Regulation of Adhesion Molecules during
No Acute Effects of Aspirin
BERND JILMA, ANDREW BLANN, THOMAS PERNERSTORFER, PETRA STOHLAWETZ,
HANS-GEORG EICHLER, BARBARA VONDROVEC, JEAN AMIRAL, VOLKER RICHTER,
and OSWALD F. WAGNER
Department of Clinical Pharmacology, The Adhesion Research Group Elaborating Therapeutics, Clinic for Blood Group Serology and
Transfusion Medicine, Department of Transfusion Medicine, University of Vienna, Vienna, Austria; Thrombosis and Vascular Biology Unit,
Department of Medicine, University of Birmingham, Birmingham, United Kingdom; Department of Clinical Chemistry and
Pathobiochemistry, University of Leipzig, Leipzig, Germany; and Hyphen BioMed, Research Organization,
Conflans Sainte Honorine, France
Gram-negative septic shock is mediated in part by endotoxin (lipopolysaccharide; LPS), and animal
models have shown that blockade of even single adhesion molecules considerably improves survival.
Thus interference with the adhesion cascade may provide a useful therapeutic approach in human
sepsis. Young healthy men (n 30) each received a bolus of 4 ng/kg LPS intravenously to study the
effects of endotoxemia on adhesion processes in humans and to identify potential targets for phar-
macologic intervention. One third of subjects received pretreatment with 1,000 mg aspirin and 1,000
mg paracetamol to study potential antiinflammatory effects of aspirin or effects of antipyresis. Circu-
lating neutrophils dropped by 80% at 67 min after LPS, monocytes by 96% at 90 min, and lym-
phocytes by 85% at 240 min. L-selectin expression decreased, particularly on monocytes. Circulat-
ing (c)E-selectin levels increased by 820%, von Willebrand factor-Ag (vWF), soluble thrombomodulin,
circulating (c)P-selectin, circulating intercellular adhesion molecule-1 (cICAM-1), and circulating vas-
cular cell adhesion molecule-1 (cVCAM-1) by a mean of 65 to 98% (p 0.001 for all), but cL-selectin
by only 15%. Urinary excretion of soluble adhesion molecules was negligible. Aspirin had no influ-
ence on the LPS-induced changes of adhesion parameters, but paracetamol blunted the relative in-
crease in vWF while having no effects on the other parameters measured. The consistent, profound,
and early upregulation of cE-selectin during endotoxemia indicates that cE-selectin may be a better
surrogate marker to monitor the activation status of endothelial cells in systemic inflammation than
the other markers measured. Although aspirin did not have any antiinflammatory effects in this
model, paracetamol lowered the relative increase in vWF. Jilma B, Blann A, Pernerstorfer T,
Stohlawetz P, Eichler H-G, Vondrovec B, Amiral J, Richter V, Wagner OF. Regulation of adhe-
sion molecules during human endotoxemia: no acute effects of aspirin.
AM J RESPIR CRIT CARE MED 1999;159:857–863.
The recruitment of circulating blood cells to sites of endothe- LPS strongly stimulates the expression of adhesion mole-
lial activation plays a crucial role in a number of pathophysio- cules on endothelial cells (EC): EC upregulate E-selectin in
logic events, including inflammation, blood coagulation, and response to LPS, which is followed by the release of soluble
the immune response (1). Animal studies indicate that these E-selectin into the bloodstrean in vivo (4, 5). As a conse-
platelet-leukocyte- endothelial interactions are also of utmost quence plasma levels of circulating (c)E-selectin are aug-
importance in the development, propagation, and outcome of mented in sepsis (6, 7). The increase in cE-selectin appears to
sepsis (2). Among other bacterial mediators (3) lipopolysac- depend on the dose of LPS (4), and is determined by the LPS-
charide (LPS, endotoxin) contributes to gram-negative sepsis. enhanced tumor necrosis factor (TNF) production (5, 8).
Interestingly, cE-selectin levels correlate with cardiovascu-
lar compromise and clinical outcome in septic patients (7, 9).
Similarly, levels of cP-selectin (10), circulating intercellular
(Received in original form May 27, 1998 and in revised form October 14, 1998)
adhesion molecule-1 (cICAM-1) (9, 11), and vascular cell ad-
Supported by a grant from the Kamillo-Eisner Stiftung, CH.
hesion molecule-1 (cVCAM-1) are increased in sepsis and
Correspondence and requests for reprints should be addressed to Bernd Jilma, may predict organ dysfunction in septic patients (6, 12, 13).
M.D., Department of Clinical Pharmacology-TARGET, The Adhesion Research
Group Elaborating Therapeutics, Vienna University Hospital School of Medicine,
Conversely, plasma levels of L-selectin, which is continuously
Währinger Gürtel 18-20, A-1090 Wien, Austria. shed from leukocytes into the bloodstream, are lowest in criti-
Am J Respir Crit Care Med Vol 159. pp 857–863, 1999 cally ill patients who progressed to adult respiratory distress
Internet address: www.atsjournals.org syndrome (14). Unfortunately the differences in study designs
858 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 159 1999
between these trials do not allow direct comparison of the
prognostic value of these markers of endothelial, platelet, and
leukocyte activation. Nor do these studies allow conclusions
on the time course and mechanisms of regulation of cAM.
