Research Service, Stratton Veterans Affairs Medical Center;
and the Center for Cardiovascular Science, The Albany Medical
College, Albany, New York 12208
The intracellular
serine/threonine kinase protein kinase C (PKC) has an important role in
the genesis of pulmonary edema. This review discusses the PKC-mediated
mechanisms that participate in the pulmonary endothelial response to
agents involved in lung injury characteristic of the respiratory
distress syndrome. Thus the paradigms of PKC-induced lung injury are
discussed within the context of pulmonary transvascular fluid exchange.
We focus on the signal transduction pathways that are modulated by PKC and their effect on lung endothelial permeability. Specifically, 
-thrombin, tumor necrosis factor (TNF)-, and reactive oxygen species are discussed because of their well-established roles in both
human and experimental lung injury. We conclude that PKC, most likely
PKC-, is a primary supporter for lung endothelial injury in response
to -thrombin, TNF-, and reactive oxygen species.
 |
INTRODUCTION |
THE VESSEL INTIMA AND
CAPILLARIES of the pulmonary and systemic circulation are lined
by a monolayer of endothelial cells that serve to control and restrict
the luminal to abluminal movement of water and protein. Specifically,
the pulmonary endothelium constitutes the interface between the blood
and extravascular tissue of the lungs. This homeostatic barrier
function of the endothelium is ultimately maintained by the dynamic
regulation of the endothelial cell shape, endothelial cell-to-cell
adherence, and endothelial-extracellular matrix adherence
(101). The homeostasis of the lung, as indicated by the
ventilation-to-perfusion ratio, exchange of blood gases, metabolism,
and transit of blood from the right to left side of the circulation, is
invariably influenced by changes in the endothelial barrier function
(101, 110, 112). Thus compromise of systemic and pulmonary
vascular endothelial function is a main component of the
pathophysiology of inflammation and its associated pernicious syndrome
sepsis (18, 44, 112, 116, 177). In the lungs, the
endothelial response is characterized by a decrease in its restrictive
barrier function, resulting in an increase in endothelial permeability
and transvascular fluid and protein flux into the interstitial space
(i.e., edema), which ultimately contributes to the pathogenesis of
multiple organ failure (MOF) associated with sepsis (18, 44, 116,
177).
| |
PULMONARY EDEMA |
The movement of water and macromolecules into the lung
interstitium, airways, and alveoli is characteristic of acute lung injury, which is pathogenetic for the respiratory distress syndrome (RDS) (110, 112, 181). The pulmonary endothelium has a
paramount role in the pathophysiology of acute lung injury because an
alteration in its function as a restrictive barrier contributes to the
changes in the forces that modulate edema (101, 110, 112, 140,
181). The forces that govern the movement of water and protein
leading to edema formation are described by the Starling equation for fluid exchange (172). A brief review of the Starling
equation for fluid exchange is presented here to facilitate
understanding of the role of PKC in lung edema.
where Kfc is the capillary
filtration coefficient, Pmv is the microvascular
hydrostatic pressure, Pi is the interstitial hydrostatic
pressure, 
is the protein reflection coefficient, 
i
is the interstitial oncotic capillary pressure, and mv
is the microvascular oncotic pressure.
In human studies and in experimental models of acute lung injury,
pulmonary edema is induced by increases in Pmv and
Kfc and by a decrease in (18, 44, 76,
81, 83, 110, 112, 141, 172). During RDS, the compromised
pulmonary endothelium maintains, at least in part, a decrease in
protein selectivity (i.e., decrease in ), an increase in
transvascular fluid flux (i.e., increase in
Kfc), and an altered metabolism of inflammatory mediators [e.g., angiotensin I, bradykinin, endothelin, prostacyclin, thromboxane A2, superoxide
( · O
) and nitric oxide
( · NO)] (Fig. 1)
(47, 49, 54, 112, 113, 152, 172, 182). The increase in
Pmv is caused by a decrease in the ratio of the
precapillary to postcapillary resistance arising from the altered
levels of a number of endothelium-dependent mediators that have been
implicated in RDS, such as angiotensin II, endothelin, thromboxane
A2, prostacyclin, reactive nitrogen species [e.g., · NO, peroxynitrite (ONOO
)], and
reactive oxygen species [e.g., hydrogen peroxide
(H2O2,), · O, hydroxyl radical
( · OH)] (Fig. 1) (110, 112). The
increase in Kfc and the decrease in are due,
in part, to the response to inflammatory mediators such as · NO (77),
H2O2 (80), thrombin (83,
101, 110), and tumor necrosis factor (TNF)- (Fig. 1)
(81, 159). Isolated lung studies indicate that the
increase in Kfc is primarily due to an
increase in permeability to water, in addition to heterogeneous
increases in surface area. In lung injury that progresses toward
alveolar flooding, the extra-alveolar and alveolar epithelium also
exhibits a decrease in barrier function and an increased generation of inflammatory mediators [e.g., inducible nitric oxide synthase (iNOS)
· NO].
