2244? The?Journal?of?Clinical?Investigation? ? ? http://www.jci.org? ? ? Volume 116? ? ? Number 8? ? ? August 2006
Anaphylactic shock depends on PI3K
and eNOS-derived NO
Anje Cauwels,1 Ben Janssen,2 Emmanuel Buys,1,3 Patrick Sips,1 and Peter Brouckaert1
1Molecular Pathophysiology and Experimental Therapy Unit, Department for Molecular Biomedical Research,
Ghent University and Flanders Interuniversity Institute for Biotechnology (VIB), Ghent, Belgium. 2Department of Pharmacology and Toxicology,
Cardiovascular Research Institute Maastricht, University of Maastricht, Maastricht, The Netherlands. 3Cardiovascular Research Center,
Massachusetts General Hospital, Charlestown, Massachusetts, USA.
Anaphylaxis is an acute, severe, and potentially fatal systemic aller-
gic reaction that usually attacks the patient in the absence of a
history of allergy. Causes of anaphylaxis include bee stings, medi-
cations, food, latex exposure, and exercise (1). Anaphylaxis is not
as rare as generally believed and may affect as much as 1.2% to 15%
of the US population (2). In addition, it is generally recognized
that the prevalence of anaphylaxis is increasing significantly. Dur-
ing an anaphylactic reaction, serious cardiovascular or pulmonary
dysfunction often leads to acute death, with severe hypotension
as the cardinal clinical feature. Intravenous fluids and parenteral
adrenaline administration are considered the only effective treat-
ment; unfortunately, severe cardiovascular collapse during ana-
phylaxis is often resistant to this treatment (1, 3).
Platelet-activating factor (PAF) is a biologically active phospho-
lipid. Although PAF was originally named after its ability to induce
platelet aggregation, it stimulates a broad range of cell types and
tissues by binding to a unique G protein–coupled 7 transmem-
brane receptor (PAF-R), activating a number of signaling pathways,
including those that elevate intracellular calcium content and acti-
vate phospholipase A2 (PLA2) and PI3K (4). As PAF antagonists or
PAF-R deficiency can prevent fatal anaphylaxis in animal models,
PAF is thought to be a critical factor in the development of ana-
phylactic shock (5–9). In addition, PAF is known to contribute to
hypotension and cardiac dysfunctions during hemorrhagic, trau-
matic, or septic shock (4). Nevertheless, the downstream molecular
pathways beyond PAF-R triggering that direct its shock-inducing
effect are still enigmatic. Both ROS and NO have been implicated
in PAF shock, but whether their involvement is detrimental or ben-
eficial is still a matter of debate (10–12). However, most studies were
performed in animals under general anesthesia, which influences
NO-mediated effects and blood pressure considerably (13–15).
Also, conflicting results in anaphylactic shock models have been
reported with regard to the possible involvement of NO (16–19).
NO may be endogenously produced by the inducible iNOS or
by the constitutively expressed isoforms eNOS and neuronal NOS
(nNOS). While the latter are believed to produce low amounts of
NO to execute physiological and/or antiinflammatory functions,
the inflammation-associated expression of iNOS and the subse-
quent overproduction of NO are assumed to be responsible for
the cardiovascular failure, morbidity, and mortality associated
with (septic) shock (20, 21). In addition, it is generally accepted
that NO-mediated activation of soluble guanylate cyclase (sGC)
is the key mechanism in the regulation of vascular tone during
(septic) shock (22, 23).
In this study, we aimed to unequivocally elucidate the involve-
ment and importance of NO, sGC, ROS, and PLA2 signaling in
shock induced by intravenous PAF or anaphylaxis in conscious,
nonanesthetized mice. We found that PAF-induced shock com-
pletely depended on PI3K signaling and on NO produced by
the constitutive enzyme eNOS and not on NO produced by the
inducible “inflammatory” iNOS. Downstream from NO, PAF
shock appeared to be independent of sGC. Also, in 2 different
models of active systemic anaphylaxis, absolute protection was
obtained by inhibition of PI3K, Akt, or NOS or by eNOS deficien-
cy. Although it is generally accepted that excessive iNOS-derived
NO underlies the cardiovascular collapse associated with shock,
Nonstandard?abbreviations?used: EB, Evans blue; l-NAME, Nω-nitro-l-arginine
methyl ester; MAP, mean arterial blood pressure; MB, methylene blue; ODQ,
1H-[1,2,4]oxadiazole[4,3-a]quinoxalin-1-one; PAF, platelet-activating factor; PLA2,
phospholipase A2; sGC, soluble guanylate cyclase; WM, wortmannin.
Conflict?of?interest: The authors have declared that no conflict of interest exists.
Citation?for?this?article: J. Clin. Invest. 116:2244–2251 (2006). doi:10.1172/JCI25426.
Related Commentary, page 2075
? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 116 Number 8 August 2006
our data provide compelling new evidence for a shock-inducing
function of eNOS-derived NO.
NO is critical for PAF-induced shock. Intravenous PAF injection in con-
scious mice elicits rapid shock and results in death within 20–30
minutes (Figure 1A). When PAF engages its specific receptor, sev-
eral intracellular signaling events are triggered, including PLA2
activation and superoxide production (4). Nevertheless, inhibition
of these events by the PLA2 inhibitor aristolochic acid or the cell-
permeable superoxide dismutase mimetic tempol did?not provide
any protection against PAF challenge (Figure 1A). In comparison,
tempol or aristolochic acid provided complete protection against
hyperacute shock induced by TNF in combination with the caspase
inhibitor zVAD-fmk (24). As PAF may cause severe hypotension and
iNOS induction (25), the vasorelaxant NO is generally assumed to
be critically involved in PAF shock. Nonetheless, NOS inhibition
cannot always preclude PAF-induced hypotension and sometimes
even aggravates fatality (10–12). Since all of the aforementioned
studies were performed in animals under general anesthesia, which
is known to influence NO-mediated effects and blood pressure
(13–15), we studied the effect of NOS inhibition on PAF hypoten-
sion and shock in conscious mice. We found that Nω-nitro-l-argi-
nine methyl ester (l-NAME) pretreatment completely protected
against PAF lethality, implicating NO in a pivotal and detrimental
function (Figure 1, A and B). Even when the PAF dose exceeded the
LD100, l-NAME still provided absolute protection (not shown).