In the current study we aimed to further characterize the
time sequence of the regulation of these cell activation mark-
ers: we therefore infused endotoxin into healthy volunteers;
this procedure presents an established model of human in-
flammation (15). We measured plasma levels of the endothe-
lial cell markers, cE-selectin, von Willebrand factor (vWF) an-
tigen, and of thrombomodulin, the leukocyte activation marker
cL-selectin, cP-selectin, a marker of platelet and/or endothelial
activation, cICAM-1, and cVCAM-1. In addition we deter-
mined differential blood counts and quantified L-selectin ex-
pression on leukocytes and P-selectin expression on platelets.
Secondly, we hypothesized that aspirin could inhibit adhesive
events because it has inhibitory effects on the LPS-induced
nuclear translocation of NF- B in vitro, which transcription-
ally regulates several “immediate early genes” such as TNF
(16). Hence, the effects of aspirin-pretreatment were com-
pared with placebo and with paracetamol. Paracetamol was
used as an active substance to control for potential effects of
inhibiting the febrile response because hyperthermia may en- Figure 1. Neutrophil (top panel) and lymphocyte counts (bottom
panel) observed before and after LPS administration (4 ng/kg).
hance L-selectin-dependent adhesion (17) and may increase
1,000 mg aspirin (closed squares), 1,000 mg paracetamol (open tri-
plasma levels of cICAM-1 (18).
angles), or placebo (open circles) were ingested after the first blood
sample. Data are expressed as means SEM. Friedman’s ANOVA,
METHOD p 0.0001 for all treatment groups. After the initial neutropenia a
second drop in neutrophils was observed (dip at 150 min), which
Study Design and Study Subjects was less pronounced in the paracetamol group.
The study was randomized, double-blind, placebo controlled in three
parallel groups (n 10/group) in 30 healthy male subjects (mainly
students). The study was approved by the Ethics Committee of the Vi- ples were predetermined and based on the results from previous stud-
enna Medical School, and all participants gave written informed con- ies (15, 20–22): plasma was obtained before LPS, at 1, 1.5, 2, 3, 4, 8,
sent prior to enrollment in the study. Mean age of the study subjects and 24 h, and additional samples for differential blood counts were
was 28 5 yr ( SD) and mean body mass index was 24 2.3 kg/m2. drawn at 7-min intervals from 45 to 90 min, at 105 and 150 min (Figure
Health status was determined by a battery of laboratory and clinical 1). Citrated plasma was obtained by centrifugation at 2,000 g for 15
tests, including evaluation of medical history, physical examination, min at 4 C and stored at 80 C until analysis as duplicates, and all
and hematologic, biochemical, virologic, and drug screening as previ- samples from individual subjects were run in the same assay.
ously described (19). Exclusion criteria were hypersensitivity to either Soluble adhesion molecules were determined from citrated plasma
aspirin or paracetamol and regular or recent (within 3 wk) intake of and urine by enzyme immunoassays (R&D Systems, Oxon, UK) as
any drugs, including over the counter medication. described previously (19–23). Intra-assay, day-to-day and within-day
variability have been described previously (19): a small but significant
diurnal decrease of cE-selectin and cP-selectin levels ( 8%) was ob-
Volunteers reported at the study ward at 8:00 A.M. after an overnight served, and the coefficient of variation for the intra-assay, day-to-day,
fast. Throughout the entire study period they had to stay in bed in su- and within-day variability averaged less than 6, 8, and 5%, respec-
pine position and were kept fasting for 8.5 h after endotoxin infusion. tively. Ofloxacin (400 mg) was added to urine containers to avoid bac-
A 5% glucose infusion (Leopold Pharma) was started at 8:30 A.M. and terial degradation of excreted adhesion molecules. Plasma levels of
continued for 8.5 h at 3ml/kg/h to ascertain constant blood glucose soluble thrombomodulin (sTM) were assayed by EIA from Diagnos-
levels and adequate urinary output. Placebo (generously provided by tica Stago (Chausson, Asnieres-sur-Seine, France), and vWF by EIA
Genericon Pharma, Lannach, Austria), 1,000 mg paracetamol (Parac- using polyclonal antisera (Dako-Patts, High Wycombe, UK) as de-
etamol; Genericon), or 1,000 mg acetylsalicylic acid (ASS; Genericon) scribed earlier (24).
were administered orally after start of the glucose infusion. At 9:00 Blood counts. Differential blood counts were performed with a cell
A.M. study subjects received 4 ng/kg LPS (National Reference Endot- counter (Sysmex, Japan). However, microscopic inspection of blood
oxin, Escherichia coli; United States Pharmacopoeial Convention Inc., smears revealed spuriously high monocyte counts at times when neu-
Rockville, MD) as an intravenous bolus infusion for 1 to 2 min. trophilia was present. Hence, monocyte counts were calculated from
Vital parameters were monitored continuously (ECG, heart rate, scatter histograms obtained with the flowcytometer.
