of LPS, which is the
domain of LPS recognized by Tlr-4 (139), protects against
LPS-induced EC apoptosis (18, 19). Downstream of
Tlr-4, MyD88 and IRAK-1 have been implicated in mediating LPS-elicited
cell death signaling. Expression of either MyD88 or IRAK-1
dominant-negative (D/N) constructs, which inhibit LPS-induced NF-
B
activation (19, 222), protects against LPS-induced EC
apoptosis (19). MyD88 has similarly been shown to
mediate Tlr-2 activation of NF-
B and proapoptotic signaling (5). Downstream of MyD88 and IRAK-1, TRAF-6 has been
implicated in promoting LPS-induced EC apoptosis
(100). TRAF-6 participation in LPS-induced
apoptosis involves c-jun NH2-terminal
kinase, the activation of which lies upstream of caspase activation.
Together, these data indicate that Tlr-4 can serve as a death receptor
for LPS and that the signaling molecules involved in LPS-induced
NF-
B activation serve a dual role in promoting LPS-induced
apoptosis (Fig. 2).
There is evidence suggesting that Tlr-4 activation of NF-
B can occur
through a MyD88- and IRAK-1-independent pathway. The LPS-induced
DNA-binding activity of NF-
B in macrophages derived from either
MyD88 or IRAK-1 knockout mice is delayed, but not inhibited, indicating
that cellular activation by LPS can occur in the absence of these
signaling molecules (115, 194). Recently, an MyD88-like
protein has been identified by two independent groups, MyD88
adapter-like protein (MAL) or TIR domain-containing adapter protein
(TIRAP), which promote LPS-induced NF-
B signaling through IRAK-2
(62, 91). MAL/TIRAP contains a COOH terminus TIR domain but lacks the NH2 terminus DD present in MyD88 (Fig. 1).
Similar to MyD88, MAL/TIRAP has an additional role in promoting
LPS-induced EC apoptosis (15). These studies
demonstrate the presence of an MyD88-independent pathway that serves a
redundant signaling role in promoting both LPS-induced NF-
B
activation and apoptosis in ECs (Fig. 2).
Although Tlr-4 and its respective intracellular binding partners MyD88
and MAL/TIRAP have been shown to mediate LPS-induced apoptosis,
questions remain as to how this signaling pathway activates the
effector proteases of apoptosis, the caspases. The
Fas-associated death domain (FADD) is a proapoptotic adapter
protein that couples death receptors to initiator caspases (24,
35, 96, 126). Its role in mediating apoptosis has been
well elucidated in death signaling initiated by the Fas and TNF-
receptors. FADD recruitment, either directly to Fas or through the
intermediary TNF receptor-associated DD protein in the case of the
TNF-
receptor, is mediated through binding between the two highly
conserved DDs found on these proteins. Similarly, FADD and procaspase-8
interact through the death effector domains (DED) contained by each
protein (Fig. 1). Procaspase-8 contains partial proteolytic activity
that enables it to cleave other procaspase-8 molecules brought into
close proximity after recruitment to FADD (155).
Activation of caspase-8 initiates a proteolytic cascade resulting in
the activation of downstream effector caspases, including caspase-3
(89). Although a role for FADD in mediating TNF-
- and
Fas-induced apoptosis has been well established, there are
conflicting reports as to whether this molecule is involved in LPS
proapoptotic signaling. Choi et al. (36) reported that
expression of a FADD D/N protected against LPS-induced EC
apoptosis, whereas another group showed that the same FADD D/N
had no effect (19). It is difficult to reconcile these
differences because both studies utilized the same EC type and a
comparable retrovirus-based system to stably express identical FADD D/N
constructs. Furthermore, both studies were able to demonstrate the
efficacy of the FADD D/N construct in protecting against
TNF-
-elicited apoptosis. One major difference between the
studies was in the assay used to measure cell death. In the study by
Choi et al. (36), mitochondrial activity was used as a
measure of cell viability. This method of viability, however,
does not discriminate between necrotic and apoptotic cell death. In
the contrasting report, a more selective assay for determining
apoptotic cell death, which involved measuring caspase
activity, was used (19). One explanation for these
differences may be that LPS induces both apoptotic
(caspase-dependent) and nonapoptotic (caspase-independent) cell
death and that FADD plays a role in mediating the latter.