NO is thought to relax vascular smooth muscle cells by activa-
tion of sGC and subsequent cyclic GMP–dependent modification
of several intracellular processes, including the phosphorylation
of proteins of the contractile apparatus and of pumps or channels
involved in modulating intracellular calcium and potassium levels
(26). Moreover, recent studies reported that methylene blue (MB),
an inhibitor of sGC activation, can reverse clinical anaphylactic
shock induced by injected contrast media and prolong survival of
rabbits in experimental anaphylaxis (27, 28). Nevertheless, MB (15
mg/kg) did not prevent PAF shock to the same degree as l-NAME
treatment, implying that sGC-independent NO activities are also
involved (Figure 1, B and C). In comparison, MB provided 100%
protection in a mouse model of TNF-induced shock as well as in a
model of hyperacute shock induced by TNF in combination with
the caspase inhibitor zVAD-fmk, as reported previously (24, 29).
Also, lower doses of MB (5 and 10 mg/kg) still provided partial
protection (data not shown). However, pretreatment with various
doses (5, 10, 15, or 20 mg/kg) of another, more specific sGC inhibi-
tor, 1H-[1,2,4]oxadiazole[4,3-a]quinoxalin-1-one (ODQ) (23, 30),
at different time points (4, 2, or 0.5 hours before challenge) did not
protect (data not shown). In addition, sGCα1–/– mice (31) deficient
for 1 of the 2 isoforms of the α subunit of the heterodimeric sGC
enzyme were also not protected against PAF-induced shock (not
shown). Together, these results suggest that, downstream from
NO, sGC-independent events are most important.
To examine the hypotensive response to PAF, mean arterial
blood pressure (MAP) was measured in unrestrained conscious
mice using a catheter placed in the femoral artery. As documented
before (8), intravenously administered PAF has an extremely fast
and drastic hypotensive effect (Figure 2, A and C). However, when
mice were pretreated with l-NAME, PAF caused only a transient
and mild drop in blood pressure, followed by a very fast recovery
in all but 1 of the mice (Figure 2B). All mice treated with PAF died
within 13 minutes (Figure 2A) while only 1 of 5 animals pretreated
with l-NAME died, after 44 minutes (Figure 2B). The MAP of the 4
surviving l-NAME–treated mice 24 hours after PAF challenge was
132 ± 18 mmHg. We therefore conclude that NO plays a dominant
role in acute PAF-induced hypotension. Nevertheless, other vaso-
active mediators, such as histamine, serotonin, or leukotrienes,
may also contribute in an NO-independent manner, albeit to a
much smaller extent.
NO that causes PAF shock is produced by eNOS, activated via PI3K. To
identify the NOS isoform involved, we used various NOS-deficient
animals. Injection of PAF in iNOS-deficient mice caused a mortal-
ity that was similar to and as rapid as that observed in WT control
animals, but age-matched eNOS-deficient animals were signifi-
cantly protected (Figure 3A), indicating an unexpected pivotal role
for eNOS in PAF shock. Studies in vitro have shown that eNOS
can be activated by direct serine phosphorylation, obviating the
need for calcium and calmodulin binding and causing increased
NO production (32, 33). This phosphorylation is mediated by the
eNOS-derived NO is critical for PAF shock. C57BL/6 mice received
various pretreatments and were challenged i.v. with 55 μg PAF; sur-
vival was monitored. (A) Pretreatment with vehicle, aristolochic acid
(ArAc), tempol, or l-NAME (–4 hours). The number of mice is indicated
between brackets. ***P < 0.0001; *P = 0.0201 compared with PAF. (B)
Pretreatment with l-NAME 4, 2, or 1 hour before treatment with PAF.
**P = 0.0165. (C) Pretreatment with MB 6, 4, 2, or 1 hour before PAF.
2246? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 116 Number 8 August 2006
PI3K/Akt pathway, which can be activated by receptor tyrosine
kinases, G protein–coupled receptors, or mechanical forces such
as shear stress. Pretreatment with wortmannin (WM), a specific
inhibitor of PI3K, completely protected both WT and iNOS–/– mice
(Figure 3A), demonstrating the critical importance of PI3K in PAF-
induced shock. In addition, Western blot analysis of WT kidney
homogenates revealed specific Ser1177 eNOS phosphorylation
after PAF challenge, which can be efficiently prevented by WM
pretreatment (Figure 3B).
Anaphylactic shock depends on PI3K and eNOS-derived NO. To evalu-
ate the pathophysiological involvement of PI3K and eNOS-derived
NO in shock in a clinically more relevant setting, we set up 2 dif-
ferent models of active systemic anaphylaxis, known to depend
predominantly on PAF signaling (8, 9). Mice were first sensitized
with BSA or OVA in the presence of adjuvants. A few weeks later,
the animals were challenged intravenously with the same antigen.
Soon after this challenge, mice developed severe hypothermia
(demonstrated in Figure 4, B and D, for BSA-induced anaphy-
laxis) and rapidly succumbed to systemic shock reaction (Figure
4A). However, when mice were pretreated with WM or l-NAME,
hypothermia was transient and less severe (Figure 4B). While 75%
of the control mice died, all 8 animals pretreated with WM or
l-NAME survived the anaphylactic challenge (Figure 4A). Similar
results were obtained in the OVA-induced anaphylaxis model (data
not shown). Western blot analysis of WT kidney homogenates
confirmed specific eNOS phosphorylation after BSA challenge in
sensitized mice, prevented by WM pretreatment (Figure 4E). Next,
we compared the 2 models of anaphylaxis in WT and eNOS-defi-
cient mice. Fatality in WT animals was 82% (9 of 11) for the BSA
model and 92% (11 of 12) for the OVA model. In contrast, none
of the eNOS-deficient mice died (Figure 4C; not shown for the
OVA model). Moreover, for both anaphylaxis models, there was
only a delayed, mild, and transient hypothermia in eNOS-deficient
mice (Figure 4D). Neither BSA nor OVA had an effect on the tem-
perature or behavior of unsensitized mice (Figure 4, B and D; not
shown for OVA). These results reveal an unexpected but pivotal
role for PI3K and eNOS-derived NO in anaphylactic shock.