and oxygen saturation) or at 20-min intervals (blood pressure) on a Flowcytometric assay. Because all samples required immediate
Care View System (Hewlett Packard, Böblingen, Germany). Tym- processing to avoid artificial activation of leukocytes or platelets, leu-
panic temperature was recorded hourly using an electronic thermome- kocytes were stained only at 30 and 90 min, and at 6 and 24 h. All
ter (Thermoscan, San Diego, CA) and 24 urine samples were collected antibodies except CD14 (Phycoerythrin labeled; DAKO) were fluoro-
before and on the study day. For safety reasons, study subjects had to isothiocyanate labeled and purchased from Immunotech International
stay at the research ward overnight until 24 h after endotoxin infusion. (Marseille, France) (CD62P, CD62L, isotypic IgG control). Flowcyto-
metric analysis was performed as described previously (23, 25, 26).
Sampling and Analysis Twenty thousand and thirty thousand events were gated for leuko-
Venous blood was drawn into vacutainer tubes before administration cytes and platelets, respectively. We did not perform analyses of
of any drug and thereafter, as indicated in the figures, until 8 h after monocytes at 90 min or 6 h after intravenously administered LPS be-
endotoxin injection by repeated venipunctures (except differential cause of the low number of circulating monocytes at 90 min and the
blood counts, which were obtained from an indwelling venous line) on highly variable monocyte counts at 6 h. The parameters of the flow cy-
the arm opposite where LPS had been administered. All blood sam- tometer did not change significantly during the study, as verified by
Jilma, Blann, Pernerstorfer, et al.: LPS and Adhesion in Men 859
means of calibration beads (Quantum Fluorescence kits; Flow Cyto- gether with the maximal neutropenia (Figure 1; p 0.05 be-
metric Standard Corp., San Juan, PR) on a daily basis. tween groups).
Monocyte counts averaged 0.52 G/L (0.45 to 0.52) at baseline
and decreased to 0.01 G/L (CI: 0.007 to 0.012) at 90 min. At 6 h
As hemodilution was only minimal ( 5% at 5 h), data were not ad- monocyte counts ranged from 0.16 to 0.93 G/L, indicating
justed for hemoglobin levels. Data are expressed as the mean and the marked variability in the duration of monocytopenia, and at 24 h
95% confidence intervals and presented as maximal percental changes
they were slightly above baseline (p 0.05 between groups).
for description in the text (n 30). Absolute values are presented in
the figures. Nonparametric statistics were applied. All statistical com- L-Selectin Expression and Shedding of L-Selectin
parisons within groups were done with Friedman’s ANOVA and the
Wilcoxon’s signed ranks test for post hoc comparisons. To test L-selectin neutrophils decreased from a baseline of 96% (CI:
changes in end points between groups for statistical significance the 94 to 97%) to 90% (CI: 87 to 93%) at 6 h. L-selectin mono-
Kruskal-Wallis ANOVA was used and post hoc comparisons were cytes decreased from a baseline of 90% (CI: 88 to 92%) to
performed by the Mann-Whitney U test. A two-tailed p value of 82% (CI: 79 to 85%) at 24 h. The MFI of neutrophils in-
0.05 was considered significant. creased slightly from baseline values of 119 (CI: 108 to 130 ar-
bitrary units) to 138 (CI: 122 to 154) at 90 min and fell to nadir
RESULTS levels of 92 (CI: 82 to 102) at 6 h. This effect was considerably
The usual side effects of endotoxin were observed (e.g., chills, more pronounced on monocytes: 107 (CI: 94 to 119) before in-
fever, headache, nausea, myalgia). Expectedly (27), paraceta- travenously adminstered LPS and 55 (CI: 48 to 62) at 24 h
mol decreased the febrile response (maximum temperature, (Figure 2; p 0.05 between groups).
38.5 C (CI: 38.1 to 38.8) and 37.6 C (CI: 37.2 to 37.9) in the A consistent but small increase in cL-selectin was observed
placebo and paracetamol groups, respectively). Aspirin de- (11%; CI: 7 to 14%), and cL-selectin levels stayed at 7 to 10%
layed the onset of fever by 1 h but it did not decrease the max- above baseline levels throughout the observation period (Fig-
imal increase in temperature seen at 4 h after LPS administra- ure 3; p 0.05 between groups).
tion. No unexpected adverse events occurred.
