Content uploaded by Cristina Casals
Author content
All content in this area was uploaded by Cristina Casals on Jan 06, 2018
Content may be subject to copyright.
EFFECT OF SURFACTANT PROTEIN A (SP-A) ON THE PRODUCTION OF
CYTOKINES BY HUMAN PULMONARY MACROPHAGES
Javier Arias-Diaz,* Ignacio Garcia-Verdugo,
†
Cristina Casals,
†
Natalia Sanchez-Rico,
‡
Elena Vara,
‡
and Jose´ L. Balibrea*
*Department of Surgery, Hospital San Carlos, Madrid, Spain
Department of Biochemistry,
†
Faculty of Biology and
‡
Medicine, Complutense University, Madrid, Spain
Received 28 Apr 2000; first review completed 5 May 2000; accepted in final form 16 Jun 2000
ABSTRACT—Surfactant protein A (SP-A) is thought to play a role in the modulation of lung inflammation
during acute respiratory distress syndrome (ARDS). However, SP-A has been reported both to stimulate and
to inhibit the proinflammatory activity of pulmonary macrophages (M
). Because of the interspecies differ-
ences and heterogeneity of M
subpopulations used may have influenced previous controversial results, in
this study, we investigated the effect of human SP-A on the production of cytokines and other inflammatory
mediators by two well-defined subpopulations of human pulmonary M
. Surfactant and both alveolar (aM
)
and interstitial (iM
) macrophages were obtained from multiple organ donor lungs by bronchoalveolar lavage
and enzymatic digestion. Donors with either recent history of tobacco smoking, more than 72 h on mechani-
cal ventilation, or any radiological pulmonary infiltrate were discarded. SP-A was purified from isolated
surfactant using sequential butanol and octyl glucoside extractions. After 24-h preculture, purified M
were
cultured for 24 h in the presence or absence of LPS (10 µg/mL), SP-A (50 µg/mL), and combinations. Nitric
oxide and carbon monoxide (CO) generation (pmol/µg protein), cell cGMP content (pmol/µg protein), and
tumor necrosis factor alpha (TNF
␣
), interleukin (IL)-1, and IL-6 release to the medium (pg/µg protein) were
determined. SP-A inhibited the lipopolysaccharide (LPS)-induced TNF
␣
response of both interstitial and
alveolar human M
, as well as the IL-1 response in iM
. The SP-A effect on TNF
␣
production could be
mediated by a suppression in the LPS-induced increase in intracellular cGMP. In iM
but not in aM
, SP-A
also inhibited the LPS-induced IL-1 secretion and CO generation. These data lend further credit to a
physiological function of SP-A in regulating alveolar host defense and inflammation by suggesting a funda-
mental role of this apoprotein in limiting excessive proinflammatory cytokine release in pulmonary M
during
ARDS.
KEYWORDS—Nitric oxide, carbon monoxide, cyclic guanosin 3⬘5⬘-monophosphate, pulmonary surfactants;
bronchoalveolar lavage fluid
INTRODUCTION
Cytokines seem to act predominantly in a paracrine way
when producing their deleterious effects during sepsis. There-
fore, the local release of cytokines by pulmonary M
would be
of importance in the pathogenesis of the adult respiratory
distress syndrome (ARDS). Recently, it has been found that
alveolar type II cells and some of their secretory products,
including surfactant-associated protein A (SP-A) and D, have
nonsurfactant-related functions and are actively associated
with various defense functions in the lung (1). SP-A is the most
abundant of the surfactant proteins. It is a calcium-dependent
lectin structurally similar to the complement component, C1q,
and the mannose-binding protein, an acute-phase protein with
opsonic function. Some of its nonsurfactant-related functions
include augmentation of alveolar macrophage migration,
enhancement of macrophage phagocytosis of opsonized or
nonopsonized bacteria or viruses (1), and regulation of reactive
oxygen species production (2).
SP-A seems also to influence the proinflammatory activity
of aM
, and it may participate this way in the modulation of
lung inflammation during the ARDS; however, it has been
reported both to stimulate (3) and to inhibit (4) the production
of cytokines by pulmonary M
.
Alveoli are lined with surfactant that consists of about 90%
lipids and 5–10% surfactant-specific proteins. The major func-
tion of this material is to reduce the surface tension in the
alveoli, thereby preventing alveolar collapse and edema.
Alveolar M
but not interstitial M
reside in this surfactant-
rich microenvironment and contain surfactant, being more
exposed to a potential influence of surfactant constituents.
However, during lung inflammation, alveolar epithelium is
disrupted and cells in the interstitium can also contact surfac-
tant. Furthermore, there have been several studies that have
detected circulating antibodies to the surfactant proteins or the
surfactant proteins themselves in the serum (5), suggesting that
these proteins may be introduced into the circulation in some
conditions and be able to interact with interstitial and even
circulating immune cells.
Besides different localizations, it is reasonable to assume
that the iM
and aM
populations exert distinct functions that
are altered upon exposure to inflammatory agents and that the
intensity and nature of these events have an impact on the net
Presented at the 20
th
annual meeting of the Surgical Infection Society, April
27–29, 2000, Providence RI.