During anaphylaxis, vasodilation and hypotension are often
associated with excessive vascular permeability and fluid extrava-
sation, resulting in a mixed distributive-hypovolemic shock. To
assess vascular permeability, protein leakage was measured with
the Evans blue (EB) technique. BSA or OVA in 1% EB was injected
into the lateral tail vein, and subsequent EB extravasation into the
ear tissue was analyzed (Figure 4F). In the BSA/BSA anaphylactic
response, vascular permeability was significantly increased in WT
mice but not in eNOS–/– animals (Figure 4F). Moreover, vascular
leakage in BSA/BSA eNOS–/– mice was 2-fold lower than in WT
mice (P < 0.01 by ANOVA). In OVA/OVA-induced shock, vascular
permeability increased only in WT animals, and extravasation was
again significantly lower in eNOS–/– OVA/OVA mice than in WT
OVA/OVA animals (P < 0.001 by ANOVA) (not shown). In addition,
pretreatment with WM or l-NAME clearly prevented EB extravasa-
tion in WT mice (P < 0.001, data not shown). Together, these data
suggest the involvement of PI3K/eNOS in vascular leakage associ-
ated with anaphylaxis.
Once activated, PI3Ks generate phosphoinositide 3,4,5-trisphos-
phate (PIP3), leading to the recruitment and activation of Akt (pro-
tein kinase B), which then activates a range of downstream targets,
MAP measurements during PAF shock. C57BL/6 mice were challenged i.v. with 55 μg PAF at t = 0. (A) Individual MAP of PAF-challenged mice. (B)
Individual MAP of mice treated with l-NAME (100 mg/kg, i.v., –2 hours) and PAF. (C) Data points show the mean MAP ± SD of the plots shown in
A and B. Error bars are plotted as vertical lines above and below the curves. The number of mice is indicated between brackets. ***P < 0.0001.
NO that causes PAF shock is produced by eNOS, activated via PI3K.
(A) C57BL/6 (WT), iNOS–/–, or eNOS–/– mice were injected i.v. with 55
μg PAF after pretreatment with vehicle or WM. Survival is presented
as a combined Kaplan-Meier plot of 2 independent, representative
experiments. The total number of mice is indicated between brackets.
***P < 0.0001 compared with PAF in WT mice. (B) Western blot for
phosphorylated eNOS (p-eNOS) in kidney homogenates of individual
mice 7 or 15 minutes after PAF injection. Where indicated, mice were
pretreated with WM.
? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 116 Number 8 August 2006
including eNOS. Akt exists as 3 isoforms (Akt1, Akt2, and Akt3)
that share extensive structural similarity and that may have unique
as well as common functions within the cell. To corroborate the
importance of the PI3K/Akt pathway in anaphylaxis, we treated
OVA-sensitized mice with perifosine, a lipid-based inhibitor of Akt
for which phase I and II clinical trials are being conducted (34, 35).
As depicted in Figure 5, A and B, all mice pretreated with perifosine
suffered much less from hypothermia and survived the otherwise
100% lethal OVA challenge. For all Akt isoforms, knockout mice
have been created (36–38). As Akt1 is the predominant Akt isoform
in endothelial cells (39), we also studied anaphylactic shock in
Akt1-deficient mice. However, Akt1–/– ani-
mals were not protected, except for a minor
(but significant) delay in mortality, as com-
pared with WT mice (Figure 5C).
To examine the effect of NOS deficiency
on anaphylaxis-induced hypotension, MAP
was measured in unrestrained conscious
mice 3 days after surgical catheter inser-
tion into the right femoral artery. In ani-
mals previously sensitized by BSA, intra-
venous BSA challenge had an immediate
precipitous hypotensive effect, resulting
in mortality, regardless of the challenging
dose of BSA used (2, 1, or 0.1 mg) (Figure
6A). When mice were challenged with 2 mg
BSA, pretreatment with l-NAME only pro-
longed survival: l-NAME mice died 40–81
minutes after challenge while all control
animals were dead within 20 minutes.
However, when 0.1 mg BSA was used as a
challenge, the BSA challenge in l-NAME
pretreated mice caused only a transient
drop in blood pressure, followed by a
very fast recovery in all but 1 mouse (Fig-
ure 6B). All mice treated with 0.1 mg BSA
died within 15 minutes (Figure 6, A and C)
while only 1 of 4 animals pretreated with
l-NAME died, after 72 minutes (Figure 6B).
NO has been shown to play a primary
and harmful role in (septic) shock, caus-
ing progressive refractory hypotension
and ultimately leading to multiple organ
dysfunction and death. Constitutive NOS
enzymes (eNOS and nNOS) are believed to
produce low amounts of NO essential for
various physiological and/or antiinflam-
matory events while higher amounts of
NO are produced by inducible iNOS dur-
ing inflammation (20, 21). It is generally
accepted that during inflammation and
shock, cardiovascular collapse is initiated
and mediated by excessive production of
NO due to the transcriptional induction
of the inducible “inflammatory” isoform
iNOS, a process that requires hours rather
than minutes. Downstream, sGC is regard-
ed as the principal vasorelaxing mediator
of NO in the cardiovascular system (22, 23, 26).
As PAF can participate in the induction of iNOS (25), induced
NO is assumed to mediate PAF-induced hypotension and shock.
