| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vol. 83, Issue 5, 1467-1475, 1997
Critical Care Medicine Department, National Institutes of Health, Bethesda, Maryland 20892-1662
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
ACKNOWLEDGEMENTS
FOOTNOTES
REFERENCES
ABSTRACT 
Freeman, Bradley D., Zenaide Quezado, Fabrice Zeni, Charles
Natanson, Robert L. Danner, Steven Banks, Marcello Quezado, Yvonne
Fitz, John Bacher, and Peter Q. Eichacker. rG-CSF reduces endotoxemia and improves survival during E. coli pneumonia. J. Appl.
Physiol. 83(5): 1467-1475, 1997.
We investigated
the effects of recombinant granulocyte colony-stimulating factor
(rG-CSF) during canine bacterial pneumonia. Beagles with chronic
tracheostomies received daily subcutaneous rG-CSF (5 µg/kg body wt)
or placebo for 14 days, beginning 9 days before intrabronchial
inoculation with E. coli. Animals
received antibiotics and fluid support; a subset received humidified
oxygen (fractional inspired O2
0.40). Compared with controls, rG-CSF increased circulating neutrophil counts (57.4 vs. 11.0 × 103/mm3,
day 1 after infection;
P = 0.0001), decreased plasma
endotoxin (7.5 vs. 1.1 EU/ml at 8 h; P < 0.01) and serum tumor necrosis factor-
(3,402 vs.
729 pg/ml at 2 h; P = 0.01) levels,
and prolonged survival (relative risk of death = 0.45, 95% confidence
interval 0.21-0.97; P = 0.038).
Also, rG-CSF attenuated sepsis-associated myocardial dysfunction
(P < 0.001). rG-CSF had no effect on
pulmonary function or on blood and lung bacteria counts (all
P = not significant). Other animals
challenged with endotoxin (4 mg/kg iv) after similar treatment with
rG-CSF had lower serum endotoxin levels (7.62 vs. 5.81 log EU/ml at 6 h; P < 0.01) and less cardiovascular
dysfunction (P < 0.05 to < 0.002)
but similar tumor necrosis factor- levels (P = not significant) compared with
controls. Thus prophylactic rG-CSF sufficient to increase circulating
neutrophils during bacterial pneumonia may improve cardiovascular
function and survival by mechanisms that in part enhance the clearance
of bacterial toxins but do not improve lung function.
neutrophil; endotoxin; recombinant granulocyte colony-stimulating
factor; sepsis; septic shock; Escherichia
coli
INTRODUCTION SEPSIS AND SEPTIC SHOCK are major causes of morbidity
and mortality in hospitalized patients in the U.S. Recent basic and clinical research has focused on inhibition of inflammation as a
therapeutic strategy (29). However, clinical trials that examined anti-inflammatory therapies in septic patients have failed to demonstrate convincing efficacy (12). As one alternative, interest has
recently centered on proinflammatory agents capable of augmenting host
defense. Recombinant granulocyte colony-stimulating factor (rG-CSF) is
one such proinflammatory agent that stimulates both circulating
neutrophil number and function.
Initial studies in immunocompromised animal models and patients, and
subsequently in immunocompetent animal models, showed that rG-CSF when
administered prophylactically reduced the risk and morbidity of
infection (3, 14, 20, 22, 25, 31, 32, 40). However, when rG-CSF was
administered therapeutically in a phase III clinical trial of
non-neutropenic patients with community-acquired pneumonia, it did not
result in any statistically significant beneficial clinical effect
(30). One interpretation of this combined experience is that rG-CSF may
be most efficacious in the immunocompetent host if it is used
prophylactically to prepare the subject who is at risk of infection or
sepsis rather than if it is used therapeutically. In a canine model of
Escherichia coli peritonitis, we have
demonstrated very different results with prophylactic vs. therapeutic
rG-CSF. Prophylactic rG-CSF enhanced endotoxin clearance, lowered tumor
necrosis factor- However, just as the prophylactic use of rG-CSF may maximize its
beneficial host-defense effects, such use may also worsen inflammatory
tissue injury. Concern exists that prophylactic rG-CSF in
non-neutropenic patients might increase lung injury in those who
ultimately develop pneumonia. A strong association exists between neutrophil activation and the pathogenesis of
inflammatory lung injury (37). In some animal models, prophylactic
rG-CSF has been clearly shown to aggravate such lung injury. In rats challenged with intrabronchial E. coli
or high O2 concentrations, we
found that rG-CSF worsened lung injury and reduced survival (18).
Therefore, increasing inflammation with rG-CSF at sites of infection,
particularly in the lung, could negate its beneficial effects on host
defense and survival.
We have developed a canine model of gram-negative pneumonia; the model
includes antibiotics, fluid, and
O2 as standard supportive therapies (35). This model simulates many of the cardiopulmonary changes that occur during gram-negative pneumonia in humans. Using this
large-animal model, we sought to better clarify the effects of
prophylactic rG-CSF administration on endotoxemia, cytokine release,
bacteremia, cardiopulmonary function, and survival in non-neutropenic
hosts during gram-negative pneumonia. To better clarify possible
mechanisms underlying the effects of rG-CSF on host defense that were
observed in our pneumonia model, we then did additional studies in
animals challenged with endotoxin alone.
MATERIALS AND METHODS
(TNF-) levels, and improved cardiovascular
function and survival, whereas therapeutic rG-CSF (i.e., administered
after the onset of infection) did not improve outcome and at very high
dosages appeared harmful (36). Thus rG-CSF may be most efficacious for host defense if used prophylactically in patients at risk of infection.
levels as well as to perform
liver-function tests and quantitative blood cultures.
Furthermore, quantitative lung cultures were obtained by a
protected-brush technique. These studies were repeated 1, 2, and 14 days after bacterial inoculation. In addition, to assess the acute
effects of infectious challenge, blood analysis was also performed
immediately before and at 2, 4, 6, and 8 h after bacterial inoculation.