Address reprint requests to Javier Arias-Diaz, MD, PhD, Departamento de
Cirugia Hospital Clinico San Carlos, Ciudad Universitaria, 28040-Madrid,
Spain.
SHOCK, Vol. 14, No. 3, pp. 300–306, 2000
300
outcome of the inflammatory response. Accordingly, the ques-
tion arises of whether SP-A affects differently both M
popu-
lations in humans. To explore this possibility we studied the
lipopolysaccharide (LPS)-induced production of tumor necro-
sis factor alpha (TNF
␣
), interleukin (IL)-1

, and IL-6 by
human interstitial and alveolar M
. To gain further insight into
the mechanisms underlying the SP-A effect, we also investi-
gated the cyclic guanosine 3⬘5⬘-monophosphate (cGMP) path-
way of LPS-induced cytokine secretion (6).
MATERIALS AND METHODS
Human lung tissue procurement
This study was approved by the review board and the ethic committee of the
Hospital San Carlos. Human lung tissue was obtained from male multiple
organ donors. A written informed consent was obtained from the next-of-kin
of each donor. Ages ranged from 22 to 46 years, and cranial trauma or spon-
taneous intracranial hemorrhage was the cause of death in all donors. Donors
with either recent history of tobacco smoking, more than 72 h receiving
mechanical ventilation or any radiological pulmonary infiltrate were excluded
from this study. Immediately after obtaining the organs that were going to be
used for transplantation, the left lung was excised. A bronchoalveolar lavage
(BAL) was performed through the main bronchus using a total of 500 mL of
0.9% normal saline at 4°C, and the lung was then placed in cold saline solution.
Cold ischemia period was less than3hinallcases.
Isolation of SP-A
SP-A was purified from the fluid obtained at lung lavage as described
previously in detail (7). Briefly, the whole surfactant pellet from lung lavage
fluids was extracted with butanol; butanol-insoluble proteins were resolubi-
lized with octylglucopyranoside (OGP); and subsequently, SP-A was solubi-
lized in 5 mM Tris-buffered water, pH 7.4. The purity of SP-A was checked
by one-dimensional SDS/PAGE in 12% acrylamide under reducing conditions
(50 mM dithiothreitol). Quantification of SP-A was carried out by amino acid
analysis in a Beckman System 6300 High Performance analyzer (7). The
protein hydrolysis was performed with 0.2 mL of 6 M HCl, containing 0.1%
(w/v) phenol in evacuated and sealed tubes at 108°C for 24 h. Norleucine was
added to each sample as internal standard.
Cell isolation and culture
Bronchoalveolar cells were separated from lavage fluid by centrifugation.
The sedimented cells were washed twice with Hank’s balanced salt solution
(HBSS) and then centrifuged (250 g, 10 min). To isolate iM
(6), pulmonary
tissue was washed with a calcium-free solution and then minced into small
fragments (1 to 2 mm). After copious washing with the same solution, to
remove most blood cells and remaining airway M
, tissue fragments under-
went an enzymatic digestion with elastase (27 orcein-elastin U/mL elastase) in
a 37°C shaking bath. Digestion was stopped by addition of 4°C fetal calf
serum, the tissue was filtered through nylon mesh (200 and 20
m), and the
cell suspension was washed twice with HBSS and centrifuged (250 g, 10 min).
The cell pellets were resuspended in RPMI-1640 medium (10% heat-
inactivated fetal calf serum, 100 IU/mL penicillin G, 50
g/mL gentamycin),
and poured into 75 cm
2
culture flasks. After 90 min of incubation (37°C, under
O
2
/CO
2
atmosphere), the supernatants were removed and the cells were
washed four times with phosphate-buffered saline (PBS) to remove contami-
nating non-adherent cells. Adherent cells were found to be 99.5 ± 0.3% viable
(Trypan blue exclusion test), and to be composed of 92.5 ± 3.1% M
,as
judged by Wright–Giemsa stained cytocentrifuge preparations, being most of
the remaining cells neutrophils or fibroblasts. The cells were gently scraped,
plated onto collagen-coated 96-well plastic dishes (5 × 10
5
cells per well) and
precultured for 24 h. Under these conditions, approximately 95% of the cell
were attached; cell viability was higher than 97% and macrophage purity was
always greater than 98%. Cells were cultured for another 24 h in the presence
or absence of LPS (Escherichia Coli 055:B5, 10
g/mL), 8-Br-cGMP (1
mmol/L), methylene blue (MB, 10
−5
mol/L), hemin (1
mol/L), SP-A (50
g/mL), and combinations. Macrophage cultures from each individual donor
were plated in duplicate wells and each series of experiments was repeated
with a minimum of at least six donors.
Biochemical determinations
Supernatants and cells were recovered, and the cGMP content of the cells
and TNF
␣
, IL-1, IL-6, carbon monoxide (CO), and nitric oxide (NO) release
to the medium were determined. Cell cGMP content was measured with
specific RIA (
125
I-RIA Kit, Radiochemical Centre, Amersham, Bucks, UK).