While some investigators have reported that NOS inhibitors can
prevent PAF-induced hypotension (40, 41), others reported that
it cannot (11, 12). However, all of these studies were performed in
animals under general anesthesia, known to influence NO-medi-
ated effects and blood pressure changes considerably (13–15). In
addition, NOS inhibitors have also been reported to exacerbate
PAF shock, including enhanced pulmonary hypertension (11),
aggravated alveolar permeability (12), and even an increase in
BSA anaphylactic shock in WT and eNOS–/– mice. (A) Survival plot of C57BL/6 mice sensitized
with BSA, treated with vehicle (controls), l-NAME, or WM, and finally challenged with BSA in
EB. The number of mice in each experimental group is shown in parentheses. **P = 0.0022. (B)
Temperature plot (mean ± SD) of mice sensitized by BSA or PBS and challenged 15 days later
with BSA in EB (BSA + EB) or EB alone. The total number of mice is shown in parentheses;
numbers beneath time points indicate the number of mice still alive at the indicated times. (C)
Survival plot of WT and eNOS–/– mice sensitized with BSA and challenged with BSA in EB. Total
number of mice is shown in parentheses. ***P = 0.0007. (D) Temperature plot (mean ± SD) of
mice sensitized by BSA or PBS and challenged with BSA in EB or EB alone. The total number of
mice is shown in parentheses, numbers above the time points indicate the number of mice still
alive at the indicated times. (E) Western blot for p-eNOS in kidney homogenates of individual,
BSA-sensitized mice 12 minutes after BSA challenge. Where indicated, mice had been pre-
treated with WM. (F) Vascular leakage assayed by EB extravasation into ear tissues. Inner ear
rims were removed 90 minutes after i.v. challenge with BSA in EB or EB only and subsequently
extracted in formamide. Plots show mean OD620 ± SD. #P < 0.001; ##P < 0.01.
2248? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 116 Number 8 August 2006
mortality of experimental animals (10, 12). In vitro, PAF stimu-
lates leukocyte superoxide production, chemotaxis, and adhesion
(42). Although it has been suggested that some biological activi-
ties of PAF are mediated by ROS, antioxidants such as butylated
hydroxytoluene, vitamin E, or N-acetylcysteine failed to inhibit
PAF-induced mortality, disseminated intravascular coagulation,
or pulmonary edema (43, 44). Furthermore, the exact role of NO
in various anaphylactic shock models also remains unclear, with
reports ranging from l-NAME–induced reduction of mortality
(16) to reduced survival (17, 19).
In the present study, we set out to unequivocally elucidate the
involvement and importance of NO, sGC, ROS, and PLA2 sig-
naling in shock induced by intravenous PAF in conscious, non-
anesthetized mice. To validate the results we obtained in the PAF-
induced shock model, we also used 2 different models of active
murine anaphylactic shock.
We found that PAF-induced mortality was not influenced by anti-
oxidant treatment or PLA2 inhibition. Intravenous PAF injection in
conscious mice elicited a very acute and drastic drop in blood pres-
sure, resulting in death within only 20–30 minutes. As iNOS activa-
tion depends on de novo synthesis of both RNA and protein, sev-
eral hours are required for its full activation. Surprisingly, however,
NOS inhibition by l-NAME prevented PAF-induced hypotension
and provided complete protection, even against PAF doses exceed-
ing the LD100. Moreover, whereas PAF injection in iNOS-deficient
mice caused equally rapid mortality, eNOS-deficient animals were
protected, indicating an unexpected pivotal role for eNOS-derived
NO in PAF shock. Inhibitors of sGC protected either partially (MB)
or not at all (ODQ), and sGCα1–/– and WT mice suffered equally
from PAF-induced shock. Catalytically active sGC heterodimers are
composed of an α1 or α2 subunit combined with a β1 subunit. Of
interest, direct pharmacological sGC activation did not cause vaso-
relaxation or hypotension in sGCα1–/– mice, indicating that sGC-
dependent cardiovascular events predominantly rely on sGCα1
(31). Our results therefore imply that, downstream from NO, sGC-
independent events are most important during PAF shock.
Traditionally, eNOS has been considered a constitutively
expressed enzyme regulated by calcium and calmodulin. However,
OVA anaphylactic shock in WT and Akt1–/– mice. (A) Survival plot of
C57BL/6 mice sensitized with OVA, pretreated with vehicle (controls)
or perifosine, and challenged 19 days later with OVA in 1% EB. The
total number of mice in each experimental group is shown in parenthe-
ses. ***P = 0.0002. (B) Temperature plot (mean ± SD) of mice sensi-
tized by OVA and challenged 19 days later with OVA in 1% EB. The
total number of mice is shown in parentheses, and numbers below the
time points indicate the number of mice still alive at the indicated times.
(C) Survival plot of WT and Akt1–/– mice, sensitized with OVA and chal-
lenged 20 days later with OVA in 1% EB. The total number of mice is
shown in brackets in the legend. **P = 0.0043.
MAP measurements during anaphylactic shock. BSA-sensitized C57BL/6 mice were challenged i.v. with BSA at t = 0. (A) Individual MAP of mice
challenged with 2, 1, or 0.1 mg BSA. (B) Individual MAP of mice pretreated with l-NAME (200 mg/kg, i.v.; –2 hours) and 0.1 mg BSA (t = 0). (C)
Data points show the mean MAP ± SD of the 0.1 mg BSA plots shown in A and B. Error bars are plotted as vertical lines above and below the
curves. Number of mice is indicated in parentheses. ***P < 0.0001.
? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 116 Number 8 August 2006
phosphorylation of a C terminal serine residue through the PI3K/
Akt pathway activates the eNOS enzyme at resting calcium levels
and/or increases its activity (32, 33, 45). Interestingly, PI3K/Akt-
dependent eNOS phosphorylation and ensuing NO production
occur much faster than iNOS-dependent NO production, gener-
ally reaching a maximum within only 10 to 15 minutes of stimula-
tion (32, 46, 47). In our study, pretreatment with WM, a specific
PI3K inhibitor, prevented PAF-induced eNOS Ser1177 phosphory-
lation and efficiently protected both WT and iNOS-deficient mice
from PAF shock. Hence, we conclude that there is a critical role for
PI3K in the rapid activation of eNOS during PAF-induced shock.
To assess the pathophysiological role of PI3K/Akt and eNOS-
derived NO in a clinically more relevant shock setting, we tested
the effects of WM, perifosine, l-NAME, and eNOS or Akt1 defi-
ciency in 2 different active and acute systemic anaphylaxis models.
The experiments revealed a critical role for PI3K/Akt and eNOS-
derived NO in anaphylactic shock. Following pretreatment with
WM, perifosine, or l-NAME or in conditions of eNOS deficiency,
anaphylaxis-induced hypothermia was mild and transient, vascu-
lar permeability was significantly lower, and no deaths occurred
at all, indicating that PI3K/Akt-activated eNOS-derived NO is the
most important vasodilator in anaphylactic shock.