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 1.
The cell types and mediators involved in the protein kinase C (PKC)
paradigm of lung injury. PMN, neutrophils; · NO,
nitric oxide; ONOO, peroxynitrite; · OH,
hydroxyl radical.
|
|
| |
DETERMINATION OF ENDOTHELIAL BARRIER FUNCTION |
The early in vivo study of transvascular fluid flux used
time-dependent lung weight measurements during different
Pmv and mv (172). Then,
with the advent of the chronic sheep lung lymph fistula model using the
online determination of fluid and protein flux, Ppm,
pm, and i permitted the estimation of (110, 172). In addition, by using the isolated lung in
situ and ex vivo, investigators were able to determine
Kfc, in addition to Pmv and .
Thus the forces that control pulmonary edema were becoming understood
but with the caveat that each technique had its limitation. Ultimately, the pursuit of a focus on the endothelium and the requirement for
specificity of experimental manipulation on a single cell type prompted
the evaluation of the transendothelial movement of molecules within an
isolated endothelial cell monolayer, in the absence of underlying
tissues. The successful isolation and culture of endothelial cells from
lung macrovessels (e.g., artery and vein) and microvessels enabled the
study of the pulmonary endothelial cell's metabolism and response to
inflammatory mediators that ultimately determine the value of
Pmv, Kfc, and (21, 38). Thus the barrier function of the endothelium is now
characterized by the microscopic visualization of tracer molecules
(e.g., horseradish peroxidase, see Ref. 180), measurement
of transendothelial clearance of molecules (e.g., albumin and dextran),
and transendothelial electrical resistance (TEER) (29, 38, 42,
86, 107, 115, 163, 173, 180).
| |
THE BASICS OF PKC |
PKC is a family of serine/threonine kinases characterized by
at least eleven different isotypes. PKC isotypes are differentially regulated by calcium (Ca2+), diacylglycerol, and
phospholipids and differ in structure, expression, intracellular
localization, substrate utilization, and mechanisms of activation
(31, 129, 130, 192). PKC is composed of four conserved
(C1-C4) and five variable (V1-V5)
domains (Fig. 2). C1 and C2 constitute
the regulatory domain and contain binding sites for phospholipids
(e.g., phosphatidylserine), Ca2+, diacylglycerol, and
phorbol esters (Fig. 2). The C3 and C4 regions contain the catalytic
domain that has binding sites for ATP and different PKC substrates
(Fig. 2). The conserved region of the regulatory domain, within
residues 19-36, has structural features of a pseudosubstrate;
therefore, it maintains PKC in the inactive form during the absence of
phospholipid activators. The activation of PKC begins with the release
of membrane phospholipids in response to phospholipase activity [e.g.,
phospholipase C (PLC)]. The phospholipids interact with the C1 and C2
domains, which provides the free energy required for the dissociation
of the NH2-terminal pseudosubstrate from the active site,
which allows substrate binding (72) (Fig. 2). The C1 and
C2 domains each have their own determinants for membrane recognition,
and the C1 domain is present in most PKC isotypes (82).
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 2.
A
molecular model for PKC isotypes and PKC activation.
C1-C4, conserved domains;
V1-V5, variable domains; DAG,
diacylglycerol; PSS, NH2-terminal pseudosubstrate;
PS, permeability-surface area product; S, substrate.
|
|
PKC isotypes are involved in signal transduction pathways that govern a
wide range of physiological processes, such as differentiation, proliferation, gene expression, brain function, membrane transport, and
the organization of cytoskeletal and extracellular matrix proteins
(22). The PKC isotypes are subdivided into three groups: the classical, novel, and atypical. This subdivision is based on the
structural and functional differences in the conserved domains
C1-C4 (Fig. 2) (128). The classical PKC-,
PKC-
1/2, and PKC-
isotypes are Ca2+ and
diacylglycerol dependent. These PKC isotypes have the conserved diacylglycerol-binding C1 domain and the Ca2+-binding C2
domain. The C1 domains consist of a tandem C1A and C1B arrangement that
can bind the endogenous diacylglycerol and exogenous phorbol esters
(Fig. 2) (167). The novel PKC-
, PKC-
, PKC-
,
PKC-
, and PKC-µ isotypes contain C2 domains that lack Ca2+-binding ability but still retain functional C1A and
C1B domains that can bind the endogenous diacylglycerol and exogenous
phorbol esters (Fig. 2). The atypical PKC-
, PKC-
, and PKC-
isotypes (31, 99, 123, 129, 137) lack a functional C2
domain and contain a single C1 domain that lacks the ability to bind
diacylglycerol and phorbol esters (25, 149). Therefore,
the mechanism of activation of the atypical PKC isotypes is both
Ca2+ and diacylglycerol independent and involves other
lipid-dependent pathways (123). Thus for example the
phorbol ester 2-O-tetradecanoylphorbol-3-acetate or the
lipid diacylglycerol would activate the classical and novel PKC
isotypes but not the atypical PKC isotypes. The activation of the novel
PKC isotypes will persist in the presence of EDTA, but the classic PKC
isotypes would not exhibit activity (22, 151).