The cytokines measurements were performed by specific ELISA kits. An
aliquot of the cell suspension was used for protein quantitation, which was
performed spectrophotometrically by the Coomassie brilliant blue dye method.
Release of NO was measured by the Griess reaction as nitrite (NO
2−
)
concentration after nitrate (NO
3−
) reduction to NO
2−
. Briefly, samples were
deproteinized by the addition of sulfosalicylic acid. They were then incubated
for 30 min at 4°C, and subsequently centrifuged for 20 min at 12000 g. After
incubation of the supernatants with E. coli NO
3−
reductase (37°C, 30 min), 1
mL of Griess reagent (0.5% naphthylethylenediamine dihydrochloride, 5%
sulfonilamide, 25% H
3
PO
4
) was added. The reaction was performed at 22°C
for 20 min, and the absorbance at 546 nm was measured, using NaNO
2
solution
as standard.
To quantify the amount of CO formed, hemoglobin (Hb) was added to bind
CO as carboxyhemoglobin (CO-Hb) and the proportion of CO-Hb was esti-
mated (6). For that, Hb (4
mol/L) was mixed gently into the sample and 1 min
was allowed to ensure maximum CO binding. Then, samples were diluted with
buffer phosphate (0.01 mol/L KH
2
PO
4
/K
2
HPO
4
, pH 6.85) containing 2 mg/mL
sodium hydrosulfite, gently mixed in stoppered cuvettes and allowed to stand
at room temperature for 10 min. Absorbance was read at 420 and 432 nm
against a matched cuvette containing only buffer.
Statistical analyses
N represents the number of separate macrophage preparations employed
(each from a different donor lung). The different assays were performed in at
least duplicates, and the means were calculated. The results are presented as
the means [± standard error of the mean (SEM)], obtained by combining the
results from each cell preparation. Mean comparison was done by the Fried-
man’s analysis of variance of ranks, followed by a two-tailed Wilcoxon’s rank
sum test for paired data to identify the source of the found differences; a
confidence level 95% or greater (P< 0.05) was considered significant.
RESULTS
LPS significantly increased the release of TNF
␣
(Fig. 1A),
IL-1 (Fig. 2A), and IL-6 (Fig. 3A) by both iM
and aM
. The
increase in TNF
␣
production was more pronounced in iM
than in aM
.
The baseline release of TNF
␣
, IL-1, and IL-6 by unstimu-
lated M
was not affected by 50
g/mL SP-A, which,
however, significantly decreased the LPS-induced TNF
␣
release by both iM
and aM
(Fig. 1A). SP-A also decreased
the LPS-induced IL-1 release by iM
. However, it did not
modify the IL-1 release by aM
(Fig. 2A). SP-A did not
modify the LPS-induced IL-6 release by aM
and iM
(Fig.
3A).
LPS also increased the content of cGMP on aM
as well as
on iM
(Fig. 4). SP-A significantly reduced the increase on
cGMP content induced by LPS (Fig. 4A).
MB, a guanylate cyclase (GC) inhibitor, was effective in
suppressing the LPS-induced increase in cGMP content (Fig.
4A). This inhibitory effect was accompanied by an inhibition
in the TNF
␣
secretion (Fig. 1A) without modifying IL-6
release (Fig. 3A). Neither LPS nor SP-A showed any detect-
able effect on the NO release to the medium (Fig. 5). In
contrast, LPS significantly increased CO release to the medium
by both iM
an aM
.IniM
, this increase was significantly
SHOCK SEPTEMBER 2000 EFFECT OF SP-A ON PULMONARY MACROPHAGES 301
reduced in the presence of SP-A, which, however, did not
modify the LPS-induced CO production on aM
(Fig. 6A).
The addition of 8-Br-cGMP to the medium increased TNF
␣
(Fig. 1B) and IL-1 (Fig. 2B) release by both aM
and iM
but
did not modify IL-6 release (Fig. 3b). The addition of hemin,
the substrate for heme oxygenase, induced similar changes.
The hemin effects on TNF
␣
and IL-1 release were accompa-
nied by an increase in cGMP content (Fig. 4B) and CO release
to the medium (Fig. 6B). SP-A did not modify hemin nor
cGMP effects.
