Differential Regulation of Cytokine Release and LeukocyteMigration by Lipopolysaccharide-Stimulated Primary HumanLung Alveolar Type II Epithelial Cells and Macrophages1
Andrew J. Thorley,* Paul A. Ford,* Mark A. Giembycz,† Peter Goldstraw,‡ Alan Young,§
and Teresa D. Tetley2*
Bacterial colonization is a secondary feature of many lung disorders associated with elevated cytokine levels and increasedleukocyte recruitment. We hypothesized that, alongside macrophages, the epithelium would be an important source of thesemediators. We investigated the effect of LPS (0, 10, 100, and 1000 ng/ml LPS, up to 24 h) on primary human lung macro-phages and alveolar type II epithelial cells (ATII; isolated from resected lung tissue). Although macrophages produced higherlevels of the cytokines TNF-� and IL-1� (p < 0.0001), ATII cells produced higher levels of chemokines MCP-1, IL-8, andgrowth-related oncogene � (p < 0.001), in a time- and concentration-dependent manner. Macrophage (but not ATII cell)responses to LPS required activation of ERK1/2 and p38 MAPK signaling cascades; phosphorylated ERK1/2 was constitu-tively up-regulated in ATII cells. Blocking Abs to TNF-� and IL-1� during LPS exposure showed that ATII cell (notmacrophage) MCP-1 release depended on the autocrine effects of IL-1� and TNF-� (p < 0.003, 24 h). ATII cell release ofIL-6 depended on autocrine effects of TNF-� (p < 0.006, 24 h). Macrophage IL-6 release was most effectively inhibited whenboth TNF-� and IL-1� were blocked (p < 0.03, 24 h). Conditioned media from ATII cells stimulated more leukocytemigration in vitro than conditioned media from macrophages (p < 0.0002). These results show differential activation ofcytokine and chemokine release by ATII cells and macrophages following LPS exposure. Activated alveolar epithelium is animportant source of chemokines that orchestrate leukocyte migration to the peripheral lung; early release of TNF-� andIL-1� by stimulated macrophages may contribute to alveolar epithelial cell activation and chemokine production. TheJournal of Immunology, 2007, 178: 463– 473.
B acterial colonization of the lung, both acutely and chron-ically, is associated with a worsening of symptoms inmany pulmonary and systemic disorders. It is estimated
that �50% of all subjects with chronic obstructive pulmonary dis-ease (COPD)3 have a persistent colonization of the lower respira-tory tract by bacteria such as Streptococcus pneumoniae, Hae-mophilus influenzae, and Moraxella catarrhalis (1), which is oftenassociated with periods of exacerbation leading to an accelerateddecline in lung function. Similarly, in the lungs of cystic fibrosissufferers there is a persistent and aggressive bacterial colonizationby the aforementioned bacteria as well as Staphylococcus aureus,Pseudomonas aeruginosa, and Burkholderia cepacia. In onestudy, 97.5% of children with cystic fibrosis were infected by the
age of 3 years with P. aeruginosa (2). In systemic diseases such asHIV and AIDS, the lung is a common site of secondary compli-cations due to opportunistic infection which is a frequent cause ofmorbidity (3). In a recent study of hospitalized HIV patients,nearly 10% had bacterial pneumonia, some of which had repeatepisodes of infection. These cases were attributable to a total of 12different Gram-negative and 5 Gram-positive strains of bacteria (4).