Tlr-4 is one member of a larger family of pattern recognition receptors
involved in the innate immune response. Another member of the Tlr
family, Tlr-2, recognizes the cell wall components peptidoglycan and
bacterial lipoproteins found in both gram-positive and -negative
bacteria as well as gram-positive-restricted lipoteichoic acid
(4, 186). Several of the intracellular signal transduction molecules involved in Tlr-4-mediated activation of NF-
B, including MyD88, IRAK, TRAF-6, NIK, and IKK, are similarly involved in Tlr-2 signaling. Interestingly, bacterial lipoprotein-elicited Tlr-2 proapoptotic signaling is reportedly mediated by MyD88 and FADD, two molecules similarly involved in LPS/Tlr-4-induced apoptosis (5). In contrast, Tlr-2-mediated apoptosis occurs
independently of TRAF-6 involvement (5), a signaling
molecule recently established to participate in LPS/Tlr-4
proapoptotic signaling (100). Thus divergence in the
apoptotic signaling pathways mediated by Tlr-2 and Tlr-4 occurs
downstream of MyD88.
| |
ROLE OF NF- B ACTIVATION IN MEDIATING LPS-INDUCED APOPTOSIS |
|---|
B ACTIVATION...
|
|---|
LPS-induced activation of NF-
B is a key signaling event that
mediates an array of EC responses, including increased 1)
IL-6 (109), IL-8 (7, 223), and IL-1
(138) production, 2) E-selectin, ICAM-1, and
VCAM-1 surface expression (34, 108), and 3)
tissue factor activity (42, 75). In addition to its role
in promoting the expression of proinflammatory gene products, NF-
B
has been implicated in both pro- (86, 178) and
antiapoptotic signaling (128, 203). In ECs, inhibition
of NF-
B activation sensitizes human ECs to direct TNF-
-induced
apoptosis in the absence of cycloheximide, suggesting an
antiapoptotic role for NF-
B (221). Evidence exists
that this sensitization is conferred by inhibition of NF-
B-dependent
expression of members of the inhibitors of apoptosis (IAP) gene
family (190). In contrast to TNF-
, inhibition of
NF-
B does not sensitize human ECs to direct LPS-induced
apoptosis (221). These data, therefore, preclude
an antiapoptotic role for NF-
B in conferring protection against
LPS-elicited apoptosis.
There are reports that NF-
B signaling has a role in proapoptotic
signaling (86, 178). On the basis of the predominant role
of NF-
B in mediating LPS-elicited EC responses and the finding that
MyD88, MAL/TIRAP, and IRAK-1 promote both LPS-induced NF-
B activation and apoptosis (Fig. 2), there has been speculation that these events are coupled. Several lines of evidence, however, suggest that these events occur independently of one another. First,
the presumed mechanism of the proapoptotic properties of NF-
B
reported by Ryan et al. (178) is through the promotion of
new gene expression, similar to its antiapoptotic role and the
induction of antiapoptotic proteins. The finding that LPS elicits
EC apoptosis in the absence of new protein synthesis argues against a proapoptotic role for NF-
B that involves new gene
expression (21). Second, inhibition of NF-
B activation
by overexpression of an I
B
superrepressor fails to block
LPS-induced EC apoptosis (19, 21). This latter
finding also demonstrates a bifurcation in the NF-
B and
apoptotic pathways that mediate LPS signaling upstream of I
B
degradation. Finally, complete inhibition of apoptosis with the
caspase inhibitor zVAD-FMK does not inhibit the ability of LPS to
activate NF-
B (19). Thus despite a commonality among upstream signaling molecules, LPS/Tlr-4-induced NF-
B activation and
apoptosis are mutually independent events (Fig. 2).
| |
ROLE OF CASPASES |
|---|
B ACTIVATION...
|
|---|
A hallmark of apoptosis is the activation of highly specific effector proteases of the caspase family. Caspases exist as inactive zymogens (proenzymes) that are activated by proteolytic processing of the procaspase molecule in one of three ways: 1) autoactivation due to low levels of intrinsic catalytic activity, 2) transactivation by other caspases within close proximity, or 3) activation by noncaspase proteases (198). Once activated, caspases cleave and activate other members of the caspase family leading to amplification of a proteolytic cascade. The end result is a series of proteolytic events that lead to the cleavage of intracellular substrates, chromatin condensation, DNA fragmentation, and eventual cell death.