In additional experiments, we studied the dose-dependent
activity of SP-A by using a lower concentration (10
g/mL)
and another one higher (100
g/mL) than 50
g/mL. Even 100
g/mL SP-A did not affect either the baseline concentrations
of cytokines, CO and cGMP, or the LPS-stimulated release of
IL-6 by M
.IniM
, SP-A dose-dependently inhibited the the
LPS-stimulated increases in TNF
␣
(to 6.54 ± 0.05 at 10
g/mL
SP-A, and 2.94 ± 0.01 at 100
g/mL SP-A; pg/
g protein; N
⳱3), IL-1 (to 4.25 ± 0.08 at 10
g/mL SP-A, and 2.87 ± 0.04
at 100
g/mL SP-A; pg/
g protein; N ⳱3), CO (to 5.15 ±
0.03 at 10
g/mL SP-A and 2.68 ± 0.09 at 100
g/mL SP-A;
pmol/
g protein; N ⳱3), and cGMP (to 0.26 ± 0.02 at 10
g/mL SP-A, and 0.11 ± 0.01 at 100
g/mL SP-A; pmol/
g
protein; N ⳱3). In aM
, SP-A did not affect the LPS-
stimulated production of either IL-1 or CO even at the maximal
concentration used. The effect on LPS-stimulated TNF
␣
and
cGMP increases was related to the SP-A concentration also in
aM
(to 4.34 ± 0.28 at 10
g/mL SP-A, and 3.02 ± 0.11 at 100
g/mL SP-A; pg/
g protein; for TNF
␣
; and to 0.17 ± 0.01 at
10
/mL SP-A, and 0.13 ± 0.01 at 100
g/mL SP-A; pmol/
g
protein; for cGMP; N ⳱3) . There were no significant differ-
ences between the effects of 50
g/mL and 100
g/mL SP-A
in any case.
DISCUSSION
The main findings of the present study indicate that SP-A
can modulate lung inflammation in humans by decreasing the
LPS-induced production of proinflammatory cytokines by both
iM
and aM
. They also show some differences in the
response to SP-A between both M
populations.
Pulmonary M
are considered to be present in at least two
anatomically different compartments (8, 9). Alveolar M
are
found in the bronchoalveolar lumen, where they act as a
primary defensive line phagocytosing inhaled material. Inter-
stitial M
, similar in number, reside within the interstitial
space and are thought to be precursors of the former. To our
knowledge, all previous studies involving interactions between
SP-A and pulmonary M
were carried out using exclusively
the alveolar variety, easily accessible by BAL. The method
used by us permits also the recovery of iM
, which, in fact,
seem to be more active from the immunological viewpoint
(10). As a source of lung tissue, we used multiple organ donors
because pulmonary M
from patients with lung cancer, whose
surgical resection specimen would have been a more accessible
source of pulmonary tissue, have shown a higher TNF
␣
and
FIG.1. Secretion of TNF
␣
by human pulmonary macrophages in
the presence or absence of LPS (10 µg/mL), SP-A (50 µg/mL), meth-
ylene blue (MB, 10
−5
mol/L), hemin (1 µmol/L), 8-Br-cGMP (1 mmol/
L), and combinations. Each column represents the mean ± SE of
duplicate samples from 6 different experiments. *
P
< 0.01 vs. all other
groups, **
P
< 0.01 vs. the rest (upper panel); *
P
<0.01vs
.
control and
SP-A, **
P
<0.05vs
.
Br-cGMP groups (lower panel).
FIG.2. Secretion of IL-1 by human pulmonary macrophages in
the presence or absence of LPS (10 µg/mL), SP-A (50 µg/mL), meth-
ylene blue (MB, 10
−5
mol/L), hemin (1 µmol/L), 8-Br-cGMP (1 mmol/
L), and combinations. Each column represents the mean ± SE of
duplicate samples from 6 different experiments. *
P
< 0.01 vs. all other
groups, **
P
< 0.01 vs. the rest (upper panel); *
P
< 0.01 vs. control and
SP-A (lower panel).
302 SHOCK VOL. 14, NO.3 ARIAS-DIAZ ET AL.
IL-1 secretion in vitro than healthy controls (11). Similarly, we
discarded the donors with recent history of tobacco smoking in
view that functional changes have been reported in aM
from
smokers (12). In bronchoalveolar fluid, the concentration of
SP-A is normally 1–2
g/mL (13). However, dilution due to
the lavage procedure probably causes a 10- to 100-fold reduc-
tion in the concentrations normally present in the alveolar
hypophase. Therefore, we believe that the SP-A concentration
we have used is in the physiological range. On the other hand,
the concentration of LPS used as a stimulus was selected
according to previous dose-response studies in human pulmo-
nary M
(14).
Leukocytes obtained from the alveolar compartment by
lavage techniques repeatedly have been shown to be hypore-
sponsive to inflammatory stimuli as compared with leukocytes
isolated from peripheral blood. This relative “dampening” of
leukocyte activity within the alveolar space is thought to
protect the host from persistent immune cell activation via
inhaled antigens. Surfactant has since long been implicated in
this suppression, since surfactant lipid mixtures and individual
lipid components were noted to inhibit lymphocyte prolifera-
tion and Ig secretion, as well as phagocyte oxygen radical
production. Recently, immunosuppressive activity has also
been demonstrated for SP-A, since this major surfactant
component inhibited lymphocyte proliferation and IL-2
production (15), and reduced TNF
␣
generation in M
(4, 16).
However, others have reported that SP-A per se stimulated
proinflammatory cytokine production in mononuclear cells,
secretion of Igs by splenocytes, and proliferation of lympho-
cytes (3, 17).