Catheters were removed at the end of each study day.
Nine days before bacterial inoculation, animals began a 14-day course
of either rG-CSF (5 µg · kg
1 · day1,
n = 28) or control protein (5 µg · kg1 · day1,
n = 28) injected subcutaneously. An
rG-CSF dose of 5 µg · kg1 · day1
was used because this dose increased circulating neutrophil
concentrations and improved survival in our canine peritonitis model
(14). On day 0, animals were
intrabronchially inoculated with E. coli [5.0 × 109 colony-forming units (CFU)/kg
(n = 8), 7.0 × 109 CFU/kg
(n = 4), 7.5 × 109 CFU/kg
(n = 16), 10 × 109 CFU/kg
(n = 18), or 15 × 109 CFU/kg
(n = 10)] in either a diffuse
distribution (inoculum divided equally among four lobes,
n = 22) or a lobar distribution
(inoculation of right lower lobe, n = 34). These techniques have been previously described.
Animals received a continuous infusion of Ringer solution (10 ml · kg1 · h1)
as hemodynamic support beginning 6 h after inoculation and continuing for 8 h. In addition, all animals received antibiotics (ceftriaxone, 100 mg · kg1 · day1
iv) beginning 6 h after bacterial inoculation and continuing for 5 days. Furthermore, a subset of animals [inoculated with either a
diffuse distribution of 5.0 × 109 CFU/kg
(n = 8) or a lobar distribution of 7.5 × 109 CFU/kg
(n = 16)] was placed into
O2 chambers (Plas Labs, Lansing, MI) and exposed to a fraction of inspired
O2 of 40%, beginning 6 h after
bacterial inoculation and continuing for 72 h. Animals in this subset
were removed from their chambers on days
1 and 2 after
bacterial inoculation and had all measurements performed after
equilibration with room air for 30 min. After these measurements were
made, these animals were returned to their chambers. All remaining
animals were considered survivors 14 days after bacterial inoculation;
they were then killed.
Techniques for studies of pneumonia caused by inoculation with
bacteria.
After the intrabronchial instillation of 1% lidocaine (10 ml), diffuse
pneumonia was produced by placing a 7.5-Fr balloon flotation catheter
(Edwards Laboratories, Santa Ana, CA) under bronchoscopic guidance
(BF1TR; Olympus, New Hyde Park, NY) sequentially into the left upper,
left lower, right upper, and right lower lobes, where the balloon was
inflated and 8 ml of bacterial solution were instilled in four equal
aliquots. To produce lobar pneumonia, the bacterial
solution was similarly instilled, but entirely into the right lower
lobe. During each week of this experiment, rG-CSF-treated animals were
studied concurrently with an equal number of controls. These concurrent
control animals were treated in a fashion identical to the treatment
group with the exception of the study drug received (i.e., control
protein).
Experimental design for endotoxin-challenge studies.
The effects of pretreatment with rG-CSF alone or in combination with a
murine MAb against canine CD11/18 neutrophil adhesion complex (MAb
R15.7) during endotoxin challenge were studied over 6 h in anesthetized
and mechanically ventilated canines. This was done to see whether
rG-CSF alone can improve endotoxin clearance independent of increasing
bacterial killing and to determine the importance, if any, of
neutrophil adhesion complex in endotoxin clearance. The study was done
over 6 h because endotoxin clearance, the primary endpoint of the
study, largely occurs within this time period. Starting 9 days before a
30-min endotoxin infusion, 14 dogs received daily subcutaneous
injections of either rG-CSF (5 µg/kg,
n = 7) or control protein (canine
serum protein, n = 7). Three hours
before the endotoxin infusion, 10 dogs received a single iv injection
of MAb R15.7 (1 mg/kg, Boehringer-Ingelheim, Ridgefield, CO;
n = 5) or control protein (murine
serum protein, 1 mg/kg; n = 5). Ten
dogs were treated simultaneously with both the rG-CSF and MAb R15.7
protocols (n = 5) or control proteins only (n = 5).
Ninety minutes before endotoxin challenge, all animals were
anesthetized (isoflurane, 3 minimum alveolar concentration
for mask induction and 0.5 minimum alveolar concentration for
maintenance), paralyzed (with the use of vecuronium, 6 µg · kg1 · h1),
intubated, and mechanically ventilated. Femoral and pulmonary arterial
thermodilution catheters were placed percutaneously. Blood analysis,
hemodynamics, and radionuclide cineangiographic studies were then
performed as in the pneumonia study. After baseline measurements were
made, all animals were challenged with endotoxin (E. coli 0111:B4, 4 mg/kg over 30 min iv). Blood analysis
was then repeated immediately after (time
0) and 30, 60, 90, 120, 150, 180, 240, 300, and 360 min after endotoxin challenge, whereas hemodynamics were measured at
60, 120, 240, and 360 min after endotoxin challenge. During the study,
animals received Ringer lactate solution iv (20 ml/kg bolus followed by
20 ml · kg1 · h1).
All animals were killed at 360 min after all hemodynamic and laboratory
studies were completed.
Laboratory analysis techniques for pneumonia and endotoxin studies.
Methods of laboratory analysis have been described previously (14, 28).
Briefly, quantitative blood and bronchial cultures were collected in
1.5 ml isolator tubes with serial dilutions of lysed samples plated for
bacterial colony counts. Serum and whole blood were analyzed by
standard automated methods (MetPath MidAtlantic Regional Laboratory,
Rockville, MD). Arterial and mixed venous blood gases were determined
by using an automated system (Ciba-Corning Diagnostic, Medfield, MA).
Plasma endotoxin concentrations were determined by using a kinetic,
chromogenic limulus lysate assay (MA Bioproducts,
Walkersville, MD) (6). Serum TNF- bioactivity was measured by using
the WEHI 164 cell line and assay (16).