Specific Markers of Endothelial Activation
Changes in Differential White Blood Cell Counts Soluble thrombomodulin increased by 86% (CI: 58 to 113%)
Neutrophil counts dropped by a maximum of 80% (95% CI at 3 h, and vWF-Ag increased by 98% (CI: 77 to 119%) at 4 h
for the nadir: 77 to 83%) at 67 min after LPS infusion, which (Figure 4). Baseline vWF levels were somewhat higher in the
was followed by a rapid recovery within 40 min (Figure 1). In paracetamol group (p 0.05). The percental increase in the
the placebo and the aspirin groups a second dip of leukocyte paracetamol group was significantly lower than in both other
counts was observed at 150 min, which was rather blunted in groups from 2 to 4 h after LPS administration (p 0.05). The
the paracetamol group. Thereafter, neutrophil counts in- increase in cE-selectin by 795% (CI: 697 to 894% at 8 h) was
creased steadily to reach a plateau at 8 h (Figure 1). considerably more pronounced than that of all other markers
Lymphocyte counts decreased continuously by a maximum (Figure 3; p 0.05 between groups).
of 83% (CI: 81 to 85%) and the nadir occurred at 280 min (CI:
P-Selectin on Platelets and Platelet/
260 to 303 min), although the most rapid fall was observed to- Endothelial-derived cP-Selectin
As expected from a previous study (28) platelet counts de-
creased from baseline values of 234 109/L (CI: 216 to 251)
Figure 2. L-selectin expression on neutrophils (top panel) and on
monocytes (bottom panel) observed before and after LPS adminis-
tration (4 ng/kg). 1,000 mg aspirin (closed squares), 1,000 mg Figure 3. Plasma levels of circulating (c)selectins (ng/ml) observed
paracetamol (open triangles), or placebo (open circles) were in- before and after LPS administration (4 ng/kg). 1,000 mg aspirin
gested after the first blood sample. Data are expressed as means (closed squares), 1,000 mg paracetamol (open triangles), or pla-
SEM. Monocyte counts averaged 0.52 G/L (CI: 0.45 to 0.58) at cebo (open circles) were ingested after the first blood sample. Data
baseline and 0.64 G/L (CI: 0.49 to 0.78) 24 h after LPS infusion. are expressed as means SEM. * p 0.05 versus baseline in all
*p 0.05 versus baseline; * p 0.05 versus baseline in all groups. groups; *p 0.05 versus baseline.
860 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 159 1999
Figure 5. Circulating intercellular (ICAM-1) and vascular cell
(VCAM-1) adhesion molecules (ng/ml) observed before and after
LPS administration (4 ng/kg); 1,000 mg aspirin (closed squares),
Figure 4. Increase in plasma levels of soluble thrombomodulin 1,000 mg paracetamol (open triangles), or placebo (open circles)
(top panel) and increase of von Willebrand antigen (bottom panel) were ingested after the first blood sample. Data are expressed as
after LPS injection; 1,000 mg aspirin (closed squares), 1,000 mg means SEM. * p 0.05 versus baseline in all groups.
paracetamol (open triangles), or placebo (open circles) were in-
gested after the first blood sample. Data are expressed as means
SEM. Baseline vWF levels were somewhat higher in the paraceta-
mol group (p 0.05), possibly accounting for the differences in Plasma levels of TNF increased by 3 log cycles in all groups
the relative increase of vWF-Ag (% vWF) (middle panel). **p (p 0.001): peak plasma levels were reached between 90 and
0.05 paracetamol versus aspirin, and p 0.01 paracetamol versus 120 min and averaged 4.1 ng/ml (range: 0.6 to 13.3) in the pla-
placebo; * p 0.05 versus baseline in all groups; *p 0.05 versus cebo group, 3.0 ng/ml (range: 0.6 to 6.8) in the paracetamol
baseline. group, and 4.7 ng/ml (range: 0.6 to 24.6) in the aspirin group
(p 0.05 between groups).
by 14% (CI: 17 to 12%) at 75 min after LPS infusion and DISCUSSION
stayed at those levels until 6 h after intravenously adminis-
tered LPS. P-selectin platelets did not increase above base- Although animal models provide evidence that adhesion mol-
line levels 0.6% (CI: 0.3 to 0.9%) at 2, 4, 6, or 24 h (p 0.05) ecules contribute to the pathogenesis of sepsis (2, 29), data on
and even decreased slightly at 2 h (0.4% CI: 0.2 to 0.6%; data the LPS-induced regulation of adhesion events in humans are
not shown). cP-selectin levels increased by 97% (73 to 122%) scarce. Hence, we have used LPS-infusions as a model to elicit
and mean cP-selectin levels of the aspirin-treated subjects al- a systemic inflammatory reaction in humans (15).
ways remained slightly above the other two groups. (Figure 3; In contrast to previous LPS trials, the more frequent sam-
p 0.05 between groups). pling in our study provided a better estimate of an 80% drop
in neutrophils (Figure 1). As LPS-infusion precipitates an
Soluble Adhesion Molecules of the early P-selectin-dependent neutropenia in rats (30), it would be
Immunoglobulin Superfamily tempting to study the pharmacodynamics of P-selectin antago-
nists in this LPS-model. Whereas the neutropenia was rapidly
cICAM-1 increased by 99% (CI: 86 to 112%) at 8 h and cV-
reversible, the concomitant lymphocytopenia was more sus-
CAM-1 by 73% (CI: 66 to 80%) at 24 h (Figure 5; p 0.05 be-
tained (Figure 1). As L-selectin mediates lymphocyte homing,
an LPS-induced upregulation of ligands for L-selectin could be
responsible for this effect. In addition, L-selectin is not only a
Urinary Excretion of Circulating
bystander mediating rolling of leukocytes but may be a signal-
ing receptor for LPS also (31). Hence, its regulation is of inter-
As indicated by the data in Table 1, soluble P-selectin was not est: only a small decrease in L-selectin leukocytes occurred
detectable in urine, and urine levels of the other soluble adhe- similar to a previous study (4). However, there was a substan-
sion molecules were also very low: when considering the esti- tial decrease in the MFI, particularly of monocytes (Figure 2).