Understanding the requirement of Akt in mammals is com-
plicated by the existence of 3 highly conserved isoforms, Akt1,
Akt2, and Akt3. Akt1-deficient mice show impaired overall
growth, Akt2 knockout mice are insulin intolerant, demonstrat-
ing a diabetes-like syndrome, and mice lacking Akt3, which has
more limited distribution, display a selective reduction in brain
size (36–38, 48). Despite the protection obtained by Akt inhibi-
tion, Akt1–/– mice were not protected against anaphylactic shock,
except for a minor delay in mortality. However, the viability of
adult mice lacking each individual Akt isoform suggests that
there is functional redundancy among these proteins. Interest-
ingly, mice lacking both Akt1 and Akt2 die shortly after birth (49).
In addition, endothelial cells of Akt1-deficient mice still show
robust Akt levels and increased eNOS phosphorylation following
VEGF treatment, which clearly provides evidence of the presence
of compensatory mechanisms (39). This redundancy may explain
why the Akt inhibitor perifosine is much more efficient than Akt1
deficiency in preventing anaphylactic shock.
Historically, the constitutively expressed eNOS enzyme was
never seriously implicated as a detrimental mediator of inflam-
mation. On the contrary, eNOS-derived NO was generally consid-
ered to exercise a rather antiinflammatory role (20). Nevertheless,
the early activation of eNOS was already linked to an early, tran-
sient vascular hyporeactivity during both endotoxic and hemor-
rhagic shock more than a decade ago (50, 51). Although eNOS was
originally believed to be constitutively expressed and modulated
exclusively via changes in intracellular calcium concentration, it
has now become clear that eNOS can be induced several-fold by
multiple stimuli (such as shear stress, estrogen, and VEGF) and
posttranslationally activated through protein phosphorylation or
protein-protein interactions (45, 52). Moreover, eNOS deficiency
or eNOS inactivation efficiently prevents the acute local inflam-
matory response induced by subplantar carrageenan injection (53,
54). Very recently, an important function for eNOS in facilitat-
ing iNOS expression and hypotension in endotoxemia was found,
demonstrating a regulatory role for eNOS-derived NO in the gene
expression of inflammatory mediators such as iNOS (55, 56). Our
present study reveals what we believe is a new and surprising con-
cept that eNOS-derived NO is the principal vasodilator in PAF-
induced and anaphylactic shock, thereby strongly supporting the
emerging belief in a possible detrimental role for eNOS in regulat-
ing inflammation and/or vascular function during shock.
In addition, MAP measurements in anaphylactic shock demon-
strated that the abrupt and precipitous drop in blood pressure,
which happens immediately after the challenge and lasts only a
couple of minutes, is clearly NOS/NO-independent. This may be
attributed to the rapid release of preformed vasoactive mediators,
such as histamine and/or serotonin. However, following this ini-
tial dramatic blood pressure drop (of ± 80 mmHg), l-NAME–pre-
treated mice quickly recovered and eventually survived while con-
trol animals suffered a continued steep drop in blood pressure,
leading to abrupt mortality. These data clearly demonstrate that
eNOS-dependent vasorelaxation is key to the sustained hypoten-
sion causing shock and mortality during anaphylaxis.
Up until lately, very little was known about the role of PI3K in
the cardiovascular system. Recently, the PAF/PI3K/eNOS path-
way was reported to exert a potent negative inotropic effect on the
heart and to be an important cause of myocardial dysfunction fol-
lowing ischemia/reperfusion (57, 58). Our data provide important
additional information on a similar detrimental effect of the PAF/
PI3K/eNOS pathway with respect to shock. In addition, we clearly
demonstrate that anaphylactic shock in conscious mice is crucially
orchestrated by PI3K/Akt and eNOS-derived NO. Although cyclic
GMP/protein kinase G–dependent phosphorylation of the con-
tractile machinery and calcium-regulating proteins is generally
assumed to be the major mechanism by which NO causes patho-
logical vascular smooth muscle relaxation, our results suggest that
sGC-independent actions of NO may contribute significantly to its
shock-inducing effect in vivo. As a general conclusion, we propose
that selective targeting of eNOS and/or PI3K or Akt may provide
new tools for the treatment of anaphylactic shock.
Mice. Female C57BL/6 mice were purchased from Elevage Janvier and Akt1–/–
mice from Jackson Laboratory, where we also obtained iNOS-deficient (59)
and eNOS-deficient (60) mice on a C57BL/6 background to breed in our
facilities. The sGCα1–/– mice were generated as reported (31), and litter-
mates were used as controls. Mice were housed in temperature-controlled,
air-conditioned facilities with 14-hour light/10-hour dark cycles and food
and water ad libitum. All data were collected using mice 8–12 weeks of age.
All experiments were approved by and performed according to the guide-
lines of the animal ethics committees of Ghent University, Belgium, and
Maastricht University, The Netherlands.
Cytokines, reagents, injections, body temperature, and MAP measurements. PAF
(1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine; Sigma-Aldrich) was
diluted in 200 μl endotoxin-free PBS supplemented with 0.25% BSA and
injected i.v. Mortality was scored up to 7 days after challenge. The NOS
inhibitor l-NAME was purchased from Novabiochem and injected i.v. at
a dose of 100 mg/kg. Tempol (Sigma-Aldrich) was injected i.p. at 6 mg/
mouse 1 hour before PAF. Aristolochic acid (Sigma-Aldrich) was admin-
istered s.c. in 50 μl DMSO 1.5 hours before PAF at 120 μg/mouse. MB in
glucose solution suitable for i.v. injection (MB; Sterop) was used at a dose
of 15 mg/kg unless otherwise stated. All of these treatments were based
on the protective effects observed in our previous studies (24, 29). ODQ
(Tocris Bioscience) was used i.p. in 50 μl DMSO at 20, 15, 10, or 5 mg/kg.