Cardiopulmonary evaluation for pneumonia and endotoxin studies.
Hemodynamic measurements, including determinations of systemic and
pulmonary arterial pressure, thermodilution cardiac output, and
radionuclide LVEF, were performed in awake, nonsedated animals as
previously described (10, 28). In addition, cardiac index; left ventricular end-diastolic, end-systolic, and stroke volume indexes
(LVEDVI, LVESVI, and LVSVI, repectively); left and right ventricular
stroke work indexes (LVSWI and RVSWI, respectively); alveolar-to-arterial O2
(AaO2) gradient; shunt fraction
(QS/QT); and systemic O2 delivery
(DO2)
were calculated by standard methods (10, 28).
Animal care.
This study protocol was performed in accordance with the
guidelines published by the National Institutes of Health (21) and was approved by the Animal Care and Use Committee of the National Institutes of Health Clinical Center. This protocol required the veterinary staff to kill any animal that experienced unexpected pain or
distress. During the pneumonia studies, animals had free access to food
and water. Every effort was made to minimize animal suffering.
Statistical methods for pneumonia and endotoxin studies.
Survival data for animals challenged with intrabronchial
E. coli were analyzed for
rG-CSF-treatment effects by using a Cox Proportional Hazards Model (5).
The Cox hazard model showed no difference in rG-CSF treatment effects
between doses of bacteria, presence or absence of supplemental
O2, or diffuse and lobar
pneumonia, so data were pooled across these various conditions to
increase statistical power. Relative risk and the 95% confidence
interval are reported.
For hemodynamic, pulmonary, and laboratory parameters, an analysis of
variance (ANOVA) (38) was performed. At baseline, a one-way ANOVA was
used to demonstrate that no significant differences existed between
control and rG-CSF animals. A four-way ANOVA was performed, with
effects for treatment, dog (nested within treatment), time, and fluid
as the main effects. This is the univariate version of the
repeated-measures ANOVA as described by Cole and Grizzle (4). In
addition, two- and three-way interactions were included in the model,
with primary attention given to the treatment-time interaction.
Higher-order interactions that included dog were pooled to form the
error term for the ANOVA. Although the data are reported from the
analysis of a four-way ANOVA, we actually carried out an analysis with
the use of a seven-way ANOVA, including type of pneumonia, dose of
bacteria, and presence of supplemental O2. In this seven-way model, we
included higher-order interaction terms to investigate whether
treatment-time interactions were altered by these three additional
factors. We found no significant higher-order interactions, including
treatment-time interactions, and have thus reported the results from
the four-way ANOVA. Frank-Starling left ventricular (LV) function data
were analyzed by using a multiple ANOVA procedure. For one parameter
(AaO2), an observation was found with
the use of a test by Dixon (8) to be an outlier, and it was
removed. The pooled sources of variability from the ANOVA
are presented as the measure of variability in the data (see Figs.
2, 3, 4). The pooled source of variability is in the form of a SE of the mean, where the estimate of the SD is the root mean
squared from the ANOVA, which is then divided by the square root of the
number of observations that are in the plotted means.
) from baseline calculated from changes in
individual animals, in %left ventricular ejection fraction (LVEF) on
days 1 (A) and
2 (B) after bacterial inoculation in animals receiving rG-CSF or placebo protein (controls) before and after
volume infusion. Relationships are shown pre- (origin of arrow) and
postvolume (tip of arrow) infusion. Pooled source of variability to
indicate the size of the error term that was used in statistical test
of significance from analysis of variance (ANOVA) is shown on
bottom left of each panel. On
days 1 and
2 after inoculation, decreases in LVEF
were significantly (P < 0.0001) less
in rG-CSF-treated animals vs. controls.
For animals challenged with iv endotoxin, the clearance of endotoxin was modeled by a two-parameter exponential decay model {y = b0[exp(b1 × time)]}. The coefficients b0 and b1 were estimated by nonlinear regression techniques for each dog. These coefficients were subsequently analyzed by a Kruskal-Wallis test. TNF-
levels were analyzed by computing the maximum level and
the maximum change from baseline for each dog and subjecting these
summary statistics to an analysis by using a Kruskal-Wallis test.
Cardiopulmonary measures and other laboratory data were analyzed by
using a four-way ANOVA with rG-CSF, MAb R15.7, dog (nested within these
two categories), and time as the main effects.
RESULTS
Cardiovascular studies with pneumonia. No differences were present in any hemodynamic variable measured at baseline in comparing control and rG-CSF-treated animals (P = NS for all). On days 1 and 2 after bacterial inoculation, all animals had significant decreases in mean arterial pressure (MAP), cardiac index, LVSVI, LVSWI, and LVEF (P < 0.001 for all). Reductions in LVEF were significantly more pronounced in controls (P < 0.0001) compared with rG-CSF-treated animals (Fig. 2). To further assess cardiovascular abnormalities, shifts from baseline to days 1 and 2 after bacterial inoculation on Frank-Starling LV function (LVEDVI vs. LVSWI) and LVESVI vs. peak systolic pressure (PSP) plots were analyzed (Fig. 3). Relative to baseline measures of both LVEDVI vs. LVSWI and LVESVI vs. PSP, rG-CSF-treated animals on days 1 and 2 showed less shift downward and to the right than controls (P < 0.001 for both). There were no significant differences between rG-CSF and control animals for any of the other hemodynamic measurements described in MATERIALS AND METHODS at 1, 2, or 14 days after bacterial challenge. Pulmonary studies with pneumonia. No differences were present at baseline in any parameter of pulmonary function comparing control and rG-CSF-treated animals (P = NS for all). On days 1 and 2 after bacterial inoculation, all animals had significant decreases, compared with baseline, in arterial O2 pressure (PaO2), and significant increases in AaO2, QS/QT, and mean pulmonary artery pressure (MPAP) (P < 0.05 for all). In rG-CSF-treated animals, changes in PaO2 (P = 0.09), AaO2 (P = 0.07) and QS/QT (P = 0.07) were not statistically different but not in a range of probability values to convincingly suggest similarity to controls (Fig. 4). No significant differences were present in comparisons of rG-CSF-treated and control animals in any parameter of pulmonary function measured at 14 days after bacterial challenge (P = NS for all). Laboratory evaluation of dogs with pneumonia. At baseline, no differences were present between treatment groups for any laboratory value studied (P = NS for all). Compared with controls, animals treated with rG-CSF developed significantly greater mean circulating concentrations of neutrophils (P < 0.0001; Fig. 5), lymphocytes (data not shown; P < 0.0001), and monocytes (data not shown; P < 0.02) at all subsequent time points with the exception of recovery (14 days after bacterial challenge). On days 1 and 2 after bacterial challenge, all animals had significant decreases in arterial pH and base excess (data not shown; P < 0.001 for both) compared with baseline. Compared with baseline, all animals developed significant elevations in serum endotoxin and TNF-
concentrations
from 2 to 8 h after bacterial challenge (P = 0.01 for both). However, these
elevations were significantly less in rG-CSF-treated animals relative
to controls (P < 0.01 for both; Fig.