mated plasma volume in our subjects (about 3,000 ml), less As the MFI of L-selectin on neutrophils inversely correlated
than 6% of cVCAM-1 was recovered in 24-h urine samples with neutrophilia during inflammation (32), it has been sug-
and substantially less for the other cAM. gested that downregulation of L-selectin by shedding might in
part explain the appearance of neutrophilia during inflamma-
Salicylate Levels and TNF Levels tion. However, there was only a 10% increase in cL-selectin in
Salicylate levels averaged 0.21 mM/L (CI: 0.13 to 0.28) at 60 plasma (Figure 3). The marginal increase in cL-selectin as
min and stayed at that level until 180 min (0.22 mM/L; CI: 0.16 compared with the other markers may be due to the high
to 0.27) after ingestion of aspirin. background noise caused by continuous shedding of cL-selec-
Jilma, Blann, Pernerstorfer, et al.: LPS and Adhesion in Men 861
CONCENTRATION OF SOLUBLE ADHESION MOLECULES IN 24-h URINE SAMPLES*
Before LPS Study Day p Values
Creatinine clearance, ml/min 141 (94–249) 139 (94–196) NS
Urine volume, ml 1,413 (580–2,300) 2,824 (1,520–3,600) p 0.001
cP-selectin, ng/ml 0.3 0.3
cE-selectin, ng/ml 0.2 ( 0.1–1.0) 0.4 ( 0.1–2.2) NS
cL-selectin, ng/ml 0.8 ( 0.3–4.6) 3.5 ( 0.3–20.7) NS
cICAM-1, ng/ml 2.9 (0.7–6.6) 6.4 (1.0–14.8) p 0.001
cVCAM-1, ng/ml 29 (2–106) 26 (3–116) NS
Definition of abbreviations: c circulating; ICAM-1 intercellular cell adhesion molecule-1; LPS lipopolysaccharide; VCAM-1 vascu-
lar cell adhesion molecule-1.
* Data are presented as means with ranges shown in parentheses. Recovery rates of standards added to urine samples ranged from 85 to
tin. Hence, cL-selectin, in contrast to neutrophil elastase (28), face-expressed activation markers because of the time lag be-
is a poor marker to detect LPS-induced neutrophil activation tween maximal expression and shedding. Yet the observation
in vivo. that cICAM-1 and cVCAM-1 levels reached a plateau at 24 h
In contrast to cL-selectin, there was an outstanding eight- (Figure 4) is in good agreement with the later upregulation of
fold increase in cE-selectin (Figure 3). This agrees with an ani- these molecules as compared with E-selectin.
mal study demonstrating that E-selectin expression is par- Another issue that merits discussion is the increase in cP-
ticularly sensitive to LPS-stimulation (33). A clinical report selectin (Figure 3), which was not accompanied by an increase
recently showed that cE-selectin levels measured on Day 1 af- in circulating P-selectin platelets. Hence, it is questionable if
ter admission to the intensive care unit predicted outcome in determination of P-selectin expression on platelets is of diag-
critically ill patients (7). As in previous studies (4, 5, 8), there nostic value in patients with septicemia. This can possibly be
was no overlap in plasma E-selectin levels of baseline samples explained by margination of activated platelets and concomi-
and those obtained at 4 to 8 h (Figure 3). Hence, compared tant shedding of P-selectin from the platelet surface into the
with the other markers of endothelial cell activation measured bloodstream (37, 38).
(vide infra), cE-selectin levels could be particularly useful to Finally, our data show that neither paracetamol nor aspirin
monitor endothelial cell activation during the early course of had an acute effect on the LPS-induced adhesion events. This
severe sepsis (9). is in contrast to the cyclooxygenase inhibitor ibuprofen (39),
Plasma levels of vWF, another specific endothelial marker, which enhances LPS-induced TNF levels. It may be worth-
doubled within 4 h (Figure 5), which is in general agreement while to investigate the reason for this difference between as-
with previous studies (28, 34). The relative increase in vWF pirin and ibuprofen in subsequent trials. Recently, it was
was lowest in the paracetamol group (Figure 5). Interestingly, shown that aspirin concentration-dependently inhibits the
the increase in vWF is at least in part mediated by TNF (28). LPS-induced nuclear translocation of NF- B in vitro (16, 40).