Our observation that SP-A decreased TNF
␣
secretion from
LPS-stimulated M
appears to contradict the data from Krem-
lev and Phelps (3), who showed that SP-A could stimulate the
production of TNF
␣
by rat aM
and TNF
␣
, IL-1, and IL-6 by
human monocytes. Although our study focused on examining
the effects of SP-A on LPS-activated M
, we did test the
ability of SP-A to induce production of TNF
␣
, IL-1, or IL-6 by
unstimulated cells and we could not detect any significant
increase in the baseline liberation of these cytokines. We
currently have no unequivocal explanation for these contrast-
ing results, although there were some differences in experi-
FIG.3. Secretion of IL-6 by human pulmonary macrophages in
the presence or absence of LPS (10 µg/mL), SP-A (50 µg/mL), meth-
ylene blue (MB, 10
−5
mol/L), hemin (1 µmol/L), 8-Br-cGMP (1 mmol/
L), and combinations. Each column represents the mean±SE of dupli-
cate samples from 6 different experiments. *
P
< 0.01 vs. non-LPS
groups.
FIG.4. Levels of cGMP in human pulmonary macrophages in the
presence or absence of LPS (10 µg/mL), SP-A (50 µg/mL), methy-
lene blue (MB, 10
−5
mol/L), hemin (1 µmol/L), and combinations.
Each column represents the mean ± SE of duplicate samples from 6
different experiments. *
P
< 0.01 vs. all other groups (upper panel); *
P
<
0.01 vs. control and SP-A (lower panel).
FIG.5. Production of nitric oxide by human pulmonary macro-
phages in the presence or absence of LPS (10 µg/mL) ± SP-A (50
µg/mL). Each column represents the mean ± SE of duplicate samples
from 6 different experiments.
SHOCK SEPTEMBER 2000 EFFECT OF SP-A ON PULMONARY MACROPHAGES 303
mental design. For example, the cells were isolated and plated
under slightly different conditions. It is possible that these
subtle differences in design induced differences in the respon-
siveness of the aM
. Major differences in the two studies are
the different origin of pulmonary M
, the different source of
human SP-A and the different methods of SP-A isolation. Vari-
ous cell types might respond differentially to SP-A, as well as
the cells from different animals, and it is well known that
interspecies variation in the mediator-induced behavior poten-
tially is a problem when extrapolating from animal studies to
humans. In an attempt to address this issue, we have employed
M
from human origin and we tested two different subpopu-
lations of pulmonary M
. The origin and chemical features of
the SP-A batches employed in these studies must also be taken
into consideration. We isolated SP-A from previously healthy
organ donors. The SP-A used in the Kremlev and Phelps study
(3) was from patients with alveolar proteinosis, whose SP-A
are known to differ from normal subjects (18), and was isolated
by a different technique. Also, SP-A appears to avidly bind
LPS, and it even may enhance presentation of LPS to aM
(19); thus, even a trace contamination of endotoxin with SP-A
might alter the M
response. In agreement with our results,
Rosseau et al. (4) have found that SP-A strongly suppress the
proinflammatory cytokine response of human aM
and mono-
cytes to Candida albicans, effecting down-regulation of proin-
flammatory cytokine synthesis at the transcriptional level. In
this study, internalization of SP-A by M
seemed necessary
for the interference with the cytokine response. In vivo,aM
are continuously in contact with surfactant and SP-A and yet
are not permanently activated. However, it is also recognized
that SP-A likely functions differently in vitro than in the
complex lipid-rich milieu of the alveolar spaces where any
potential proinflammatory effect could be overcome by an
antagonistic effect of another surfactant component. In fact, the
previous in vivo surfactant exposure could also influence the in
vitro behavior of aM
after isolation. This potential influence
of previous surfactant exposure should be negligible in the
iM
subpopulation.
Our findings, that exogenously added cGMP increases the
release of TNF
␣
and IL-1 and that LPS-stimulated TNF
␣
and
IL-1 release can be abolished by the GC inhibitor MB, are in
agreement with the hypothesis of cGMP playing a role in the
regulation of both TNF
␣
and IL-1 production by macrophages
(20). In contrast, a different regulatory mechanism seems to be
present for IL-6. Cyclic GMP is generated by the enzyme GC,
and the best known activator of the soluble isoform of this
enzyme is NO (21). Nevertheless, a high output, inducible
NO-generating system similar to that reported in rats and mice
has not been found in human M
under a variety of conditions
(22, 23). In fact, we did not find any increase in nitrite concen-
tration in the supernatants of LPS-stimulated macrophages
despite their showing an increased production of cGMP and
cytokines, confirming previous findings of our group (6) as
well as others (22, 23). Since GC can be activated also by CO
(21) and we have found this gas can participate in the LPS-
induced cytokine production in human M
(6), we chose to
examine the production of CO in our system. We found an
LPS-induced increase in CO generation by both iM
and aM
.
Hemin, a precursor of CO elicited the same effect serving as a
control. SP-A was able to suppress the LPS-induced produc-
tion of CO on iM
but not on aM
.