6). At baseline and throughout the study,
quantitative blood and lung cultures were similar between control and
rG-CSF-treated animals (data not shown;
P = NS for all). There were
no other changes from baseline, nor were there any differences between treatment groups, for any other hematological or metabolic variables studied during the course of this experiment.
(TNF; B) 2-8 h after bacterial
inoculation in animals receiving rG-CSF or placebo protein (controls).
Compared with controls, animals receiving rG-CSF had significant
decreases in endotoxin and TNF- levels
(P < 0.01 for both).
Challenge with iv endotoxin. Animals treated with rG-CSF had significant increases in numbers of circulating neutrophils compared with controls. Numbers of circulating neutrophils were 61 ± 7 vs. 10 ± 1 × 103 cells/mm3, treated vs. control, respectively, 1 h before endotoxin challenge compared with 17.9 ± 1.6 vs. 5.5 ± 2.6 × 103 cells/mm3, treated vs. control, respectively, 6 h after endotoxin challenge (P = 0.001 at both times). Animals treated with rG-CSF also had significant decreases in serum endotoxin levels after endotoxin challenge (P < 0.04, Fig. 7). Of note, after rG-CSF and before endotoxin (
1 h), TNF- levels were lower, and immediately
after endotoxin challenge (1 to 0 h), TNF- levels were
greater (P < 0.05) in animals
treated with rG-CSF compared with controls. Overall, however, this
resulted in no significant differences from 1 to 6 h in TNF- levels
with rG-CSF compared with controls (P = NS; Fig. 7).
levels
(bottom, A-C) after endotoxin
infusion. A: demonstrated mean values
for each of 4 individual treatment groups. 
and dashed line, rG-CSF only (a); 
and dashed line, monoclonal antibody (MAb) R15.7 only (b); 
and dotted line, rG-CSF + MAb R15.7 (c); 
and solid line, control (d). B: groups that received
rG-CSF (a and c; and dashed line) are averaged together and
compared with those that did not receive rG-CSF (b and d; and solid
line). C: groups that received MAb
R15.7 (b and c; and dashed line) are averaged together and compared
with those that did not (a and d; and solid line). rG-CSF
significantly lowered endotoxin levels after endotoxin challenge
(top, B), but MAb R15.7 had no
effect on endotoxin or cytokine levels. (Bottom
left of each panel), pooled source of variability for
treatments.
rG-CSF from 1 to 6 h after endotoxin challenge was associated with a higher MAP (P = 0.02), LVSWI (P = 0.008), RVSWI (data not shown; P = 0.03), and DO2 (P < 0.05) compared with controls (Fig. 8). rG-CSF from 1 to 6 h after endotoxin challenge was associated with higher MPAP (P = 0.05) and arterial and venous lactates (P = 0.0003 and 0.006, respectively) compared with controls (data not shown).
MAb R15.7 had no significant effect on endotoxin or TNF-
levels
throughout (P = NS; Fig. 7). From 1 to
6 h after endotoxin challenge, MAb R15.7 was associated with
significant reductions in
DO2
(P = 0.002) and cardiac output
(P = 0.04) compared with controls
(Fig. 8). rG-CSF and MAb R15.7 given alone were not associated with any
other significant effects (P = NS).
The effects of rG-CSF and MAb R15.7 given in combination on endotoxemia
and all cardiopulmonary parameters measured were equal to the sum of
their individual effects (i.e., no interaction;
P = NS).
DISCUSSION
Using a non-neutropenic canine model of pneumonia, we found that rG-CSF
given prophylactically decreases endotoxemia and TNF- levels,
attenuates cardiovascular dysfunction, and prolongs survival. In a
previous study, we similarly found that prophylactic rG-CSF administration during canine peritonitis decreased serum endotoxin and
cytokine levels and favorably affected cardiac function and survival
(14). These studies in our canine model add to the growing number of
investigations that suggest that prophylactic rG-CSF improves survival
during infection and sepsis in both immunocompromised and normal
subjects (3, 14, 20, 22, 25, 31, 32, 39, 40). However, in the present
study in animals with pneumonia, increased endotoxin clearance and
improved cardiovascular function with rG-CSF were not associated with
increases in bacterial clearance or improved pulmonary function.
Furthermore, in another set of animals challenged with iv endotoxin
alone, rG-CSF increased endotoxin clearance and improved cardiovascular
function without altering cytokine levels. These results suggest that
rG-CSF in an immunocompetent host can enhance the clearance of
microbial toxins, such as endotoxin, and improve outcome independent of
killing microbes and lowering circulating cytokine levels. Finally, the
beneficial effects of rG-CSF during pneumonia appear to be independent
of improving lung function.