WM (Sigma-Aldrich, 30 μg/mouse in 1.5% DMSO) was given i.p. 1.3
hours before the PAF challenge. Perifosine (octadecylphosphopiperidine;
Keryx Biopharmaceuticals) was injected s.c. (1 mg/100 μl PBS/mouse) for
2250? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 116 Number 8 August 2006
4 consecutive days, the last treatment being 2 hours before challenge. Rec-
tal body temperature was recorded with an electronic thermometer (model
2001; Comark). Blood pressure and heart rate were measured continuously
in conscious, permanently catheterized C57BL/6 mice, as described (24,
29). Briefly, a heat-stretched piece of polyethylene tubing was inserted into
the right femoral artery. Similarly, a catheter was placed in the jugular vein
for drug injections. After surgery, animals received Ringer solution (Baxter
BV) and analgesic and were kept at 30°C to improve recovery. Experiments
were performed 3 days after surgery, at which time the catheter was con-
nected to a low-volume pressure transducer to record MAP and HR con-
tinuously throughout the study. The setup was allowed to stabilize for at
least 40 minutes before the experiment was started.
Active systemic anaphylaxis models. Two different sensitization models were
used: (a) Mice were given a single i.p. injection of 1 mg BSA (Sigma-Aldrich)
mixed with 300 ng pertussis toxin (Sigma-Aldrich). Anaphylaxis was elic-
ited 15 days later by i.v. injection of 2 mg of BSA. In the hemodynamic
studies, the dose of BSA had to be lowered from 2 to 0.1 mg because the
surgical procedures led to a sensitization to the BSA challenge. (b) Mice
were sensitized by i.p. injection of 100 μg OVA (Sigma-Aldrich), aluminum
hydroxide (Sigma-Aldrich, 1 mg) and pertussis toxin (300 ng). Mice were
challenged 19–20 days later by i.v. injection of 150 μg OVA. When appro-
priate, mice were pretreated with l-NAME (i.v., –2.0 hours), WM (i.p., –1.3
hours), or perifosine (s.c.). Sensitizing i.p. injections were given in 0.5 ml of
endotoxin-free PBS. To permit analysis of vascular permeability, challenges
were injected into the lateral tail vein in 0.2 ml of 1% EB in endotoxin-free
PBS, and ear biopsies were taken 90 minutes later. Control mice were chal-
lenged with EB only. As long as vascular permeability is unaffected, EB
remains confined to the intravascular space. To quantify vascular leakage,
ear biopsies were incubated in 0.3 ml formamide at 60°C for 4 days to
extract EB from the tissue; absorption was measured at 620 nm.
Western blot analysis. Mice were pretreated with WM or vehicle and 1.3
hours later challenged with a lethal dose of PAF, BSA, or OVA. They were
euthanized by cervical dislocation at the indicated time. Organs were dis-
sected, snap-frozen in liquid nitrogen, and homogenized in a lysis buffer
suitable for detecting phosphorylated proteins containing 50 mM HEPES,
5 mM EDTA, 250 mM NaCl, 0.1% Nonidet P-40, 0.15 μM aprotinin, 2.1 μM
leupeptin, 1 mM Pefabloc, 25 mM β-glycerolphosphate, 0.2 mM sodium
vanadate, and 10 mM sodium fluoride. After centrifugation of cell debris,
protein concentrations were determined (Bio-Rad), and 20 mg were used for
immunoprecipitation with mouse anti-eNOS (BD Biosciences — Pharmin-
gen). Purified eNOS was resolved in a 4–12% Bis-Tris gel (Invitrogen) and
transferred to nitrocellulose (Schleicher & Schuell BioScience). Filters were
blocked for 1 hour at room temperature in TBS containing 0.1% Tween 20
(TBS/T) supplemented with 5% milk powder. The filters were incubated
overnight at 4°C with a mAb against phosphorylated eNOS (Ser1177) (Cell
Signaling Technology) in TBS/T with 5% BSA, washed 5 times with TBS/T,
and incubated for 1 hour at room temperature with a goat anti-rabbit
IgG conjugated with horseradish peroxidase (Cell Signaling Technology).
Afterwards, 5 washes with TBS/T preceded development with the Chemi-
luminescent Reagent Plus (ECL) detection reagent (PerkinElmer).
Statistics. Statistics were performed using GraphPad Prism version 4.0
(GraphPad Software; log-rank test for comparison of survival curves,
c2 test to compare total mortality percentages, unpaired 2-tailed Student’s
t test to compare mean MAP data, and 1-way ANOVA with Bonferroni’s
multiple comparison test for EB extravasation data).
The authors would like to thank Linda Van Geert, Joris De
Backer, and Geert Versporten for animal care; Jacques Debets
and Peter Leenders for preparation of mice in the hemodynamic
studies; Elke Rogge for help with immunoblotting; and Enrique
Poradosu (Keryx Biopharmaceuticals) for the generous gift of
perifosine. Research was supported by grants from the Fonds
Wetenschappelijk Onderzoek–Vlaanderen (FWO-Vlaanderen),
the Geconcerteerde Onderzoeksacties, and the Interuniversitaire
Attractiepolen. A. Cauwels is a postdoctoral fellow of the FWO-
Vlaanderen; P. Sips is an Instituut voor de aanmoediging van
Innovatie door Wetenschap en Technologie fellow.
Received for publication April 21, 2005, and accepted in revised
form May 2, 2006.
Address correspondence to: Anje Cauwels, Department for Molec-
ular Biomedical Research, Technologiepark 927, B-9052 Ghent
(Zwijnaarde), Belgium. Phone: 32-9-3313712; Fax: 32-9-3313609;
1. Kemp, S.F., and Lockey, R.F. 2002. Anaphylaxis: a
review of causes and mechanisms. J. Allergy Clin.
2. Neugut, A.I., Ghatak, A.T., and Miller, R.L. 2001.
Anaphylaxis in the United States: an investigation
into its epidemiology. Arch. Intern. Med. 161:15–21.
3. Brown, S.G. 2005. Cardiovascular aspects of ana-
phylaxis: implications for treatment and diagnosis.
Curr. Opin. Allergy Clin. Immunol. 5:359–364.
4. Montrucchio, G., Alloatti, G., and Camussi, G.
2000. Role of platelet-activating factor in cardiovas-
cular pathophysiology. Physiol. Rev. 80:1669–1699.