At least two sources of CO are available, one of which is the
metabolism of heme (24), catalysed by heme oxygenase. A
second source is lipid peroxidation. Induction of heme oxygen-
ase has been observed in the presence of several oxidant
stresses, especially agents that deplete glutathione (25), and
LPS has been found to induce heme oxygenase activity in rat
macrophages and Kupffer cells (26). Our finding that hemin
increases CO, TNF
␣
, and IL-1 production and cGMP content
indicates that endogenous heme oxygenase-dependent CO
production is effective in triggering the production of both
TNF
␣
and IL-1 by macrophages, and SP-A seems not to inter-
fere with this pathway.
On the other hand, an increase in the generation of oxygen
free radicals is a well-known component in the pathophysiol-
ogy of sepsis. Several specific intracellular signaling mecha-
nism have been found to be influenced by redox changes,
including calcium shifts, tyrosine kinase activation, and the
nuclear transcription factor NF-
B. It has been shown that the
LPS stimulatory pathway in the alveolar macrophage possesses
a signal transduction mechanism that is sensitive to redox
changes (27). Given that SP-A did not interfere with the effect
of exogenously added cGMP or hemin, the inhibitory effect of
SP-A on LPS-induced CO production in iM
might be
explained by an SP-A-mediated attenuation of the LPS-
induced oxygen radical production in iM
. Although SP-A
does not have a scavenger effect for superoxide anion (28),
FIG.6.Production of carbon monoxide by human pulmonary
macrophages in the presence or absence of LPS (10 µg/mL), SP-A
(50 µg/mL), methylene blue (MB, 10
−5
mol/L), hemin (1 µmol/L),
8-Br-cGMP (1 mmol/L), and combinations. Each column represents
the mean ± SE of duplicate samples from 6 different experiments. *
P
<
0.01 vs. all other groups.
304 SHOCK VOL. 14, NO.3 ARIAS-DIAZ ET AL.
several studies suggest that SP-A alters oxygen radical produc-
tion (28). Alveolar M
incubated with SP-A have a decrease in
superoxide production, indicating a dampening of the respira-
tory burst (15, 28) and suggesting that SP-A has a protective
role against the oxidant injury caused by aM
in the lung. In
contrast with these results, van Iwaarden et al. (2) found an
increase in the oxygen radical production by rat aM
induced
by human SP-A.
Collectively, our findings suggest functional differences
between iM
and aM
. The LPS-induced increase in IL-1
release and CO generation was inhibited by SP-A in iM
, but
not in aM
. These findings are also in agreement with earlier
reports of the functional differences between aM
and M
from digested lung tissue (9). Nevertheless, we cannot exclude
a possible influence of the previous in vivo surfactant exposi-
tion of the aM
subpopulation. In addition, there exists the
possibility that the enzyme treatment per se could have influ-
ence the phenotype of the cells, although Johansson et al. (8)
showed that enzyme treatment did not modify the receptor
expression of rat iM
.
The LPS-elicited synthesis of IL-1 in aM
, and IL-6 in both
types of M
was not suppressed by SP-A. This selective effect
together with the finding that SP-A did not affect the baseline
levels of cytokine generation in our M
suggest a distinct
immunomodulatory rather than a general inhibitory effect of
SP-A on M
inflammatory responsiveness.
In conclusion, the major surfactant protein SP-A was found
to inhibit the LPS-induced TNF
␣
response of both interstitial
and alveolar human M
, as well as the IL-1 response in iM
.
The SP-A effect on TNF
␣
production could be mediated by a
suppression in the LPS-induced increase in intracellular
cGMP. In iM
, but not in aM
, SP-A also inhibited the LPS-
induced IL-1 secretion and CO generation. These data lend
further credit to a physiological function of SP-A in regulating
alveolar host defense and inflammation and suggest a funda-
mental role of this apoprotein in limiting excessive proinflam-
matory cytokine release in pulmonary M
during the ARDS.
ACKNOWLEDGMENTS
This report was partially financed by grants from Fondo de Investigacio´nde
la Seguridad Social (FISS 98/1397), and Direccio´n General De Investigacio´n
Cientı´fica Y Te´cnica (DGICYT PB98-0769-C02). We thank the transplant
coordinators and transplant team staff nurses of the Hospital Clinico “San
Carlos” for their kind cooperation in obtaining the lung tissue.
REFERENCES
1. McCormack F: The structure and function of surfactant protein-A. Chest
111:114S–119S, 1997.
2. van Iwaarden F, Welmers B, Verhoef J, Haagsman HP, van Golde LM:
Pulmonary surfactant protein A enhances the host-defense mechanism of
rat alveolar macrophages. Am J Respir Cell Mol Biol 2:91–98, 1990.
3. Kremlev SG, Phelps DS: Surfactant protein A stimulation of inflammatory
cytokine and immunoglobulin production. Am J Physiol 267:L712–L719,
1994.
4. Rosseau S, Hammerl P, Maus U, Gu¨ nther A, Seeger W, Grimminger F,
Lohmeyer J: Surfactant protein A down-regulates proinflammatory cyto-
kine production evoked by Candida albicans in human alveolar macro-
phages and monocytes. J Immunol 163:4495–4502, 1999.
5. Chida S, Phelps DS, Soll RF, Taeusch HW: Surfactant proteins and anti-
surfactant antibodies in sera from infants with respiratory distress
syndrome with and without surfactant treatment. Pediatrics 88:84–89,
1991.