Of the 14 known caspases expressed in mammalian cells, LPS has been
reported to activate caspase-1 (152, 185), caspase-3 (15, 49, 100, 111, 152), caspase-6 (111), and
caspase-8 (100, 111). Activation of caspase-1
(interleukin-1 converting enzyme) is primarily responsible for IL-1
processing and has limited involvement in apoptosis
(43). In contrast, caspase-3, -6, and -8 all have been
established to prominently participate in apoptosis. Caspase-8,
also known as FADD-like interleukin converting enzyme protease (FLICE),
is classically described as an initiator caspase that is recruited to
such death receptors as TNFR and Fas. Procaspase-8 has
intrinsically low levels of proteolytic activity that enables it
to cleave other procaspase-8 molecules brought into close proximity
after recruitment to the receptor (155). Activation of
caspase-8 initiates a proteolytic cascade resulting in the activation
of downstream effector caspases and the onset of apoptosis
(89). A key target of caspase-8 is procaspase-3. Proteolytic processing of procaspase-3 results in an active caspase-3 molecule that serves as a key effector caspase responsible for much of
the proteolysis associated with apoptosis (89,
107). The effects of LPS-induced caspase activation on ECs are
quite dramatic, resulting in the cleavage of nuclear proteins as well as structural proteins that mediate cell-cell and cell-substrate adhesion (18, 21). In vivo studies support a role for
caspases in mediating LPS-induced EC apoptosis (49,
116). Liver ECs derived from mice administered intravenous LPS
display enhanced capase-3 activity (49). Furthermore,
inhibition of caspase activation with the cell-permeable, pan-caspase
inhibitor peptide zVAD protects against EC apoptosis elicited
by LPS in a murine model of endotoxin shock (116).
The mechanism by which LPS/Tlr-4 signaling activates caspases remains unknown. A proximal event in LPS signaling is MyD88 recruitment of IRAK through reciprocal binding of the DD regions on each of these molecules (48). One possibility is that FADD or a FADD-like protein may bind either MyD88 and/or IRAK-1 through homophilic DD-DD interactions (Fig. 2). By a mechanism similar to that reported for other death receptor signaling pathways, including those involving Fas and TNFR1 (11), FADD, which contains an additional DED binding domain, may recruit procaspase-8 via DED-DED interactions to the Tlr-4 signaling complex. The induced proximity of procaspase-8 molecules could lead to respective activation and the onset of apoptosis. In support of this hypothesis, overexpressed MyD88 has been reported to bind FADD via DD-mediated interactions (5, 92). Whether this interaction occurs when these proteins are expressed at physiological levels remains unknown.
| |
ANTIAPOPTOTIC SIGNALING AND ROLE OF FLICE-LIKE INHIBITOR PROTEIN |
|---|
B ACTIVATION...
|
|---|
FLICE-like inhibitor protein (FLIP) is an antiapoptotic protein with significant homology to caspase-8 (FLICE) (106). In FLIP, a substitution of two amino acids that corresponds to the caspase-8 catalytic site renders it catalytically inactive. The role of FLIP in the inhibition of Fas death signaling has been well elucidated. On Fas ligand binding, the adapter protein FADD is recruited to the Fas receptor via an interaction between the DD of each protein (199). FLIP and procaspase-8 are, in turn, recruited to FADD via DED protein-protein interactions contained within all three proteins (Fig. 1). It has been proposed that FLIP can inhibit activation of upstream initiator caspases, including caspase-8, by competitively binding to FADD and blocking assembly of a functional death signaling complex (188). Alternatively, FADD recruitment of FLIP, which lacks intrinsic proteolytic activity, may prevent transactivation of procaspase-8 (182). A role for FLIP in inhibiting TNFR1-elicited apoptosis through a similar mechanism has been reported (97).