However, paracetamol did not significantly suppress TNF lev- NF- B transcriptionally regulates several “immediate early
els. Further, it seems unlikely that paracetamol had an effect genes” such as TNF or E-selectin. This does not appear to
on endotoxin signaling because vWF was the only molecule translate into a measurable effect on any downstream events
affected by pretreatment with paracetamol. It is of note that in vivo because E-selectin shedding and neutrophilia, which
basal vWF levels were somewhat higher in the paracetamol are critically dependent on TNF (5), were unaffected by ad-
group. This could be due to continuous vWF release from its ministration of 1,000 mg aspirin. This could be explained by
intraendothelial storage site, and those subjects may therefore the average plasma salicylate levels of 0.2 mM/L in our study,
secrete less vWF in response to an inflammatory stimulus such which indicate that the IC50 of salicylate (1.5 mM in vitro) (16)
as LPS. Alternatively, muscular exercise is known to increase was not reached in vivo. Yet, aspirin was administered at the
vWF antigen levels (35), and the muscular work needed to in- maximum single dose approved. The usual dose for antipyre-
crease body temperature to maximal levels may have been sis is 300 to 600 mg and for analgesia 300 to 900 mg, every 4 to
greater in the placebo group than in the paracetamol group. 6 h, with a maximum daily dose of 4 g. However, it is conceiv-
Hence, the relatively lower increase in vWF may also be at- able that repeated aspirin intake, which leads to an accumula-
tributable to less muscle rigor. Finally, vWF is a well-known tion of aspirin, may yield different results. This must be re-
acute-phase reactant (36). Although paracetamol is known to garded as a limitation of our study.
have only weak antiinflammatory potential, it is conceivable Finally, we have to address an important issue: while endo-
that a yet unidentified antiinflammatory action of paraceta- toxemia may provide insights into gram-negative septic shock,
mol selectively dampened the increase in vWF. there may be significant differences between clinical septic
Like vWF, plasma levels of sTM increased rapidly and shock and experimental endotoxemia. In particular LPS is a
reached maximal levels at 3 h. Thrombomodulin is downregu- noninfectious mediator, with self-limited consequences in our
lated rather than upregulated on endothelial cells by cytokines setting. Hence, it may be very difficult to extrapolate results
in vitro (33). Thus, sTM seems to be shed into plasma either from studies such as the present one to the clinical situation.
by endothelial sloughing or proteolytic cleavage from the en- In conclusion, the profound LPS-induced leukopenia could
dothelial cell surface. It is interesting that the kinetics of sTM provide a suitable surrogate end point for testing the efficacy
were faster (Figure 5) than for the other markers measured, of antiadhesive drugs. The consistent, profound, and early up-
further supporting the different regulation, i.e. downregula- regulation of cE-selectin during endotoxemia indicates that
tion rather than upregulation. Unfortunately, determination cE-selectin may be a better suitable surrogate end point to
of plasma levels does not allow for direct quantification of sur- monitor the activation status of endothelial cells in systemic
862 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 159 1999
inflammation than the other markers measured. Although as- 15:1–8.
pirin did not have any antiinflammatory effects in this model, 19. Jilma, B., T. Szalay, E. Dirnberger, H. G. Eichler, P. Stohlawetz, I.
Schwarzinger, S. Kapiotis, and O. F. Wagner. 1997. Effect of endothe-
paracetamol lowered the relative increase in vWF.
lin-1 on circulating adhesion molecules in man. Eur. J. Clin. Invest. 27:
Acknowledgment : The writers are indebted to Inge Wallner, RN, and Alex- 850–856.
andra Rauch, RN, for their excellent assistance. 20. Bloom, J. N., A. F. Suffredini, J. E. Parrillo, and A. C. Palestine. 1990.
Serum neopterin levels following intravenous endotoxin administra-
tion to normal humans. Immunobiology 181:317–323.
References 21. Pajkrt, D., A. Manten, T. van der Poll, M. M. Tiel van Buul, J. Jansen, J.
1. Celi, A., R. Lorenzet, B. Furie, and B. C. Furie. 1998. Platelet-leukocyte- Wouter ten Cate, and S. J. van Deventer. 1997. Modulation of cyto-
endothelial cell interaction on the blood vessel wall. Semin. Hematol. kine release and neutrophil function by granulocyte colony-stimulat-
34:327–335. ing factor during endotoxemia in humans. Blood 90:1415–1424.
2. Tedder, T. F., D. A. Steeber, and P. Pizcueta. 1995. L-selectin-deficient 22. Suffredini, A. F., J. H. Shelhamer, R. D. Neumann, M. Brenner, R. J.
mice have impaired leukocyte recruitment into inflammatory sites. J. Baltaro, and J. E. Parrillo. 1992. Pulmonary and oxygen transport ef-
Exp. Med. 181:2259–2264. fects of intravenously administered endotoxin in normal humans. Am.