5. Terashita, Z., Imura, Y., Nishikawa, K., and Sumida,
S. 1985. Beneficial effects of (RS)-2-methoxy-3-(o
phosphate, a specific PAF antagonist, in endo-
toxin and anaphylactic shock. Adv. Prostaglandin
Thromboxane Leukot. Res. 15:715–717.
6. Sanchez-Crespo, M., Fernandez-Gallardo, S., Nieto,
M.L., Baranes, J., and Braquet, P. 1985. Inhibition
of the vascular actions of IgG aggregates by BN
52021, a highly specific antagonist of paf-acether.
7. Darius, H., Smith, J.B., and Lefer, A.M. 1986. Inhi-
bition of the platelet activating factor mediated
component of guinea pig anaphylaxis by recep-
tor antagonists. Int. Arch. Allergy Appl. Immunol.
8. Ishii, S., et al. 1998. Impaired anaphylactic respons-
es with intact sensitivity to endotoxin in mice lack-
ing a platelet-activating factor receptor. J. Exp. Med.
9. Fukuda, Y., et al. 2000. Effect of human plasma-
type platelet-activating factor acetylhydrolase in
two anaphylactic shock models. Eur. J. Pharmacol.
10. Takano, H., et al. 1997. Effects of aminoguanidine on
systemic inflammatory response syndrome induced
by platelet activating factor and by lipopolysaccha-
ride in rats. Eur. J. Pharmacol. 323:107–110.
11. Wang, H.G., Shibamoto, T., Matsuda, Y., and Koya-
ma, S. 1998. The role of endogenous nitric oxide
in the sympathetic and hemodynamic response to
platelet activating factor-induced hypotension in
anesthetized dogs. Shock. 9:58–64.
12. Yoshikawa, T., et al. 1997. Role of superoxide and
nitric oxide in platelet-activating factor-induced
acute lung injury, hypotension, and mortality in
rats. Crit. Care Med. 25:286–292.
13. Muldoon, S.M., Hart, J.L., Bowen, K.A., and Freas,
W. 1988. Attenuation of endothelium-mediated
vasodilation by halothane. Anesthesiology. 68:31–37.
14. Hart, J.L., et al. 1993. Effects of halothane on
EDRF/cGMP-mediated vascular smooth muscle
relaxations. Anesthesiology. 79:323–331.
15. Ma, J., et al. 1996. L-NNA-sensitive regional cere-
bral blood flow augmentation during hypercap-
nia in type III NOS mutant mice. Am. J. Physiol.
16. Amir, S., and English, A.M. 1991. An inhibitor of
nitric oxide production, NG-nitro-L-arginine-
methyl ester, improves survival in anaphylactic
shock. Eur. J. Pharmacol. 203:125–127.
17. Mitsuhata, H., et al. 1995. Nitric oxide synthase
inhibition is detrimental to cardiac function and
promotes bronchospasm in anaphylaxis in rabbits.
18. Mitsuhata, H., et al. 1995. An inhibitor of nitric
oxide synthase, N omega-nitro-L-arginine-methyl
ester, attenuates hypotension but does not improve
cardiac depression in anaphylaxis in dogs. Shock.
3:447–453; discussion 454.
19. Bellou, A., et al. 2003. Constitutive nitric oxide
synthase inhibition combined with histamine
and serotonin receptor blockade improves the ini-
tial ovalbumin-induced arterial hypotension but
decreases the survival time in brown norway rats
anaphylactic shock. Shock. 19:71–78.
20. Moncada, S. 1997. Nitric oxide in the vasculature:
physiology and pathophysiology. Ann. N. Y. Acad.
Sci. 811:60–67; discussion 67–69.
research article Download full-text
? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 116 Number 8 August 2006
21. Kirkeboen, K.A., and Strand, O.A. 1999. The role of
nitric oxide in sepsis–an overview. Acta Anaesthesiol.
22. Murad, F. 1996. The 1996 Albert Lasker Medical
Research Awards. Signal transduction using nitric
oxide and cyclic guanosine monophosphate. JAMA.
23. Ortega Mateo, A., and Amaya Aleixandre de Arti-
ñano, M. 2000. Nitric oxide reactivity and mecha-
nisms involved in its biological effects. Pharmacol.
24. Cauwels, A., Janssen, B., Waeytens, A., Cuvelier, C.,
and Brouckaert, P. 2003. Caspase inhibition causes
hyperacute tumor necrosis factor-induced shock
via oxidative stress and phospholipase A2. Nat.
25. Szabo, C., et al. 1993. Platelet-activating factor con-
tributes to the induction of nitric oxide synthase by
bacterial lipopolysaccharide. Circ. Res. 73:991–999.
26. Lucas, K.A., et al. 2000. Guanylyl cyclases and sig-
naling by cyclic GMP. Pharmacol. Rev. 52:375–414.
27. Oliveira Neto, A.M., Duarte, N.M., Vicente, W.V.,
Viaro, F., and Evora, P.R. 2003. Methylene blue: an
effective treatment for contrast medium-induced
anaphylaxis. Med. Sci. Monit. 9:CS102–CS106.
28. Buzato, M.A., Viaro, F., Piccinato, C.E., and Evora,
P.R. 2005. The use of methylene blue in the treat-
ment of anaphylactic shock induced by compound
48/80: experimental studies in rabbits. Shock.
29. Cauwels, A., et al. 2000. Protection against TNF-
induced lethal shock by soluble guanylate cyclase
inhibition requires functional inducible nitric
oxide synthase. Immunity. 13:223–231.
30. Zingarelli, B., Hasko, G., Salzman, A.L., and
Szabo, C. 1999. Effects of a novel guanylyl cyclase
inhibitor on the vascular actions of nitric oxide
and peroxynitrite in immunostimulated smooth
muscle cells and in endotoxic shock. Crit. Care Med.
31. Sips, P., Buys, E., Rogge, E., Dewerchin, M., and
Brouckaert, P. 2005. Functional knockout of the
soluble guanylate cyclase alpha1 subunit leads to
gender-specific hypertension while retaining sensi-
tivity to nitric oxide. Circulation. 112:A747.