Recent studies indicate that FLIP protects against LPS-induced apoptosis. First, decreased expression of FLIP parallels in both a dose- and time-dependent manner with EC sensitization to LPS-induced apoptosis (21, 60). Second, inhibition of proteasome-mediated FLIP degradation protects against LPS-induced apoptosis (21). Third, specific inhibition of FLIP expression with antisense oligonucleotides sensitizes ECs to LPS-induced apoptosis (21). Finally, overexpression of FLIP protects against LPS-evoked apoptosis (60). Together, these data indicate that FLIP confers protection against LPS-induced apoptosis. The mechanism of protection remains unknown. A previous report has implicated a role for FADD in mediating LPS-induced apoptosis (36). One may speculate that Tlr-4 recruitment of FADD or a FADD-like molecule leads to the recruitment of upstream initiator caspases and subsequent caspase activation. In a manner analogous to the Fas and TNFR1 pathways, FLIP recruitment to FADD would be expected to prevent caspase recruitment and/or activation and block the onset of apoptosis (Fig. 2).
Both LPS and TNF-
directly induce EC apoptosis in bovine
ECs. Human EC sensitization to apoptosis elicited by either of
these agents is dependent on the absence of de novo gene expression (164). EC sensitization is conferred by either
cycloheximide or actinomycin D, protein or mRNA synthesis inhibitors,
respectively. The requisite inhibition of gene expression required for
sensitization suggests that either a constitutively expressed protein
with a relatively short half-life or an inducible protein confers EC resistance to LPS- and TNF-
-induced apoptosis. LPS and
TNF-
each upregulate the expression of several antiapoptotic
proteins, including A1 (99, 112, 113), A20 (58, 99,
220), and cellular inhibitor of apoptosis proteins
(cIAPs) (190) in an NF-
B-dependent manner.
Blocking NF-
B activation sensitizes ECs to direct TNF-
-induced apoptosis and correlates with the inhibition of expression of these NF-
B-inducible antiapoptotic proteins (221).
In contrast, inhibition of NF-
B activation fails to sensitize ECs to
direct LPS-induced apoptosis, suggesting a distinct mechanism
by which inhibition of new gene expression sensitizes ECs to LPS- and
TNF-
-induced apoptosis. Recently, inhibition of
proteasome-mediated protein degradation was shown to protect against
LPS-induced apoptosis in the presence of cycloheximide
(21). That LPS-evoked apoptosis was prevented in
the absence of new gene expression precludes a role for an inducible
gene product. Interestingly, inhibition of de novo protein synthesis
leads to a marked decrease in the expression of FLIP, a constitutively
expressed protein implicated in conferring resistance to LPS-induced
apoptosis (21, 60). This suggests that
cycloheximide sensitization of human ECs to LPS-induced
apoptosis is mediated by both the inhibition of de novo
synthesis of FLIP and the rapid degradation of preexisting FLIP
molecules via the proteasome (Fig. 3).
|
Although cycloheximide or actinomycin D have been used as tools to
investigate sensitization to LPS-induced apoptosis by
inhibiting de novo gene expression, it is difficult to imagine a
scenario in which these agents would be present in vivo with LPS.
However, naturally occurring agents that inhibit protein synthesis may be expected to confer similar sensitivity. Shiga toxin and Shiga-like toxin-1 (SLT-1), produced by Shigella dysenteriae serotype 1 and certain strains of E. coli, respectively, inhibit
protein synthesis (141, 162, 213). Both of these toxins
have been implicated in the pathogenesis of hemolytic uremic syndrome
and its attendant EC injury (158, 167, 170). Shiga toxin
and SLT-1 inhibit protein synthesis by cleaving a specific bond in the
28S rRNA component of the 60S ribosomal subunit, resulting in the
release of a single adenine base and the inhibition of aminoacyl tRNA
binding to the ribosome (79, 158). Recent findings suggest
that SLT-1 sensitizes human ECs to LPS-induced apoptosis by
virtue of its protein synthesis inhibitory properties (60)
(Fig. 3). SLT-1 inhibition of FLIP expression sensitizes ECs to
LPS-induced apoptosis in both a dose- and time-dependent
manner. Furthermore, sustained expression of FLIP in the presence of
SLT-1 abrogates SLT-induced sensitization of human ECs to LPS-induced
apoptosis. Thus strains of E. coli that produce
SLT-1 have all the necessary components to elicit EC apoptosis,
namely a sensitizing agent, SLT-1, and an inducer of apoptosis,
LPS. Other naturally occurring agents reported to have protein
synthesis inhibitory properties include Pseudomonas aeruginosa toxin (101), interferons
(39), and TNF-
(129). The role of these
agents in sensitizing ECs to LPS-induced apoptosis during mixed
infection and inflammation should be considered. Furthermore, patients
in severe sepsis often are in a profound catabolic state
(85). Whether this influences the ability of cells to
maintain adequate levels of antiapoptotic proteins while being
exposed to LPS and/or apoptotic-inducing cytokines remains unknown.