In addition to the classic activation mechanisms indicated above, other
PKC activation mechanisms have been proposed. Slater et al.
(167) have shown that PKC- activation is also dependent on lack of a C1-C2 domain interaction, corresponding to a
transition of PKC- from a closed inactive state to an open active
state (Fig. 2). They showed that PKC- isotype bound specifically and with high affinity to an C1A-C1B fusion protein of PKC-. The C1A-C1B domain activated the isozyme in a phorbol ester- and diacylglycerol-dependent manner comparable with activation
resulting from membrane-phosphatidylserine association. Interestingly,
the C1A-C1B domain also activated the classical PKC-1/2, and
PKC- isotypes, but not the novel PKC- or PKC- isotypes that
were each activated by their own C1 domains. PKC-, PKC-1/2, and
PKC- isotypes were unaffected by the C1 domain of the PKC-
isotype and only slightly activated by that of PKC- isotypes. Thus
the activation mechanism of the novel PKC isotypes may be similar to
that of the classic isotypes. PKC- isotype activity was unaffected by its own C1 domain and those of the other PKC isotypes. Another key
determinant of PKC activity is the phosphorylation of the PKC molecule,
its intracellular localization, and proteolytic degradation.
Phosphorylation is controlled by PKC-mediated autophosphorylation and
the phosphorylation mediated by other kinases such as
3-phosphoinositide-dependent kinase-1 and tyrosine kinases (40,
94).
The distribution of PKC activity is regulated by its direct interaction
with accessory proteins (e.g., receptor for activated C protein) that
target the movement of the PKC molecule to different intracellular
compartments, which confers selectivity by associating individual
isotypes with specific substrates (149, 156, 169). PKC
activity is also determined by its degradation. The literature indicates that the calcium-lipid-dependent protease calpain-µ can
degrade PKC to a catalytically active PKM by cleaving off the
regulatory domain (43, 46, 154, 161). PKM may be a
constitutively active enzyme that mediates long-term phosphorylation
activity of PKC. Yet the regulation of PKC activity is maintained
because of further degradation into inactive degradation products by
calpain-m (43, 46, 154, 161).
Identification and characterization of the PKC isotypes were restricted
by the availability of PKC isotype-specific pharmacological inhibitors
and activators. However, recent technology offers better isotype
specificity, which includes oligonucleotide antisense (AS)-induced
inhibition of expression, expression of wild-type and dominant-negative
PKC vectors, PKC isotype knockout mice, and peptide fragments to either
inhibit or promote translocation of PKC isotypes to specific anchoring
proteins (9, 12, 22, 184, 189).
| |
PKC IN LUNG INJURY |
PKC is implicated in many cellular responses associated with lung
injury, including endothelial permeability (104, 107, 165,
164), cell contraction (100, 196), migration
(40), proliferation (40, 197),
apoptosis (162), mucous secretion (84), gene expression (147), and the
organization of cytoskeletal and extracellular matrix proteins
(4, 195, 196). Throughout the last five decades,
investigation into the pathogenesis of the increased endothelial
permeability associated with RDS has indicated a role for many
mediators, such as cytokines (e.g., IL-1, TNF-) (52, 93,
146), growth factors [e.g., vascular endothelial growth factor
(VEGF)] (93), peptides (substance P, bradykinin)
(7, 150, 183), proteases (e.g., elastase) (19), complement activation (e.g., C5a) (125,
188), intravascular coagulation (e.g., thrombin) (59, 102,
103, 110, 170), reactive oxygen and nitrogen species (e.g.,
H2O2,
· O, · OH, · NO,
ONOO) (20, 80, 89, 164, 165), and lung
sequestration of neutrophils (95) (Fig. 1). The
intracellular signal pathways that cause an increase in endothelial
permeability are still not completely defined despite extensive study
of the effect of the above mediators of endothelial dysfunction.
Importantly, however, PKC is now known to be a necessary part in the
regulation of endothelial permeability and edema formation induced by
three known mediators of RDS: H2O2, thrombin,
and TNF- (7, 23, 50, 54, 87, 107, 164, 165). This
review highlights the role of PKC in the endothelial barrier alteration
induced by H2O2, thrombin, and TNF-. We
discuss the role of PKC in the vascular permeability of acute lung
injury as assessed in in vivo, ex vivo, and in vitro studies using
cultured endothelial cells of pulmonary origin.
| |
PKC IN PHORBOL ESTER-INDUCED LUNG INJURY |
The involvement of PKC in lung injury began with studies using the
phorbol esters 12-O-tetradecanoylphorbol-13-acetate (TPA) and phorbol 12-myristate 13-acetate (PMA). TPA and PMA bind strongly to
the C1 diacylglycerol regulatory domain, thereby activating the classic
and novel PKC isotypes (Fig. 2) (130, 151). Conversely, long-term activation of the PKC by phorbol esters results in
degradation of PKC, which depletes the PKC activity (130,
151). Thus the use of PMA and TPA holds the caveat that each
isotype of PKC is differentially activated and/or depleted, which will
give an indication only to the class of PKC involved in the response
(9, 15, 99, 117, 151). Despite these disadvantages,
phorbol esters are still used as a "screen" for indicating a role
for PKC function in lung endothelial injury.