Other recent studies from our laboratory, in combination with the present study, suggest that the beneficial effects on survival associated with rG-CSF may actually vary, depending on experimental conditions. In our canine peritonitis model, we observed that, in contrast to the beneficial effect when rG-CSF was given prophylactically, rG-CSF given therapeutically (i.e., at the time of infection) had no effect and in high doses was harmful (36). Furthermore, using a non-neutropenic, antibiotic-treated rat model, we found that, although prophylactic rG-CSF improved survival when animals were challenged intrabronchially with low or high doses of bacterial inocula, rG-CSF was harmful when animals were challenged with an intermediate dose of bacteria and actually worsened both lung injury and survival (13, 18). In contrast, in the present study of canine pneumonia, rG-CSF had a beneficial effect on survival for all doses of bacteria administered. Taken together, these studies suggest that several factors, including the host species, the route and burden of bacterial infection, and the dose and schedule of rG-CSF administration, influence the efficacy of this therapy. Others have shown that additional conditions, including the level of immunocompetence and the use of antibiotics, may also alter the effects of rG-CSF (32, 33, 40).
There continues to be concern regarding the potential for rG-CSF to aggravate pulmonary inflammatory injury (7, 17). We showed previously that rG-CSF administration can worsen lung injury and survival after intrabronchial E. coli challenge in the rat (18). Others have similarly demonstrated exacerbation of pulmonary injury with rG-CSF administration after intrabronchial instillation of endotoxin or hydrochloric acid (24, 39). In the present study of canines challenged with intrabronchial bacteria, the effects of rG-CSF on lung injury were not significant. In a nonpulmonary infection (bacterial peritonitis), we also found that rG-CSF had no effect on pulmonary injury (14). During iv endotoxin challenge in the present study, rG-CSF did not affect measurements of pulmonary functions. In a rat model, rG-CSF has been demonstrated to attenuate iv endotoxin-induced pulmonary injury (23). Overall, these studies indicate that, although rG-CSF has the potential to aggravate lung injury under certain circumstances, this is not an inevitable consequence of its use. These studies suggest that the pulmonary effects of rG-CSF during sepsis are variable and may depend on a number of factors. During nonpulmonary infection, rG-CSF treatment may augment neutrophil host defense without aggravating inflammatory pulmonary injury. However, during intrapulmonary challenge, the presence of increased numbers of activated neutrophils may in some but not all cases (e.g., the present study) significantly exacerbate pulmonary injury.
A potential mechanism by which rG-CSF mediates its beneficial effects on host defense during sepsis is through enhanced bacterial killing. In our canine peritonitis model, we found that rG-CSF pretreatment significantly decreased quantitative blood cultures. Similarly, others using immunocompromised, nonantibiotic-treated pneumonia models have found that rG-CSF administration decreased the recovery of organisms from blood, tracheobronchial lymph nodes, and pulmonary parenchymal tissue (22, 31). In contrast, in our present study, as well as in our previous investigation of pneumonia in immunocompetent, antibiotic-treated rats, rG-CSF had no discernible effect on lung or blood cultures (18). Our inability to show enhanced bacterial clearance with rG-CSF treatment in these pneumonia models may be related to differences in techniques, sampling times, or burden of bacterial inocula. Alternatively, in the immunocompetent antibiotic-treated host, the ability of rG-CSF to further enhance bacterial killing during pneumonia may be marginal.
Although rG-CSF did not appear to significantly facilitate bacterial clearance in our pneumonia model, rG-CSF did decrease levels of circulating endotoxin. While these reductions may have been related to greater bacterial killing with rG-CSF that was undetectable microbiologically, these findings are also consistent with a direct effect of rG-CSF on the clearance of endotoxin. Such a direct effect on endotoxin clearance is supported by our findings that rG-CSF lowered endotoxin levels in canines challenged with iv endotoxin alone.
The mechanisms by which rG-CSF might enhance endotoxin clearance are not known. Increased production of endotoxin-neutralizing proteins by neutrophils or enhanced neutrophil-mediated endotoxin degradation by rG-CSF are two possible mechanisms (15, 26). One leukocyte surface protein that has been implicated in the clearance of endotoxin is neutrophil CD11/18 adhesion complex, also termed complement receptor 3 (11, 26a). In the present study, we found in dogs that MAbs directed against CD11/18 worsened hemodynamic function but had no effect on endotoxemia. In previous studies in both canine peritonitis and mouse pneumonia models, we also found that CD11/18-directed MAbs worsened hemodynamic function and survival (11, 18). However, in contrast to the present study with bacterial infection, MAb R15.7 was associated with increased endotoxemia (11). In combination, these data suggest that, although CD11/18 may modify endotoxemia during bacterial infection, possibly via its role in the phagocytosis of microbes, its ability to reduce endotoxin-induced cardiovascular injury is related to mechanisms other than the direct clearance of endotoxin.
rG-CSF administration in our pneumonia studies was associated with
significant reductions in circulating TNF- concentrations. It is
unclear whether this effect, which we also observed in our canine
peritonitis model, results from decreases in circulating bacterial
mediators (e.g., endotoxin) or from direct suppression of TNF-
release (1, 14, 20). After iv endotoxin challenge in the present study,
TNF- levels were not altered overall by rG-CSF. These data suggest
that the effects of rG-CSF on TNF- are variable, depending on the
inflammatory stimulus. Moreover, after endotoxin challenge, the
beneficial effects of rG-CSF are not necessarily dependent on lowering
TNF- levels.