| |
ROLE OF BCL-2 FAMILY MEMBERS |
|---|
B ACTIVATION...
|
|---|
Cell commitment to apoptosis is governed by both pro- and antiapoptotic signaling pathways. The Bcl-2 family of proteins plays a central role in mediating these two opposed pathways (2, 8). The Bcl-2 family is composed of both pro- and antiapoptotic members. A key mechanism by which these proteins mediate apoptosis is through the regulation of cytochrome c release from the mitochondrion (2, 8). Cytochrome c is a cofactor in the activation of caspase-9, the latter of which activates downstream effector caspases, including caspase-3. Proapoptotic and antiapoptotic members of the Bcl-2 family facilitate and restrict, respectively, cytochrome c release from the mitochondrion. The various Bcl-2 family members are characterized by the presence of one or more of four distinct Bcl-2 homology (BH) domains that facilitate protein-protein interactions. Antiapoptotic Bcl-2 family members exert their effect by forming heterodimers via these BH domains with the proapoptotic members, thereby impairing the ability of the proapoptotic members to induce cytochrome c release.
Several lines of evidence suggest that members of the Bcl-2 family mediate sepsis- and LPS-induced apoptosis. First, in a murine model of sepsis induced by cecal ligation and puncture, enhanced endothelial apoptosis parallels a decrement in the expression of the antiapoptotic protein Bcl-2 (217). Second, LPS upregulates EC expression of the proapoptotic Bcl-2 family members Bax (81, 118, 152), Bad (147), and Bak (147). LPS has also been reported to upregulate expression of A1 (99) and to downregulate levels of Bcl-2 (81, 118) and Bcl-xL (147), all of which are antiapoptotic members of the Bcl-2 family. Third, vascular endothelial growth factor (VEGF) (152) and other agents (81, 118) that inhibit LPS-induced upregulation of proapoptotic Bcl-2 family members protect against LPS-induced EC apoptosis. Inhibition of LPS-induced downregulation of antiapoptotic Bcl-2 homologs similarly confers protection (81, 118). Finally, overexpression of A1 or Bcl-xL protects against LPS-induced EC apoptosis (99). Together, these data implicate a role for Bcl-2 homologs in determining EC fate after LPS exposure.
| |
ROLE OF NITRIC OXIDE |
|---|
B ACTIVATION...
|
|---|
Nitric oxide (NO) is a biological messenger molecule with profound
influence on the vasculature. NO has been implicated in the regulation
of vasomotor tone, inhibition of platelet aggregation and leukocyte
adhesion to ECs, activation of transcription factors leading to new
gene expression, and the regulation of apoptosis (140). In ECs, NO is generated by both a constitutive
nitric oxide synthase and an inducible nitric oxide synthase (iNOS)
(168). LPS upregulates the expression of iNOS and the
production of NO in ECs (130, 156, 175, 179). These EC
responses are further potentiated by LPS coadministration with other
proinflammatory cytokines, including interferon-
(70,
208).
NO has been reported to inhibit LPS-induced EC apoptosis. ECs exposed to the NO donors S-nitroso-N-acetylpenicillamine or (Z)-1-[2-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate demonstrate reduced sensitivity to LPS-induced apoptosis (51, 196). Enhanced production of endogenous NO after overexpression of iNOS has been similarly shown to block LPS-elicited EC apoptosis (31, 200). The mechanism by which NO inhibits LPS-induced EC apoptosis remains unclear. Increased levels of NO have been reported to suppress LPS-induced activation of caspase-3 in ECs (31, 200). In vitro studies using purified human recombinant caspases have shown that NO reversibly inhibits seven distinct members of the caspase family, including caspase-3, through direct S-nitrosylation of a cysteine residue required for caspase catalytic activity (55, 119, 136). Thus the protective effect of NO on LPS-exposed ECs may be mediated by NO inactivation of caspases.