In studies using whole animals, PMA induces a route-, time-, and
dose-dependent pathophysiology (e.g., edema, increased
vascular/endothelial permeability, hypertension, and hypoxia) similar
to acute lung injury and RDS (179). Alveolar instillation
of PMA results in alveolar edema supposedly due to intense activation
of alveolar macrophages and pulmonary epithelium in anesthetized rats
and sheep (65). Vascular PMA infusion causes an early
decrease in endothelial cell angiotensin converting-enzyme activity
before and independently of changes in capillary surface area and edema (45, 118). The early functional change induced by PMA is
time dependent because it is followed by frank pulmonary edema due to
activation of the many cell types in the lung's vasculature (45,
118).
The presence of many cell types in the lung and edema formation in vivo
confounded accurate study of endothelial barrier function; therefore,
investigators began focusing on ex vivo techniques and cell culture. In
isolated perfused guinea pig lungs without polymorphonuclear leukocytes
(PMN), PMA increased pulmonary arterial pressure (Ppa),
pulmonary capillary pressure (Ppc), and lung weight; decreased the ratio of arterial-to-venous vascular resistance; but had
no effect on Kfc (76). A caveat to
this study is that the severe increase in vascular resistance did not
undermine the measurement in Kfc
(76). Thus experiments using cultured endothelial cells
from various species and vascular sites were performed (63, 66,
160, 185). The results indicated that, in bovine pulmonary arterial endothelial cell monolayers, short-term treatment with PMA
(108-106 M) increased endothelial
permeability to albumin (107). The negative controls
4--phorbol-didecanoate and 1-mono-oleoyl glycerol, which did
not activate PKC, had no effect on endothelial permeability to albumin
(107). The early use of PKC inhibitors indicated that 1-(5-isoquinolinylsulfonyl)-2-methylpiperazine (H7) (68)
reduced the PMA-induced increase in endothelial permeability,
whereas the isoquinoline
N-(2-guanidinoethyl)-5-isoquinolinesulfonamide hydrochloride
(HA1004) (68), which had no affect on PKC activity, also
had no effect on endothelial permeability in response to PMA.
| |
ISOTYPES OF PKC IN PHORBOL ESTER-INDUCED LUNG INJURY |
The specific PKC isotypes involved in the PMA-induced barrier
dysfunction have also been studied. In the human dermal microvascular endothelial cell line HMEC-1, the overexpression of PKC-1 augmented the PMA-induced increase in endothelial permeability. PMA stimulated the translocation of PKC-1 from the cytosol to the membrane, indicating PKC-1 activation (126). In another study, a
decrease in PKC-1 expression reduced the effect of PMA on barrier
function (187). Interestingly, downregulation of PKC with
PMA can delay the normalization of endothelial permeability that
follows the altered barrier function in response to thrombin (see
PKC IN THROMBIN-INDUCED LUNG ENDOTHELIAL INJURY),
indicating that PKC isotypes have disparate roles in changing
endothelial permeability that are dependent on the agonist used to
induce the inflammatory response (105).
| |
LEUKOCYTES IN PKC-MEDIATED LUNG INJURY IN RESPONSE TO PHORBOL
ESTERS |
There are many potential mechanisms for PMA-induced, PKC-dependent
lung injury due to the myriad of signal transduction pathways affected
by PKC. A role for PMN is noted because in studies using isolated rat
lungs not depleted of PMN, PMA added to the perfusate increased
microvascular permeability as expressed by increases in
Ppa, Kfc, lung weight gain, lung
wt/body wt ratio, and the protein concentration of the bronchoalveolar
lavage fluid (27). The PMA-induced effect on endothelial
permeability was attributed to activated PMN and their release of
inflammatory mediators, including reactive oxygen species and proteases
(Fig. 1) (5, 83, 134, 141). However, other leukocytes can
mediate acute lung injury, as observed in PMA instillation of
PMN-depleted lungs in the presence of mononuclear leukocytes
(142).
| |
REACTIVE OXYGEN SPECIES IN PKC-MEDIATED LUNG INJURY IN RESPONSE TO
PHORBOL ESTERS |
In animal studies, reactive oxygen species mediate, at least in
part, the response to PMA because the reactive oxygen species scavenger
dimethylthiourea (26) prevents pulmonary hypertension, and
the enzyme CuZn superoxide dismutase (SOD) (121)
attenuates the increase in vascular permeability (26).