Although the neutrophil is thought to be a principal mediator of organ injury during sepsis, rG-CSF-mediated increases in circulating neutrophil counts in this study were associated with improved myocardial and peripheral cardiovascular vascular function (2, 37). In animals with pneumonia and sepsis, myocardial dysfunction was improved by rG-CSF as assessed by three independent parameters: LVEF, Frank-Starling LV function plots, and peak end-systolic volume/peak systolic pressure LV function plots. In addition, in endotoxin-challenged animals, both myocardial function (as assessed by RVSWI and LVSWI and systemic DO2) and peripheral vascular function (as assessed by MAP) were significantly improved. We previously found that rG-CSF improved LVEF and MAP in our canine peritonitis model (14). Although the neutrophil has been proposed as a possible mediator of the organ dysfunction occurring during sepsis, these data suggest that rG-CSF-induced augmentation of neutrophil number and function can have a net beneficial effect on some types of organ injury.
In summary, prophylactic rG-CSF administration reduced endotoxemia and
serum TNF- levels and also improved cardiac function and survival in
our canine model of bacterial pneumonia. In addition, similar treatment
with rG-CSF accelerated endotoxin clearance and improved cardiovascular
function without altering TNF- levels in canines challenged with iv
endotoxin. Our findings suggest that, in non-neutropenic patients who
subsequently develop pneumonia, the administration of prophylactic
rG-CSF in conjunction with supportive therapy may augment host defense,
accelerate the clearance of microbial toxins such as endotoxin, and
result in improved cardiovascular function and survival despite having
no direct salutory effect on lung injury.
ACKNOWLEDGEMENTS
The authors thank Donald Dolan, Alan Hilton, Dan Madden, and Steven Richmond for technical support during this study; Dr. Victoria Hampshire for veterinary care; Julie Friedman for manuscript preparation; and Dr. Robert Cunnion for editorial suggestions.
FOOTNOTES
Address for reprint requests: P. Q. Eichacker, Critical Care Medicine Dept., National Institutes of Health, Bldg. 10, Rm. 7D43, 9000 Rockville Pike, Bethesda, MD 20892-1662.
Received 19 August 1996; accepted in final form 23 May 1997.
REFERENCES
| 1. | Baram, D., P. Madara, S. Culpepper, L. Yan, H. L. Malech, and R. L. Danner. Human neutrophils regulate TNF concentrations in monocyte cultures and whole blood (Abstract). Am. J. Respir. Crit. Care Med. 153: A54, 1996. |
| 2. | Bersten, A., and W. J. Sibbald. Acute lung injury in septic shock. Crit. Care Clin. 5: 49-77, 1989[Medline]. |
| 3. | Cairo, M. S., D. Mauss, S. Kommareddy, K. Norris, C. Van De Ven, and H. Modanlou. Prophylactic or simultaneous administration of recombinant human granulocyte colony stimulating factor in the treatment of group B streptococcal sepsis in neonatal rats. Pediatr. Res. 27: 612-615, 1990[Medline]. |
| 4. | Cole, J. W. L., and J. E. Grizzle. Application of multivariate analysis of variance to repeated measures experiments. Biometrics 22: 810-828, 1966. |
| 5. | Cox, D. R., and D. Oakes. Analysis of Survival Data. London: Chapman and Hall, 1984. |
| 6. |
Danner, R. L.,
R. J. Elin,
J. M. Hosseini,
R. A. Wesley,
J. M. Reilly,
and
J. E. Parillo.
Endotoxemia in human septic shock.
Chest
99:
169-175,
1991 |
| 7. | Demuynck, H., P. Zachee, G. E. G. Verhoef, M. Schetz, G. Van Den Berghe, P. Lauwers, and M. A. Boogaerts. Risks of rhG-CSF treatment in drug-induced agranulocytosis. Ann. Hematol. 70: 143-147, 1995[Medline]. |
| 8. | Dixon, W. J. Ratios involving extreme valves. Ann. Math. Stat. 22: 68-78, 1951. |
| 9. | Eichacker, P. Q., A. Farese, W. D. Hoffman, S. M. Banks, T. Mouginis, S. Richmond, G. C. Kuo, T. J. MacVittie, and C. Natanson. Leukocyte CD11b/18 antigen-directed monoclonal antibody improves early survival and decreases hypoxemia in dogs challenged with tumor necrosis factor. Am. Rev. Respir. Dis. 145: 1023-1029, 1992[Medline]. |
| 10. |
Eichacker, P. Q.,
W. D. Hoffman,
A. Farese,
S. M. Banks,
G. C. Kuo,
T. J. MacVittie,
and
C. Nathanson.
TNF but not IL-1 in dogs causes lethal lung injury and multiple organ dysfunction similar to human sepsis.
J. Appl. Physiol.
71:
1979-1989,
1991 |
| 11. |
Eichacker, P. Q.,
W. D. Hoffman,
A. Farese,
R. L. Danner,
A. F. Suffredini,
Y. Waisman,
S. M. Banks,
T. Mouginis,
L. Wilson,
R. Rothlein,
R. J. Elin,
J. M. Hosseini,
T. J. MacVittie,
and
C. Natanson.
Leukocyte CD18 monoclonal antibody worsens endotoxemia and cardiovascular injury in canines with septic shock.
J. Appl. Physiol.
74:
1885-1892,
1993 |
| 12. | Eichacker, P. Q., and C. Natanson. Rethinking the anti-inflammatory response to septic shock. In: Yearbook of Intensive Care and Emergency Medicine, edited by J. L. Vincent. Frankfurt, Germany: Springer-Verlag, 1996, p. 165-176. |
| 13. | Eichacker, P. Q., M. Quezado, F. Zeni, R. Correa, J. Fontana, S. Banks, Y. Fitz, R. L. Danner, and C. Natanson. G-CSF increases or decreases the relative risk of death dependent on site and severity of infection in rats (Abstract). Am. J. Respir. Crit. Care Med. 153: A252, 1996. |
| 14. |
Eichacker, P. Q.,
Y. Waisman,
C. Natanson,
A. Farese,
W. D. Hoffman,
S. M. Banks,
and
T. J. MacVittie.
Cardiopulmonary effects of granulocyte colony-stimulating factor in a canine model of bacterial sepsis.