In contrast to a cytoprotective role for NO, other studies have reported that suppression of iNOS induction and inhibitors of NOS activity protect against LPS-elicited EC injury (87, 160, 180), suggesting that NO promotes LPS-induced apoptosis. The differential effects of NO on mediating LPS-induced EC apoptosis appear to be dose dependent. DeMeester et al. (51) have reported that moderate levels of NO confer protection, whereas higher concentrations of NO enhance LPS-induced EC apoptosis. Although the mechanism by which NO potentiates LPS-elicited apoptosis remains unknown, high concentrations of NO can inhibit protein synthesis (214). Because protein synthesis inhibition has been established to sensitize human ECs to LPS-induced apoptosis by inhibiting the expression of the antiapoptotic protein FLIP (21), one may speculate that high levels of NO may result in decreased expression of this cytoprotective protein. Alternatively, increased production of NO after LPS induction of iNOS may result in NO reaction with superoxide anion to form peroxynitrite. EC injury resulting from the generation of peroxynitrite, a potent oxidizer, may synergistically enhance the EC apoptosis elicited by LPS. Further studies are needed to clarify the potential proapoptotic role of NO in mediating LPS-induced EC apoptosis.
| |
ROLE OF REACTIVE OXYGEN SPECIES |
|---|
B ACTIVATION...
|
|---|
In response to stress and/or injury, ECs generate reactive oxygen species (ROS) and nitrogen intermediates (165). Generation of these free radicals can lead to an alteration in the balance of the pro- and antioxidant states, resulting in oxidative stress and potential cell injury (32). Increases in ROS or depletion of antioxidants can lead to apoptosis in a variety of cell types, including ECs (54, 68, 125, 134, 135). LPS elicits an increase in the generation of EC ROS (28, 117, 118, 177, 216). Interestingly, LPS-induced generation of ROS correlates with the onset of apoptosis, and agents that inhibit the formation of ROS confer protection against LPS-evoked EC apoptosis (1, 28, 117, 118, 150). Although the mechanism by which ROS contribute to LPS-induced EC apoptosis remains unknown, ROS have been shown to increase mitochondrial membrane permeability, resulting in the release of proapoptotic factors (10, 29, 63, 195, 224). One of these factors is cytochrome c, a cofactor that drives the assembly of the caspase-9-activating apoptosome (169). Furthermore, ROS have been reported to increase both the expression and activation of caspase-8 (151, 224) as well as the activation of caspase-3 (10, 224).
In addition to the generation of ROS, LPS induces the upregulation of manganese superoxide dismutase (MnSOD) (150, 206) and copper, zinc superoxide dismutase (Cu,Zn-SOD) (123). SODs serve as an important component of the EC antioxidant defense system by accelerating the conversion of superoxide to H2O2. Although both superoxide and H2O2 are potent oxidants, the latter reacts much more slowly with cellular substrates (165). LPS-induced upregulation of MnSOD is mediated, in part, by increased generation of ROS, thus SOD may act to counterbalance the proapoptotic effects of ROS (150). Consistent with this hypothesis, upregulation of SOD has been reported to inhibit apoptosis in a variety of cell types, including ECs (56, 120, 215). Furthermore, overexpression of Cu,Zn-SOD protects against reactive oxygen intermediates (ROI)-elicited induction of caspase-8, activation of caspase-3, and mitochondrial release of cytochrome c after ischemia (151, 191). Finally, overexpression of MnSOD decreases the sensitivity of ECs to LPS-induced apoptosis (31). Clearly, one possible mechanism by which SOD confers protection is by restricting the accumulation of potent oxidizing molecules that may promote apoptosis. Another mechanism may be through SOD's role in accelerating the conversion of superoxide to H2O2. Although H2O2 has been reported to induce apoptosis in some systems, others have reported that high levels of H2O2 inhibit caspase activation and apoptosis (25, 83, 131). Whether LPS-induced upregulation of SODs generates H2O2 at a level that will promote or inhibit LPS-evoked EC apoptosis remains unknown.