Similarly, in cultured endothelial cells, PMA is used as a model for
reactive oxygen species-induced activation of nuclear transcription
factors such as NF-
B and activator protein (AP)-1 (73).
The importance of the effect of reactive oxygen species on NF-B and
AP-1 is dictated by DNA promoter-driven expression of many substances
that have a role in endothelial barrier dysfunction such as the
intercellular adhesion molecules (ICAM), prostaglandin E2
(PGE2), TNF-, and IL-8 (1, 49, 50, 97,
138). PMA causes depletion of glutathione in pulmonary arterial
endothelial cells that is inhibited by CuZn-SOD, supporting the notion
that reactive oxygen species participate in the effect of PKC
activation (144). Glutathione depletion markedly enhances
PMA-induced expression of ICAM and PGE2 in human umbilical
vein endothelial cells, which supports the notion of reactive oxygen
species-driven expression for mediators of inflammation. The PKC
inhibitor H7 prevents most of PMA-induced ICAM-1 expression, PGE2 production, and the effect of glutathione depletion.
The glutathione effect is also inhibited by the antioxidant quercetin (73). Thus PKC-induced glutathione depletion enhances
susceptibility of vascular endothelial cells to the effects of reactive
oxygen species generation and its downstream effect on gene expression (73). The increase in reactive oxygen species
concentration is probably due to the activation of NADPH oxidase,
because PKC phosphorylates p22phox and p47phox
(2, 10, 32, 34). It is well established that reactive oxygen species mediate endothelial barrier dysfunction indirectly via
downstream signal molecules and/or by direct effects on the cell cytoskeleton.
| |
INTRACELLULAR SIGNAL PATHWAYS IN PKC-MEDIATED LUNG INJURY IN
RESPONSE TO PHORBOL ESTERS |
The downstream signaling pathways by which PKC activation
increases endothelial permeability have been studied using cultured endothelial cells (185) (Fig.
3). PMA induces rapid phosphorylation of
the Rho-GDP guanine nucleotide dissociation inhibitor (GDI) in human
umbilical vein endothelial cells. Phosphorylated GDI permits
Rho-GDP/GTP exchange resulting in Rho-GTP as a co-actor for Rho kinase
(ROCK). ROCK is implicated in modulation of myosin-actin-ATPase activity (24, 108, 180) and therefore in cell contraction (Fig. 3). Bovine microvessel endothelial cells grown on a flexible substrate contract on addition of agents that cause increased endothelial permeability in vivo or in vitro, including
angiotensin II, thrombin, bradykinin, and the stable analog of
thromboxane A2 U-44069 (122). PKC depletion by
preincubation with PMA prevented the contraction by angiotensin II. The
inactive analog 4--phorbol 12,13-didecanoate did not inhibit
contraction, providing direct evidence that contraction of microvessel
endothelial cells may mediate the increase in endothelial permeability
in response to activation of PKC (122). Similarly,
PMA-induced depletion of PKC prevents thrombin-induced GDI
phosphorylation, Rho activation, and thrombin-induced decrease in TEER
(119), supporting a PKC GDI Rho-GDP/GTP Rho-GTP ROCK myosin-actin-ATPase pathway (Fig. 3). Evidence for
another signaling pathway indicates that PMA triggers a Ras-dependent
signal transduction in primary human umbilical vein endothelial cells
and in the permanent endothelial cell line ECV304 (119)
(Fig. 3). Selective inhibition of the mitogen-activated protein
kinase (MAPK) and extracellular signal-regulated kinases (ERK)
significantly attenuates the PMA-induced reduction in TEER, consistent
with PKC- and ERK-mediated endothelial cell barrier regulation
(185). PMA also produced a time-dependent increase in the
activity of Raf-1, a serine/threonine kinase known to activate MAPK
kinase (MEK), and increased the activity of Ras, which binds
and activates Raf-1. Inhibition of Ras completely abolished PMA-induced
Raf-1 activation, suggesting that the sequential activation of PKC Ras Raf-1 MEK ERK is also involved in endothelial barrier
regulation by PMA (185) (Fig. 3).
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 3.
PKC-mediated activation of signaling pathways that target
critical proteins that impact on endothelial barrier function. GDI,
guanine nucleotide dissociation inhibitor; ROCK, rho kinase; VE,
vascular endothelial; JAM, junctional adhesion molecule; FAK, focal
adhesion kinase.