J. Appl. Physiol.
77:
2366-2373,
1994 |
| 15. | Erwin, A. L., and R. S. Munford. Processing of LPS by phagocytes. In: Bacterial Endotoxin Lipopolysaccharides, edited by D. C. Morrison, and J. Ryan. Boca Raton, FL: CRC, 1992, p. 405-430. |
| 16. | Espevik, T., and J. Nissen-Meyer. A highly sensitive cell line, WEHI 164 clone 13, for measuring cytotoxic factor/tumor necrosis factor from human monocytes. J. Immunol. 95: 99-105, 1986. |
| 17. | Foster, P. F., D. Mital, H. N. Sankary, L. P. McChesney, J. Marcon, G. Koukoulis, K. Kocess, S. Leurgans, J. F. Whiting, and J. W. Williams. The use of granulocyte colony-stimulating factor after liver transplantation. Transplantation 59: 1557-1563, 1995[Medline]. |
| 18. |
Freeman, B. D.,
R. Correa,
W. Karzai,
C. Natanson,
M. Patterson,
S. Banks,
Y. Fitz,
R. L. Danner,
L. Wilson,
and
P. Q. Eichacker.
Controlled trials of rG-CSF and CD11b-directed MAb during hyperoxia and E. coli pneumonia in rats.
J. Appl. Physiol.
80:
2066-2076,
1996 |
| 20. | Gorgen, I., T. Hartung, M. Leist, M. Niehorster, G. Tiegs, S. Uhlig, F. Weitzel, and A. Wendel. Granulocyte colony-stimulating factor treatment protects rodents against lipopolysaccharide-induced toxicity via suppression of systemic tumor necrosis factor-alpha. J. Immunol. 149: 918-924, 1992[Abstract]. |
| 21. | Guide for the Care and Use of Laboratory Animals. Institute of Laboratory Resources, National Research Council, Department of Health, Education, and Welfare, 1974. (Publ. no. 74-23) |
| 22. |
Hebert, J. C.,
M. O'Reilly,
and
R. L. Gamelli.
Protective effect of recombinant human granulocyte colony-stimulating factor against pneumococcal infections in splenectomized mice.
Arch. Surg.
125:
1075-1078,
1990 |
| 23. | Kanazawa, M., A. Ishizaka, N. Hasegawa, Y. Suzuki, and T. Yokoyama. Granulocyte colony-stimulating factor does not enhance endotoxin-induced acute lung injury in guinea pigs. Am. Rev. Respir. Dis. 145: 1030-1035, 1992[Medline]. |
| 24. | King, J., B. P. Deboisblanc, C. M. Mason, J. M. Onoforio, G. Lipscomb, D. E. Mercante, W. R. Summer, and S. Nelson. Effect of granulocyte colony-stimulating factor on acute lung injury in the rat. Am. J. Respir. Crit. Care Med. 151: 302-309, 1995[Abstract]. |
| 25. |
Lindemann, A.,
F. Herrmann,
W. Oster,
G. Haffner,
W. Meyenburg,
and
L. M. Souza.
Hematologic effects of recombinant human granulocyte colony-stimulating factor in patients with malignancy.
Blood
74:
2644-2651,
1989 |
| 26. | Luchi, M., and R. S. Munford. Binding, internalization and deacylation of bacterial lipolysaccharide by human neutrophils. J. Immunol. 151: 959-969, 1993[Abstract]. |
| 26a. | Lynn, W. A., C. R. Raetz, N. Qureshi, and D. Golenbock. Lipopolysaccharide-induced stimulation of CD11b/CD18 expression on neutrophils. Evidence of specific receptor-based response and inhibition by lipid A-based antagonists. J. Immunol. 147: 3072-3079, 1991[Abstract]. |
| 27. |
Natanson, C.,
P. W. Eichenholz,
R. L. Danner,
P. Q. Eichacker,
W. D. Hoffman,
G. C. Kuo,
S. M. Banks,
T. J. MacVittie,
and
J. E. Parrillo.
Endotoxin and tumor necrosis factor challenges in dogs simulate the cardiovascular profile of human septic shock.
J. Exp. Med.
169:
823-832,
1989 |
| 28. | Natanson, C., M. P. Fink, H. K. Ballantyne, T. J. MacVittie, J. J. Conklin, and J. E. Parillo. Gram-negative bacteremia produces both severe systolic and diastolic cardiac dysfunction in a canine model that simulates human septic shock. J. Clin. Invest. 78: 259-270, 1986. |
| 29. |
Natanson, C.,
W. D. Hoffman,
A. F. Suffredini,
P. Q. Eichacker,
and
R. L. Danner.
Selected treatment strategies for septic shock based on proposed mechanisms of pathogenesis.
Ann. Intern. Med.