The role of ROS in mediating apoptosis is further complicated
by ROS activation of signal transduction pathways that inhibit EC
apoptosis (105, 165). ROS-activated pathways
include: 1) the mitogen-activated protein kinases, which
have been implicated in VEGF-induced EC survival (78),
2) Akt, a kinase involved in phosphoinositide 3-kinase
signaling that is linked to shear stress and VEGF-induced survival
(53, 66, 71), and 3) NF-
B, a transcription
factor that upregulates the antiapoptotic proteins A1, A20, and
IAPs (44, 99, 113, 183, 190, 209, 221). Whether
LPS-induced generation of ROS activates these pathways resulting in
resistance to LPS proapoptotic signaling remains unknown.
| |
ROLE OF APOPTOTIC SIGNALING MOLECULES IN MEDIATING LPS-INDUCED EC ACTIVATION |
|---|
B ACTIVATION...
|
|---|
A key marker of EC activation is NF-
B activation and nuclear
translocation, a requisite event for many EC responses, including increased expression of cytokines and adhesion molecules. Recent studies demonstrate that several signaling molecules originally described as mediators of apoptosis also contribute to the
regulation of NF-
B activation (33, 98, 102, 103, 114, 184,
207). Evidence supporting this dual function for apoptotic
signaling molecules includes: 1) enhanced caspase activation
inhibits TNF-
-induced EC NF-
B nuclear translocation and VCAM-1
expression (157), 2) transient overexpression
of FADD, FLIP, or caspase-8 augments basal levels of NF-
B activation
(33, 98, 114), 3) overexpression of FADD
upregulates monocyte chemotactic protein-1 and IL-8 expression, the
transcription of which is mediated by NF-
B (184), and
4) FADD-deficient Jurkat cells display impaired
activation of NF-
B after exposure to TNF-
, TNF-related
apoptosis-inducing ligand (TRAIL), or Fas (207).
Similarly, evidence exists that apoptotic signaling molecules also
influence the ability of LPS to induce NF-
B activation. LPS-insensitive human embryonic kidney 293T cells are sensitized to
LPS-induced NF-
B activation on induced expression of Nod1, a
cytoplasmic protein with structural homology to the apoptosis regulator, Apaf-1 (104). Furthermore, overexpression of
Bcl-2 and Bcl-xL inhibits LPS-induced NF-
B activation
and NF-
B-dependent gene expression in ECs (13). This
inhibition of NF-
B activation corresponds with Bcl-2-mediated
inhibition of I
B
degradation. The mechanism by which Bcl family
members inhibit LPS-induced NF-
B activation remains unknown.
Identification of a caspase consensus site within I
B
has led to
speculation that caspases may be involved in its proteolysis
(13). Thus Bcl-mediated inhibition of caspase activation
may be expected to block I
B
degradation and subsequent activation
of NF-
B.
Further evidence for cross talk between apoptotic signaling
molecules and LPS-induced NF-
B activation was provided in a recent report demonstrating that the proapoptotic adapter protein FADD downregulates LPS-induced NF-
B activation (20). In that
study, overexpression of either full-length FADD or the DD of FADD in ECs blocked LPS-induced NF-
B activation. Furthermore, mouse embryo fibroblasts (MEF) derived from FADD
/
mice embryos displayed enhanced NF-
B activity relative to FADD+/+ MEF after LPS treatment. The production of IL-6 and KC, two NF-
B-dependent gene
products, were similarly enhanced in FADD
/
MEF after LPS exposure
relative to FADD+/+ MEF. Reconstitution of FADD in FADD
/
MEF
abrogated the enhanced NF-
B activation and IL-6 and KC production
elicited by LPS. Together, these data suggest that changes in FADD
expression can affect LPS-induced activation of NF-
B. The role of
FADD in inhibiting this event under basal conditions remains unclear as LPS activates EC NF-
B in the presence of physiological levels of FADD.
The mechanism by which FADD downregulates NF-
B-dependent gene
expression remains unknown. However, the enhanced NF-