|
|
| |
INTRACELLULAR TARGETS IN PKC-MEDIATED LUNG INJURY IN RESPONSE TO
PHORBOL ESTERS |
In bovine pulmonary artery endothelial cell monolayers, PMA and
-thrombin induce rapid and concentration-dependent activation and
translocation of PKC that are temporarily associated with agonist-mediated endothelial cell contraction and increased in endothelial permeability. PMA and -thrombin induce
phosphorylation of the cytoskeletal protein actinin, the
calmodulin-binding protein caldesmon77, and the
intermediate filament protein vimentin. The inhibition of PKC
prevents the -thrombin- and PMA-induced phosphorylation of the above
cytoskeletal proteins, attenuates the cell contraction, and reduces the
increase in endothelial permeability (57, 168). In human
umbilical vein endothelial cells, thrombin increases intracellular Ca2+ concentration
([Ca2+]i), endothelial permeability, and
activation of PKC- and causes alterations in vascular
endothelial (VE)-cadherin junctions (153). The
thrombin-induced alteration in VE-cadherin junctions occurred in
association with actin stress fiber formation and 20-kDa myosin light
chain (MLC20) phosphorylation. Inhibition of PKC prevented the disruption of VE-cadherin and the increase in endothelial permeability caused by the thrombin. This supports the notion that the
permeability response to thrombin is mediated by PKC-induced cell
contraction (via actin/myosin) activity and cell-cell adherence (via
VE-cadherins) (Fig. 3). However, in human umbilical vein endothelial
cells, thapsigargin-induced discontinuities in VE-cadherin junctions
occurred without formation of actin stress fibers and phosphorylation of the MLC20 (153).
The inhibition of PKC prevented the disruption of
VE-cadherin and the increase in endothelial permeability in response to
thapsigargin. These results suggest that PKC-mediated barrier
dysfunction occurs via MLC-dependent and -independent mechanisms that
depend on the agonist used (Fig. 3).
| |
PARACELLULAR TARGETS IN PKC-MEDIATED LUNG INJURY IN RESPONSE TO
PHORBOL ESTERS |
The increase in endothelial permeability that occurs in response
to PMA-induced PKC activation is associated with disruption of
endothelial cell monolayer integrity manifested as endothelial contraction and intercellular gap formation (63, 185). The endothelial transport of solutes and fluids occurs via transcellular and paracellular pathways. The paracellular pathways are regulated by
intercellular junctional proteins [i.e., junctional adhesion molecules
(JAM)] (11, 36, 37, 135), claudins (178),
occludins (127), and the transmembrane adhesion junction
proteins cadherins (e.g., VE-cadherin) (36, 37).
JAM regulates tight junctions and is also implicated in the processes
of neutrophil and monocyte transendothelial migration, because it is
constitutively expressed on circulating monocytes, neutrophils,
lymphocytes subsets, and platelets (36, 37, 135). In
epithelial cells from mucosal tissue sampled from patients with
inflammatory bowel disease, there was downregulation of occludin,
zonula occludens (ZO)-1, claudin-1, JAM, -catenin, and E-cadherin
primarily in regions of actively transmigrating PMN
(92). Thus adherent/migrating leukocytes can
directly modify junctional proteins, which may contribute to
endothelial barrier dysfunction. The vascular endothelial junction-associated molecule is prominently expressed on high endothelial venules and is present on the endothelia of other vessels
(135). Claudins are components of the tight junctional complex in endothelial cells. In endothelial cells, claudin-5 is
expressed more than claudin-1 and -2. A role for PKC in the expression
of tight junction proteins is noted in the choroid plexus, because PMA
increased immunoreactivity of claudin-1 and reduced immunoreactivity of
claudin-2 and -5 (98). The phosphorylation state of
occludin is thought to be important in both tight junction assembly and
regulation. Clarke et al. (28) have shown that PKC
activation leads to dephosphorylation of occludin and increases in
permeability in the epithelial cell line LLC-PK1. In epithelial cells,
PMA-induced PKC activation results in decreases in threonine phosphorylation of occludin, which correlated closely with the rapid
decreases in transepithelial electrical resistance, indicating a role
of a serine/threonine phosphatase in response to PKC activation (Fig.
3).
The adhesion of cells at adherence junctions is also achieved through
the calcium-dependent homotypic interaction of
cadherins (Fig. 3). The cadherins are associated with the cytoplasmic
catenins 120(cas)/p120(ctn) and the splice variant p100. Endothelial
cells have PKC-dependent and -independent pathways that regulate
the serine/threonine phosphorylation of p120/p100, further
demonstrating a connection between PKC and cadherins in
endothelial cells (148). In human dermal microvascular
endothelial cells exposed to hypoxia/aglycemia, increases in
endothelial permeability were associated with significant decreases in
the concentrations of occludin and cadherin. The hypoxia/aglycemia-mediated permeability changes and decreases in
junctional proteins were blocked by chelation of intracellular Ca2+ and by inhibition of PKC, PKG, and p38 MAPK. Thus the
altered permeability may occur through PKC-, PKG-, MAPK-, and
Ca2+-mediated dissociation of cadherin-actin and
occludin-actin junctional bonds (136). Angiopoietin-1, a
ligand for the endothelium-specific tyrosine kinase receptor Tie-2
(35), supports junctional localization of platelet
endothelial cell adhesion molecule-1 (PECAM-1) and decreases the
phosphorylation of PECAM-1 and VE-cadherin. Angiopoietin-1 induces
tightening of the endothelial junctions, thereby reducing endothelial
permeability. Angiopoietin-1 inhibits thrombin- and VEGF-induced
increases in endothelial permeability (56). Thus PKC-mediated phosphorylation of tight and adherence-junctional proteins can contribute to increased endothelial permeability (Fig. 3).