6. Arias-Diaz J, Vara E, Garcia C, Villa N, Balibrea JL: Evidence for a cyclic
guanosine monophosphate-dependent, carbon monoxide-mediated, signal-
ing system in the regulation of TNF
␣
production by human pulmonary
macrophages. Arch Surg 130:1287–1293, 1995.
7. Casals C, Miguel E, Perez-Gil J: Tryptophan fluorescence study on the
interaction of pulmonary surfactant protein A with phospholipid vesicles.
Biochem J 296(Pt 3):585–593, 1993
8. Johansson A, Lundborg M, Sko¨ld CM, Lundahl J, Tornling G, Eklund A,
Camner P: Functional, morphological, and phenotypical differences
between rat alveolar and interstitial macrophages. Am J Respir Cell Mol
Biol 16:582–588, 1997.
9. Zetterberg G, Johansson A, Lundahl J, Lundborg M, Sko¨ld CM, Tornling
G, Camner P, Eklund A: Differences between rat alveolar and interstitial
macrophages 5 wk after quartz exposure. Am J Physiol 274:L226–L234,
1998.
10. Weissler JC, Lyons CR, Lipscomb MF, Toews GB: Human pulmonary
macrophages. Functional comparison of cells obtained from whole lung
and by bronchoalveolar lavage. Am Rev Respir Dis 133:473–477, 1986.
11. Siziopikou KP, Harris JE, Casey L, Nawas Y, Braun DP: Impaired tumori-
cidal function of alveolar macrophages from patients with non-small cell
lung cancer. Cancer 68:1035–1044, 1991.
12. Sherman MP, Campbell LA, Gong HJr, Roth MD, Tashkin DP: Antimi-
crobial and respiratory burst characteristics of pulmonary alveolar macro-
phages recovered from smokers of marijuana alone, smokers of tobacco
alone, smokers of marijuana and tobacco, and nonsmokers. Am Rev Respir
Dis 144:1351–1356, 1991.
13. Phelps DS, Rose RM: Increased recovery of surfactant protein A (SP-A)
in AIDS related pneumonia. Am Rev Respir Dis 143:1072–1075, 1991.
14. Balibrea JL, Arias-Diaz J, Garcia C, Vara E: Effect of pentoxifylline and
somatostatin on tumour necrosis factor production by human pulmonary
macrophages. Circ Shock 43:51–56, 1994.
15. Borron P, Veldhuizen RA, Lewis JF, Possmayer F, Caveney A, Inchley K,
McFadden RG, Fraher LJ: Surfactant associated protein-A inhibits human
lymphocyte proliferation and IL-2 production. Am J Respir Cell Mol Biol
15:115–121, 1996.
16. McIntosh JC, Mervin-Blake S, Conner E, Wright JR: Surfactant protein A
protects growing cells and reduces TNF
␣
activity from LPS-stimulated
macrophages. Am J Physiol 271:L310–L319, 1996.
17. Kremlev SG, Umstead TM, Phelps DS: Effects of surfactant protein A and
surfactant lipids on lymphocyte proliferation in vitro.Am J Physiol
267:L357–L364, 1994.
18. Voss T, Melchers K, Scheirle G, Schaefer KP: Structural comparison of
recombinant pulmonary surfactant protein SP-A derived from two human
coding sequences: implications for the chain composition of natural
human SP-A. Am J Respir Cell Mol Biol 4:88–92, 1991.
19. Stamme C, Wright JR: Surfactant protein A enhances the binding and
deacylation of E. coli LPS by alveolar macrophages. Am J Physiol
276:L540–L547, 1999.
20. Renz H, Gong JH, Schmidt A, Nain M, Gemsa D: Release of tumor
necrosis factor-
␣
from macrophages. Enhancement and suppression are
dose-dependently regulated by prostaglandin E
2
and cyclic nucleotides. J
Immunol 141:2388–2393, 1988
21. Furchgott RF, Jothianandan D: Endothelium-dependent and -independent
vasodilation involving cyclic GMP: relaxation induced by nitric oxide,
carbon monoxide and light. Blood Vessels 28:52–61, 1991.
22. Sakai N, Milstien S: Availability of tetrahydrobiopterin is not a factor in
the inability to detect nitric oxide production by human macrophages.
Biochem Biophys Res Commun 193:378–383, 1993.
23. Schneemann M, Schoedon G, Hofer S, Blau N, Guerrero L, Schaffner A:
Nitric oxide synthase is not a constituent of the antimicrobial armature of
human mononuclear phagocytes. J Infect Dis 167:1358–1363, 1993.
24. Coburn RF, Williams WJ, White P, Kahn SB: The production of carbon
monoxide from hemoglobin in vivo.J Clin Invest 46:346–356, 1967.
25. Lautier D, Luscher P, Tyrrell RM: Endogenous glutathione levels modu-
late both constitutive and UVA radiation/hydrogen peroxide inducible
expression of the human heme oxygenase gene. Carcinogenesis 13:227–
232, 1992.