120:
771-783,
1994 |
| 30. | Nelson, S., S. Farkas, N. Fotheringham, H. Ho, T. Marrie, and H. Movahhed. Filgrastim in the treatment of hospitalized patients with community acquired pneumonia (CAP) (Abstract). Am. J. Respir. Crit. Care Med. 151: A535, 1996. |
| 31. | Nelson, S., W. Summer, G. Bagby, C. Nakamura, L. Stewart, G. Lipscomb, and J. Andresen. Granulocyte colony-stimulating factor enhances pulmonary host defenses in normal and ethanol-treated rats. J. Infect. Dis. 164: 901-906, 1991[Medline]. |
| 32. | O'Reilly, M., G. M. Silver, D. G. Greenhalgh, R. L. Gamelli, J. H. Davis, and J. C. Hebert. Treatment of intra-abdominal infection with granulocyte colony-stimulating factor. J. Trauma 33: 679-682, 1992[Medline]. |
| 33. | Polak-Wyss, A. Protective effect of human granulocyte colony stimulating factor (hG-CSF) on Cryptococcus and Aspergillus infections in normal and immunosuppressed mice. Mycoses 34: 205-215, 1990. |
| 34. | Quezado, M. J. N., Z. Quezado, F. Zeni, C. Natanson, Y. Fitz, S. M. Banks, and P. Q. Eichacker. G-CSF improves and CD11/18 Mab worsens hemodynamics independently during endotoxemia in canines (Abstract). Am. J. Respir. Crit. Care Med. 153: A838, 1996. |
| 35. | Quezado, Z. M. N., C. Natanson, W. D. Hoffman, S. M. Banks, R. L. Danner, C. A. Koev, and P. Q. Eichacker. Cardiopulmonary effects of inhaled nitric oxide in canine lobar and diffuse gram-negative pneumonia (Abstract). Crit. Care Med. 21: S210, 1993. |
| 36. | Quezado, Z. M. N., C. Natanson, W. D. Hoffman, W. Karzai, C. Koev, R. L. Danner, S. M. Banks, L. Wilson, and P. Q. Eichacker. Granulocyte colony stimulating factor (GCSF) treatment at the onset of E. coli peritonitis-induced septic shock in canines (Abstract). Am. J. Respir. Crit. Care Med. 151: A447, 1995. |
| 37. | Rinaldo, J. E., and J. W. Christman. Mechanisms and mediators of the adult respiratory distress syndrome. Clin. Chest Med. 11: 621-632, 1990[Medline]. |
| 38. | Snedcor, G. W., and W. G. Cochran. Statistical Methods. Ames, IA: Iowa State Univ. Press, 1989. |
| 39. | Terashima, T., M. Kanazawa, K. Sayama, A. Ishizaka, T. Urano, F. Skamaki, H. Nakamura, Y. Waki, and S. Tasaka. Granulocyte colony-stimulating factor exacerbates acute lung injury induced by intratracheal endotoxin in guinea pigs. Am. J. Respir. Crit. Care Med. 149: 1295-1303, 1994[Abstract]. |
| 40. |
Yasuda, H.,
Y. Ajiki,
T. Shimozato,
M. Kasahara,
H. Kawada,
M. Iwata,
and
K. Shimizu.
Therapeutic efficacy of granulocyte colony-stimulating factor alone and in combination with antibiotics against Pseudomonas aeruginosa infections in mice.
Infect. Immun.
58:
2502-2509,
1990 |
This article has been cited by other articles:
|
|
 |
|
  P. C. Minneci, K. J. Deans, B. Hansen, C. Parent, C. Romines, D. A. Gonzales, S.-X. Ying, P. Munson, A. F. Suffredini, J. Feng, et al. A canine model of septic shock: balancing animal welfare and scientific relevance Am J Physiol Heart Circ Physiol, October 1, 2007; 293(4): H2487 - H2500. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. S. Michalopoulos, S. Geroulanos, and S. D. Mentzelopoulos Determinants of Candidemia and Candidemia-Related Death in Cardiothoracic ICU Patients Chest, December 1, 2003; 124(6): 2244 - 2255. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Rassias, A. L. Givan, C. A. S. Marrin, K. Whalen, J. Pahl, and M. P. Yeager Insulin Increases Neutrophil Count and Phagocytic Capacity After Cardiac Surgery Anesth. Analg., May 1, 2002; 94(5): 1113 - 1119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Quezado, C. Parent, W. Karzai, M. Depietro, C. Natanson, W. Hammond, R. L. Danner, X. Cui, Y. Fitz, S. M. Banks, et al. Acute G-CSF therapy is not protective during lethal E. coli sepsis Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2001; 281(4): R1177 - R1185. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. WOLLIN, S. UHLIG, R. NÜSING, and A. WENDEL Granulocyte-Macrophage Colony-Stimulating Factor Amplifies Lipopolysaccharide-induced Bronchoconstriction by a Neutrophil- and Cyclooxygenase 2-Dependent Mechanism Am. J. Respir. Crit. Care Med., February 1, 2001; 163(2): 443 - 450. [Abstract] [Full Text]
|
|
|
|
|
|
G. MATUTE-BELLO, C. W. FREVERT, O. KAJIKAWA, S. J. SKERRETT, R. B. GOODMAN, D. R. PARK, and T. R. MARTIN Septic Shock and Acute Lung Injury in Rabbits with Peritonitis . Failure of the Neutrophil Response to Localized Infection Am. J. Respir. Crit. Care Med., January 1, 2001; 163(1): 234 - 243. [Abstract] [Full Text]
|
|
|
|
 |
|
 F. Zeni, C. Parent, R. Correa, C. Natanson, B. Freeman, J. Fontana, M. Quezado, R. L. Danner, Y. Fitz, S. Richmond, et al. ICAM-1 and CD11b inhibition worsen outcome in rats with E. coli pneumonia J Appl Physiol, July 1, 1999; 87(1): 299 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. KARZAI, B. U. von SPECHT, C. PARENT, J. HABERSTROH, K. WOLLERSEN, C. NATANSON, S. M. BANKS, and PETER. Q. EICHACKER G-CSF during Escherichia coli versus Staphylococcus aureus Pneumonia in Rats Has Fundamentally Different and Opposite Effects Am. J. Respir. Crit. Care Med., May 1, 1999; 159(5): 1377 - 1382. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Weiss, L. L. Moldawer, and E. M. Schneider Granulocyte Colony-Stimulating Factor to Prevent the Progression of Systemic Nonresponsiveness in Systemic Inflammatory Response Syndrome and Sepsis Blood, January 15, 1999; 93(2): 425 - 439. [Full Text] [PDF]
|
|
|
|
|
|
Z. M. N. Quezado, C. Natanson, W. Karzai, R. L. Danner, C. A. Koev, Y. Fitz, D. P. Dolan, S. Richmond, S. M. Banks, L. Wilson, et al. Cardiopulmonary effects of inhaled nitric oxide in normal dogs and during E. coli pneumonia and sepsis J Appl Physiol, January 1, 1998; 84(1): 107 - 115. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