Cell-matrix adherence associated with focal adhesion plaques and
protease activity is also influenced by PKC-mediated events (Fig. 3).
Stimulation of coronary venular endothelial cells with PMA enhanced
tyrosine phosphorylation of paxillin and focal adhesion kinase
(p125FAK), suggesting a possible involvement of protein
tyrosine kinases and their associated focal adhesion plaques in the
control of PMA-induced endothelial barrier dysfunction (Fig. 3)
(193). PMA treatment of primary human umbilical vein
endothelial cells and the transformed endothelial cell line ECV304
induces increased expression (at the level of the promoter, mRNA, and
protein activity) of matrix metalloproteinase-9 (MMP-9) (Fig.
4). MMP-9 degrades native type IV
collagen and is implicated in barrier dysfunction, decreased
cell-matrix adhesion, and angiogensesis (60).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 4.
A model for PKC-mediated endothelial cell barrier
dysfunction during respiratory distress syndrome (RDS). Sepsis via
endotoxin causes the release of TNF- and
H2O2, primarily from macrophages and PMN, and
the generation of thrombin. The data indicate that PKC- is a major
isotype that promotes the vascular derangements associated with the
inflammatory mediators known to cause RDS. PKC- has a repressive
role at least with respect to -thrombin-induced endothelial injury.
TNF-, H2O2, and -thrombin activate
endothelial PKC that leads to at least 4 downstream pathways as
indicated: 1) PKC causes generation of reactive oxygen
species (ROS) and reactive nitrogen species (RNS), probably via the
activities of NADPH oxidase and nitric oxide synthase (NOS),
2) PKC indirectly/directly affects the phosphorylation of
cytoskeletal targets, 3) PKC modifies gene expression by
altering the activity of transcription factors such as NF-B and
activator protein (AP)-1, and 4) PKC can increase activity
of matrix metalloproteases such as MMP-9. The "positive-feedback"
loop results from increased expression of intercellular adhesion
molecules (ICAM), tissue factor (TF), and plasminogen activator
inhibitor (PAI) and decreased activity of plasminogen activator (PA).
The loop promotes leukocyte sequestration, the generation of
-thrombin and TNF-, and formation of ROS and RNS.
|
|
| |
PKC IN THROMBIN-INDUCED LUNG ENDOTHELIAL INJURY |
Thrombin is a serine protease that was initially described as a
procoagulant because it activates platelets and catalyzes the
conversion of fibrinogen to fibrin (51), yet thrombin also has direct effects on endothelial cells because of its specific binding
to the protease-activated receptor (PAR) on the endothelium (185). The thrombin receptor has seven transmembrane
domains composed of an extracellular NH2 terminus, three
extracellular loops, three intracellular loops, and an intracellular
COOH terminus. The thrombin receptor is coupled to the pertussis
toxin-sensitive and -insensitive heterotrimeric G proteins
Gi and Gq. PAR is activated by
thrombin-mediated cleavage of the NH2 terminus between
Arg41 and Ser42, generating a new
NH2 terminus beginning at Ser42 that functions
as a "tethered ligand" as a result of its binding to specific sites
on the PAR receptor (185).
The thrombin-PAR paradigm is implicated in endothelial injury and the
formation of pulmonary edema in diseases such as sepsis (70,
174), RDS (109, 110, 112, 139), and pulmonary
emboli (78, 109, 110, 112). Thrombin challenge in vivo
causes lung edema by increasing Pmv and i
with a decreased . In the isolated lung, studies confirm the in vivo
data and also indicate an increase in Kfc. In
morphological studies of the dog and sheep, thrombin produces focal
disruption of the microvascular endothelium associated with
intravascular fibrin and PMN sequestration and increased pulmonary
vascular permeability to proteins (110-112, 120). In
bovine pulmonary arterial endothelial cell monolayers, -thrombin
causes a dose-dependent increase in endothelial permeability to albumin (6, 7, 58, 59, 102, 145, 153). Interestingly,
-thrombin, which lacks the fibrinogen recognition site, also causes
a dose-dependent increase in endothelial permeability to albumin
(6, 39, 174). In bovine pulmonary arterial endothelial
cell monolayers, the -thrombin-active catalytic site is required for
the increase in transendothelial permeability to albumin
(7).
There are many postulated mechanisms for the thrombin-induced effect on
lung edema (111, 114). With regard to our discussion specific to endothelial function, thrombin causes an increase in
cytosolic [Ca2+]i, which is a critical
participant in intracellular signaling (102, 103). The
increase in [Ca2+]i is connected to important
downstream effects implicated in regulation of endothelial permeability
to protein such as: 1) phospholipase activation,
2) generation of prostaglandin, leukotrienes, reactive
oxygen species, and reactive nitrogen species, and 3) cell
contraction (Fig. 2). Specifically, thrombin activates endothelial cell
phosphatidylinositol-specific PLC that catalyzes the hydrolysis of
phosphatidylinositol 4,5-bisphosphate (PIP2).