SHOCK SEPTEMBER 2000 EFFECT OF SP-A ON PULMONARY MACROPHAGES 305
26. Gemsa D, Woo CH, Fudenberg HH, Schmid R: Stimulation of heme
oxygenase in macrophages and liver by endotoxin. J Clin Invest 53:647–
651, 1974.
27. Mendez C, Garcia I, Maier RV: Oxidants augment endotoxin-induced
activation of alveolar macrophages. Shock 6:157–163, 1996.
28. Katsura H, Kawada H, Konno K: Rat surfactant apoprotein A (SP-A)
exhibits antioxidant effects on alveolar macrophages. Am J Respir Cell
Mol Biol 9:520–525, 1993.
DISCUSSION
DR. MAIER: I appreciate the opportunity to discuss this well-performed,
extensive study. As the authors have shown, there is an impact of human SPA
on pulmonary macrophage function in the lung with a selective effect on the
intra-alveolar versus the interstitial populations. A significant advantage to
their studies is that they used human tissues as a source for cells and awesome
products. They used normal tissue and, as such, avoided many of the potential
artifacts and complicating factors that have been reported in the literature.
They show that SPA inhibits the production of both macrophage population’s
production of TNF and a selective inhibition of IL-1 in the interstitial macro-
phage, while there is no effect on IL-6 or nitric oxide production.
As published previously, the authors have demonstrated a carbon monox-
ide-dependent induction in cGMP, which appears to be the mechanism for
upregulating TNF in the alveolar macrophage. In the current study SPA blocks
cGMP and TNF in the alveolar macrophage population, yet does not block
carbon monoxide production. Therefore inhibition of CO production in the
alveolar macrophage is not the cause of the decreased GMP. Do the authors
have any suggestions or data to help us understand at what point SPA is
working in the alveolar macrophage that inhibits TNF production but not
carbon monoxide.
I would also ask the authors to extend their comments on the LPS interac-
tions in their experiments. The amount of LPS used was extremely high, at 15
mcg/ml. We and others have shown maximal production after 10 to 100 ng of
LPS in human cells rather than these enormous doses. There are recent data
showing that at extremely high doses, LPS may signal the macrophage through
both the CD-14-dependent pathway and also a Toll protein-independent path-
way. Do the authors have data at lower doses of LPS which would avoid this
complicating dual pathway signaling to confirm SPA has an effect on the
CD-14-dependent pathway? In addition, in their experimental methods I could
not tell if they provided serum and therefore a source for lipopolysaccharide-
binding protein (LPB). Was there serum present in their experimental condi-
tions? Without LBP, it is highly likely that LPS was not signaling through
CD-14.
Lastly, while SPA appears to have an effect, SPA, as they pointed out,
constitutes only 5 to 10% of total surfactant in vivo. It is well known that the
other phospholipids in surfactant also have significant LPS-binding and
immunomodulatory effects. Have the authors added back increasing doses of
SPA to depleted surfactant to establish the relative contribution of SPA to
intact surfactant? This would establish a clinical relevance to SPA separate
from the effect of surfactant overall?
I would like to thank the Society for the opportunity to discuss this paper.
DR. ARIAS-DIAZ: Thank you, Dr. Maier, for your interesting comments.
In the first place, previous studies of our group about carbon monoxide regu-
lating TNF production were performed exclusively using the interstitial variety
of macrophages. So the different behavior of carbon monoxide in alveolar
macrophages was an unexpected finding. SP-A was able to suppress the LPS-
induced production of carbon monoxide on interstitial macrophages, but not on
alveolar macrophages. However, SP-A antagonized the LPS-induced increase
in cGMP and TNF in both types of cells. In our opinion, other guanylate-
cyclase-stimulating molecule, different from nitric oxide and carbon monox-
ide, could be the responsible for the cGMP increase in alveolar macrophages.
We currently haven’t any other explanation for this finding.
With respect to the LPS concentration used by us, I agree it is a high LPS
concentration, but we think it may be closer to the one present in pathological
conditions. In the purulent exudate of a gram-negative pneumonia, LPS
concentration is known to be very high. At 10 micrograms per mL, LPS might
bind to membranes and different receptors than CD-14. Nevertheless, it contin-
ues to be a valid conclusion that SP-A is able to overcome some of the LPS
effects either CD-14-specific or unspecific.
With regard to the presence of serum in the cultures, we add 10% fetal calf
serum to the RPMI-1640 media.
Finally, regarding whether we have studied the effect of SP-A in combi-
nation with the lipid component of surfactant, yes we did. We have repeated
some of these experiments with SP-A with and without either phosphatidyl-
choline or dipalmitoylphosphatidylcholine. We found that neither phosphati-
dylcholine nor dipalmitoylphosphatidylcholine had any effect on TNF produc-
tion in the absence of SP-A. Interestingly, they did not interfere with the effect
of SP-A on the production of cytokines by LPS-stimulated pulmonary macro-
phages, in spite that we had previously found that SP-A binds and aggregates
phospholipid vesicles in the RPMI-1640 medium.
306 SHOCK VOL. 14, NO.3 ARIAS-DIAZ ET AL.