32. Dimmeler, S., et al. 1999. Activation of nitric oxide
synthase in endothelial cells by Akt-dependent
phosphorylation. Nature. 399:601–605.
33. Fulton, D., et al. 1999. Regulation of endothelium-
derived nitric oxide production by the protein
kinase Akt. Nature. 399:597–601.
34. Van Ummersen, L., et al. 2004. A phase I trial of
perifosine (NSC 639966) on a loading dose/main-
tenance dose schedule in patients with advanced
cancer. Clin. Cancer Res. 10:7450–7456.
35. Granville, C.A., Memmott, R.M., Gills, J.J., and Den-
nis, P.A. 2006. Handicapping the race to develop
inhibitors of the phosphoinositide 3-kinase/Akt/
mammalian target of rapamycin pathway. Clin.
Cancer Res. 12:679–689.
36. Chen, W.S., et al. 2001. Growth retardation and
increased apoptosis in mice with homozygous dis-
ruption of the Akt1 gene. Genes Dev. 15:2203–2208.
37. Cho, H., et al. 2001. Insulin resistance and a diabetes
mellitus-like syndrome in mice lacking the protein
kinase Akt2 (PKB beta). Science. 292:1728–1731.
38. Easton, R.M., et al. 2005. Role for Akt3/protein
kinase Bgamma in attainment of normal brain size.
Mol. Cell. Biol. 25:1869–1878.
39. Chen, J., et al. 2005. Akt1 regulates pathological
angiogenesis, vascular maturation and permeabil-
ity in vivo. Nat. Med. 11:1188–1196.
40. Takekoshi, K., et al. 1993. Effect of NG-nitro-L-
arginine on shock induced by endotoxin and by
platelet activating factor in dogs. Eur. J. Pharmacol.
41. Yamada, T., Yukioka, H., Hayashi, M., Asada, A.,
and Inoue, M. 1998. Effects of inhaled nitric oxide
on platelet-activating factor-induced pulmonary
hypertension in dogs. Acta Anaesthesiol. Scand.
42. Snyder, F. 1990. Platelet-activating factor and relat-
ed acetylated lipids as potent biologically active cel-
lular mediators. Am. J. Physiol. 259:C697–C708.
43. Choi, J.H., et al. 2000. Selective involvement of reac-
tive oxygen intermediates in platelet-activating fac-
tor-mediated activation of NF-kappaB. Inflamma-
44. Herbert, J.M., et al. 1993. Biochemical and pharma-
cological properties of SR 27388, a dual antioxi-
dant and PAF receptor antagonist. J. Lipid. Mediat.
45. Fleming, I., and Busse, R. 2003. Molecular mecha-
nisms involved in the regulation of the endothelial
nitric oxide synthase. Am. J. Physiol. Regul. Integr.
Comp. Physiol. 284:R1–R12.
46. Duran, W.N., et al. 2000. Stimulation of NO produc-
tion and of eNOS phosphorylation in the microcir-
culation in vivo. Microvasc. Res. 60:104–111.
47. Hisamoto, K., et al. 2001. Estrogen induces the Akt-
dependent activation of endothelial nitric-oxide
synthase in vascular endothelial cells. J. Biol. Chem.
48. Cho, H., Thorvaldsen, J.L., Chu, Q., Feng, F., and
Birnbaum, M.J. 2001. Akt1/PKBalpha is required
for normal growth but dispensable for mainte-
nance of glucose homeostasis in mice. J. Biol. Chem.
49. Peng, X.D., et al. 2003. Dwarfism, impaired skin
development, skeletal muscle atrophy, delayed bone
development, and impeded adipogenesis in mice
lacking Akt1 and Akt2. Genes Dev. 17:1352–1365.
50. Szabo, C., Mitchell, J.A., Thiemermann, C., and
Vane, J.R. 1993. Nitric oxide-mediated hyporeac-
tivity to noradrenaline precedes the induction of
nitric oxide synthase in endotoxin shock. Br. J. Phar-
51. Szabo, C. 1995. Alterations in nitric oxide produc-
tion in various forms of circulatory shock. New
52. Cirino, G., Fiorucci, S., and Sessa, W.C. 2003.
Endothelial nitric oxide synthase: the Cinderella of
inflammation? Trends Pharmacol. Sci. 24:91–95.
53. Bucci, M., et al. 2000. In vivo delivery of the cave-
olin-1 scaffolding domain inhibits nitric oxide
synthesis and reduces inflammation. Nat. Med.
54. Bucci, M., et al. 2005. Endothelial nitric oxide syn-
thase activation is critical for vascular leakage dur-
ing acute inflammation in vivo. Proc. Natl. Acad. Sci.
U. S. A. 102:904–908.
55. Vo, P.A., Lad, B., Tomlinson, J.A., Francis, S.,
and Ahluwalia, A. 2005. Autoregulatory role of
endothelium-derived nitric oxide (NO) on lipo-
polysaccharide-induced vascular inducible NO
synthase expression and function. J. Biol. Chem.
56. Connelly, L., Madhani, M., and Hobbs, A.J. 2005.
Resistance to endotoxic shock in endothelial nitric-
oxide synthase (eNOS) knock-out mice: a pro-
inflammatory role for eNOS-derived NO in vivo.
J. Biol. Chem. 280:10040–10046.
57. Alloatti, G., et al. 2003. Phosphoinositide 3-kinase
gamma-deficient hearts are protected from the
PAF-dependent depression of cardiac contractility.
Cardiovasc. Res. 60:242–249.
58. Alloatti, G., Montrucchio, G., Lembo, G., and
Hirsch, E. 2004. Phosphoinositide 3-kinase gamma:
kinase-dependent and -independent activities in
cardiovascular function and disease. Biochem. Soc.
59. Laubach, V.E., Shesely, E.G., Smithies, O., and
Sherman, P.A. 1995. Mice lacking inducible nitric
oxide synthase are not resistant to lipopolysac-
charide-induced death. Proc. Natl. Acad. Sci. U. S. A.
60. Shesely, E.G., et al. 1996. Elevated blood pressures
in mice lacking endothelial nitric oxide synthase.
Proc. Natl. Acad. Sci. U. S. A. 93:13176–13181.