PIP2 hydrolysis results in generation of the second
messengers inositol 1,4,5-triphosphate (IP3) and
diacylglycerol. 1,4,5-Triphosphate regulates intracellular [Ca2+]i by mobilizing the release of
Ca2+ from internal cellular stores. In addition, the
phosphorylated metabolite of IP3 inositol
1,3,4,5-tetrakisphosphate may stimulate external Ca2+ entry
into the cell. The increase in [Ca2+]i and
diacylglycerol are known to activate PKC (Fig. 1) (130).
PKC participates in -thrombin-induced increases in endothelial
permeability (7, 105, 119, 153, 168, 187). -Thrombin induces a rapid and dose-dependent translocation of PKC activity from
the cytosol to the membrane, as assessed by -[32P]ATP
phosphorylation of H1 histone in bovine pulmonary arterial endothelial
cell monolayers. Thrombin-induced PKC activation is temporally
associated with endothelial cell contraction demonstrated by changes in
cell morphology, similar at least in part to the effects of PMA
described above (168). It is also possible that the
-thrombin-induced increase in endothelial permeability occurs independently of PLC activation and increased
[Ca2+]i, despite the fact that the
-thrombin-induced increase in endothelial permeability requires a
PKC-dependent pathway (7). In similar studies,
inhibition of PKC activity prevents -thrombin-induced phosphorylation of the cytoskeletal protein caldesmon77 and
the intermediate filament vimentin and attenuates the endothelial cell
contraction as indicated above (168). In other studies, -thrombin induced a PKC--dependent increase in stress fibers, a
disruption of junctional VE-cadherin, and a decrease in TEER (153). Again, these results demonstrate that
-thrombin-induced PKC activity results in alteration of the
cytoskeleton, an event resulting in endothelial barrier dysfunction
(168) (Fig. 3). The phosphorylation of GDI and
Rho-dependent endothelial barrier dysfunction is a potential mechanism
for the effect of -thrombin-induced activation of PKC- as seen
with PMA (Fig. 3) (119). In the same study
(119), PKC- did not have a role in the
-thrombin-induced increase in endothelial permeability, indicating a
PKC isotype-specific role in response to -thrombin. Interestingly, a
role for another PKC isotype is indicated in the response to
-thrombin in endothelial cells. In a human dermal microvascular
endothelial cell line (HMEC-1), PKC-1 downregulates the -thrombin
receptor and suppresses the increase in endothelial permeability in
response to -thrombin (187). HMEC-1 transduced with
full-length PKC-1 AS cDNA or control pLNCX vector created the stable
cell lines HMEC-AS and HMEC-pLNCX, respectively. In the HMEC-AS, there
was expression of the AS-PKC-1 transcript and a decreased PKC-1
protein level without a change in PKC- or PKC-. The baseline
endothelial permeability of the HMEC-1, HMEC-pLNCX, and HMEC-AS were
comparable. -Thrombin induced a similar increase in permeability in
HMEC-1 and HMEC-pLNCX. In contrast, -thrombin stimulation of HMEC-AS
enhanced the increase in endothelial permeability compared with HMEC-1
and HMEC-pLNCX (187). Thus PKC-1, via a negative
feedback loop, modulates endothelial monolayer injury in response to
-thrombin (187). However, the role of PKC-1 changes
with the mediator used to alter the endothelial function, because
overexpression of the PKC-1 isotype augments the increase in
endothelial permeability in response to PMA (as indicated above), as
opposed to the suppressive role of PKC-1 during the response to
-thrombin (126).
In addition to phosphorylation events, the action of phosphatases has
an impact on the response to PKC activation. Specifically, the
serine/threonine protein phosphatases (PPs), including PP1, PP2A, and
PP2B, are implicated in PKC-mediated endothelial injury (105). Bovine pulmonary microvessel endothelial cells
express three major PPs: PP1, PP2A, and PP2B (105).
Inhibition of PP2B, but not of PP1 and PP2A, potentiated
-thrombin-induced increases in PKC- activity but not PKC-
activity. The inhibition of PP2B prevented normalization of the
thrombin-induced decrease in TEER; therefore, PP2B, via its effect
on PKC- activity, has a role in restoring thrombin-induced
endothelial barrier dysfunction, i.e., thrombin 
PP2a activity
PKC- activity (Fig. 3) (105).
TOP
ABSTRACT
INTRODUCTION
-INDUCED LUNG...
![= <IT>K</IT><SUB>fc</SUB> [(P<SUB>mv</SUB> − P<SUB>i</SUB>) + &sfgr;(&pgr;<SUB>i</SUB> − &pgr;<SUB>mv</SUB>)]](/content/vol284/issue3/fulltext/L435/img002.gif)