3. Danner, R. L., C. Natanson, R. J. Elin, J. M. Hosseini, S. Banks, T. J. Rev. Respir. Dis. 145:1398–1403.
MacVittie, and J. E. Parrillo. 1990. Pseudomonas aeruginosa com- 23. Jilma, B., J. Voltmann, S. Albinni, P. Stohlawetz, I. Schwarzinger, C.
pared with Escherichia coli produces less endotoxemia but more car- Gleiter, A. Rauch, H. G. Eichler, and O. F. Wagner. 1997. Dexameth-
diovascular dysfunction and mortality in a canine model of septic asone down-regulates the expression of L-selectin on the surface of
shock. Chest 98:1480–1487. neutrophils and lymphocytes in humans. Clin. Pharmacol. Ther. 62:
4. Kuhns, D. B., W. G. Alvord, and J. I. Gallin. 1995. Increased circulating 562–568.
cytokines, cytokine antagonists, and E-selectin after intravenous ad- 24. Blann, A. D., G. Y. Lip, D. G. Beevers, and C. N. McCollum. 1997. Solu-
ministration of endotoxin in humans. J. Infect. Dis. 171:145–152. ble P-selectin in atherosclerosis: a comparison with endothelial cell
5. van der Poll, T., S. M. Coyle, M. Levi, P. M. Jansen, M. Dentener, K. and platelet markers. Thromb. Haemost. 77:1077–1080.
Barbosa, W. A. Buurman, C. E. Hack, J. W. ten Cate, J. M. Agosti, 25. Pernerstorfer, T., P. Stohlawetz, G. Stummvoll, S. Kapiotis, T. Szekeres,
and S. F. Lowry. 1997. Effect of a recombinant dimeric tumor necrosis H. G. Eichler, and B. Jilma. 1998. Low dose aspirin does not prevent
factor receptor on inflammatory responses to intravenous endotoxin platelet activation in healthy smokers. Br. J. Haematol. 102:1229–1239.
in normal humans. Blood 89:3727–3734. 26. Stohlawetz, P., N. Hergovich, G. Stiegler, H. G. Eichler, S. Kapiotis, P.
6. Cowley, H. C., D. Heney, A. J. Gearing, I. Hemingway, and N. R. Web- Höcker, and B. Jilma. 1998. Differential induction of P-selectin ex-
ster. 1994. Increased circulating adhesion molecule concentrations in pression on platelets by two cell separators during plateletpheresis
patients with the systemic inflammatory response syndrome: a pro- and gender effects on the release of soluble P-selectin. Transfusion 38:
spective cohort study. Crit. Care Med. 22:651–657. 24–30.
7. Cummings, C. J., C. N. Sessler, L. D. Beall, B. J. Fisher, A. M. Best, and 27. Vargas, R., T. Maneatis, L. Bynum, C. Peterson, and F. G. McMahon.
A. A. Fowler III. 1997. Soluble E-selectin levels in sepsis and critical 1994. Evaluation of the antipyretic effect of ketorolac, acetamino-
illness: correlation with infection and hemodynamic dysfunction. Am. phen, and placebo in endotoxin-induced fever. J. Clin. Pharmacol. 34:
J. Respir. Crit. Care Med. 156:431–437. 848–853.
8. Suffredini, A. F., D. Reda, S. M. Banks, M. Tropea, J. M. Agosti, and R. 28. DeLa Cadena, R. A., A. Majluf-Cruz, A. Stadnicki, M. Tropea, D. Reda,
Miller. 1995. Effects of recombinant dimeric TNF receptor on human J. M. Agosti, R. W. Colman, and A. F. Suffredini. 1998. Recombinant
inflammatory responses following intravenous endotoxin administra- tumor necrosis factor receptor p75 fusion protein (TNFR:Fc) alters
tion. J. Immunol. 155:5038–5045. endotoxin-induced activation of the kinin, fibrinolytic and coagulation
9. Kayal, S., J. P. Jais, J. Chaudiere, and J. Labrousse. 1998. Elevated circu- systems in normal humans. Thromb. Haemost. 114–118.
lating E-selectin, intercellular adhesion molecule-1, and von Wille- 29. Ridings, P. C., A. C. Windsor, M. A. Jutila, C. R. Blocher, B. J. Fisher,
brand factor in patients with severe infection. Am. J. Respir. Crit. Care M. M. Sholley, H. J. Sugerman, and A. A. Fowler, III. 1995. A dual-
Med. 776–784. binding antibody to E- and L-selectin attenuates sepsis-induced lung
10. Fijnheer, R., C. J. Frijns, J. Korteweg, H. Rommes, J. H. Peters, J. J. injury. Am. J. Respir. Crit. Care Med. 152:247–253.
Sixma, and H. K. Nieuwenhuis. 1997. The origin of P-selectin as a cir- 30. Coughlan, A. F., H. Hau, L. C. Dunlop, M. C. Berndt, and W. W. Han-
culating plasma protein. Thromb. Haemost. 77:1081–1085. cock. 1994. P-selectin and platelet-activating factor mediate initial en-
11. Moss, M., M. K. Gillespie, L. Ackerson, F. A. Moore, E. E. Moore, and dotoxin-induced neutropenia. J. Exp. Med. 179:329–334.
P. E. Parsons. 1996. Endothelial cell activity varies in patients at risk 31. Bird, M. I., M. R. Foster, R. Priest, and R. Malhotra. 1997. Selectins:
for the adult respiratory distress syndrome. Crit. Care Med. 24:1782– physiological and pathophysiological roles. Biochem. Soc. Transact.
12. Sessler, C. N., A. C. Windsor, M. Schwartz, L. Watson, B. J. Fisher, H. J. 32. Rogowski, O., Y. Sasson, M. Kassirer, D. Zeltser, N. Maharshak, N. Ar-
Sugerman, and A. A. Fowler III. 1995. Circulating ICAM-1 is in- ber, P. Halperin, J. Serrov, P. Sorkin, A. Eldor, and S. Berliner. 1998.
creased in septic shock. Am. J. Respir. Crit. Care Med. 151:1420–1427. Down-regulation of the CD62L antigen as a possible mechanism for
13. Sakamaki, F., A. Ishizaka, M. Handa, S. Fujishima, T. Urano, K. neutrophilia during inflammation. Br. J. Haematol. 101:666–669.
Sayama, H. Nakamura, M. Kanazawa, T. Kawashiro, M. Katayama, 33. Drake, T. A., J. Cheng, A. Chang, and F. B. Taylor. 1993. Expression of
and Y. Ikeda. 1995. Soluble form of P-selectin in plasma is elevated in tissue factor, thrombomodulin, and E-selectin in baboons with lethal
acute lung injury. Am. J. Respir. Crit. Care Med. 151:1821–1826. Escherichia coli sepsis. Am. J. Pathol. 142:1458–1470.
14. Donnelly, S. C., C. Haslett, I. Dransfield, C. E. Robertson, D. C. Carter, 34. Gralnick, H. R., L. P. McKeown, O. M. Wilson, S. B. Williams, and R. J.
J. A. Ross, I. S. Grant, and T. F. Tedder. 1994. Role of selectins in de- Elin. 1989. von Willebrand factor release induced by endotoxin. J.
velopment of adult respiratory distress syndrome. Lancet 344:215– Lab. Clin. Med. 113:118–122.
219. 35. Jilma, B., E. Dirnberger, H. G. Eichler, B. Matulla, L. Schmetterer, S.
15. Martich, G. D., A. J. Boujoukos, and A. F. Suffredini. 1993. Response of Kapiotis, W. Speiser, and O. F. Wagner. 1997. Partial blockade of ni-
man to endotoxin. Immunobiology 187:403–416. tric oxide synthase blunts the exercise-induced increase of von Wille-
16. Osnes, L. T. N., K. B. Foss, G. B. Joo, C. Okkenhaug, A. B. Westvik, R. brand factor antigen and of factor VIII in man. Thromb. Haemost. 78:
Ovstebo, and P. Kierulf. 1996. Acetylsalicylic acid and sodium salicy- 1268–1271.
late inhibit LPS-induced NF- B/c-Rel nuclear translocation, and syn- 36. Pottinger, B. E., R. C. Read, E. M. Paleolog, P. G. Higgins, and J. D.
thesis of tissue factor (TF) and tumor necrosis factor alpha (TNF- ) in Person. 1989. von Willebrand factor is an acute phase reactant in man.
human monocytes. Thromb. Haemost. 76:970–976. Thromb. Res. 53:387–394.
17. Wang, W. C., L. M. Goldman, D. M. Schleider, M. M. Appenheimer, J. R. 37. Michelson, A. D., M. R. Barnard, H. B. Hechtman, H. MacGregor, R. J.
Subjeck, E. A. Repasky, and S. S. Evans. 1998. Fever-range hyper- Connolly, J. Loscalzo, and C. R. Valeri. 1996. In vivo tracking of plate-
thermia enhances L-selectin-dependent adhesion of lymphocytes to lets: circulating degranulated platelets rapidly lose surface P-selectin
vascular endothelium. J. Immunol. 160:961–969. but continue to circulate and function. Proc. Natl. Acad. Sci. U.S.A.
18. Nakayama, J., T. Nakao, T. Mashino, S. Nagae, Y. Hori, H. Mayumi, R. 93:11877–11882.
Tominaga, H. Yasui, K. Matsuda, and S. Takahashi. 1997. Kinetics of 38. Stohlawetz, P., M. Horwath, H. Nguyen, B. Vondrovec, A. Robisch, H. G.
immunological parameters in patients with malignant melanoma Eichler, S. Kapiotis, and B. Jilma. 1999. Effects of nitric oxide on
treated with hyperthermic isolated limb perfusion. J. Dermatol. Sci. platelet activation during plateletpheresis, and in-vivo tracking of bi-
Jilma, Blann, Pernerstorfer, et al.: LPS and Adhesion in Men 863
otinylated platelets in humans. Transfusion (In press) mia. J. Infect. Dis. 163:89–95.
39. Spinas, G. A., D. Bloesch, U. Keller, W. Zimmerli, and S. Cammisuli. 40. Weber, C., W. Erl, A. Pietsch, and P. C. Weber. 1995. Aspirin inhibits
1991. Pretreatment with ibuprofen augments circulating tumor necro- nuclear factor-kappa B mobilization and monocyte adhesion in stimu-
sis factor-alpha, interleukin-6, and elastase during acute endotoxine- lated human endothelial cells. Circulation 91:1914–1917.