The cycle of inflammation that is likely to play a key role in thedeterioration of the lung during bacterial infection is, withoutdoubt, complex and multifactorial, (5) involving a number of cellswithin the lung such as neutrophils, macrophages, and epithelialcells (6). The neutrophil is a rich source of proinflammatory me-diators which can be released acutely and rapidly following acti-vation of the cell by both exogenous and intrinsic factors (7). Sim-ilarly, the macrophage can release high levels of cytokines andchemokines following exposure to proinflammatory agents whichincludes exposure to bacterial cell wall products (8). Previous stud-ies in animals, including humans, have also shown that macro-phages, as well as phagocytosing bacteria, can release a number ofantimicrobial substances to aid bacterial killing and clearance (9–11). Unlike the neutrophil, the macrophage has a much longerhalf-life within the lung, continually producing mediators and re-sponding to repeated stimulation over long periods of time (12).The role of the alveolar epithelium, however, has been less wellcharacterized in this respect. Historically, the alveolar epitheliumwas considered to have only two primary functions: that of a bar-rier and as the source of pulmonary surfactant. In addition, studiesin animals and cell lines have shown that these cells are able torelease a variety of mediators following stimulation with proin-flammatory agents (13, 14). Previous studies by us, using primaryhuman alveolar epithelial type II (ATII) cells, have also shown that
*Lung Cell Biology, Section of Airways Disease, National Heart and Lung Institute,Imperial College, London, United Kingdom; †Department of Pharmacology and Ther-apeutics, Institute of Infection, Immunity and Inflammation, University of Calgary,Calgary, Alberta, Canada; ‡Department of Thoracic Surgery, Royal Brompton andHarefield National Health Service Trust, London, United Kingdom; and §AstraZenecaR&D, Loughborough, United Kingdom
Received for publication November 18, 2005. Accepted for publication October17, 2006.
The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 A.J.T. and P.A.F. were supported by a grant from AstraZeneca.2 Address correspondence and reprint requests to Dr. Teresa D. Tetley, Lung CellBiology, Section of Airways Disease, National Heart and Lung Institute, ImperialCollege, London, U.K. SW3 6LY. E-mail address: [email protected] Abbreviations used in this paper: COPD, chronic obstructive pulmonary disease;ATII, alveolar epithelial type II; GRO, growth-related oncogene; PSG, penicillin/streptomycin/glutamine; KO, knockout.
the alveolar epithelium is a rich source of chemokines (15, 16).Indeed, ATII cell chemokine release increases dramatically fol-lowing exposure to LPS, reflecting the cell surface expression ofTLR4 (17).
There have been numerous investigations that focused on thefunctional contribution to pulmonary homeostasis of a single celltype (e.g., the macrophage, epithelial cell). However, to ourknowledge, there are no comparative studies of human lung mac-rophages and epithelial cells from the same subjects; yet, it isimportant to understand how these cells complement each other inmaintaining homeostasis of the alveolar unit. Thus, we isolatedprimary human alveolar macrophages and ATII cells from thespecimens of lung tissue obtained from sequential subjects to com-pare the response of these cells to LPS and to determine how thismay influence leukocyte recruitment. We hypothesized that thesecells would differentially release cytokines (IL-1�, TNF-�, andIL-6) and chemokines (IL-8, MCP-1, and growth-related oncogene� (GRO�)) following LPS exposure and this, in turn, would gov-ern peripheral blood leukocyte migration in an in vitro model.
Materials and MethodsIsolation of primary human alveolar macrophages and type IIepithelial cells
ATII cells and macrophages were isolated from lungs of grossly normalappearance following resection for lung carcinoma, with the approval ofthe Royal Brompton and Harefield Ethical Committee, as previously de-scribed (n � 16 consecutive subject samples; Ref. 15). Briefly, lung sec-tions were perfused by injection of sterile saline until the cell count was�1 � 104 cells/ml. The draining lavage was then collected and centrifuged(290 � g, 10 min, 20°C). The cell pellet was resuspended in serum-freeDCCM-1 (React Scientific) containing 1% penicillin/streptomycin/glutamine (PSG; Invitrogen Life Technologies) and plated in 24-well tissueculture plates (0.5 � 106 macrophages/well; VWR). After 3 h, macro-phages had adhered and the medium was removed and the wells werewashed to remove nonadherent cells. The macrophages were maintained inserum-free DCCM-1 plus 1% PSG. We have shown previously that �95%of these cells are CD68� and are, therefore, macrophages (16). These cellsare also phagocytic and show macrophage morphology on scanning elec-tron microscopy (18).
To isolate ATII cells, tissue was perfused and inflated with trypsin(0.25% in HBSS; Sigma-Aldrich) and incubated at 37°C for 45 min; tryp-sin was replaced twice during this time. The tissue was finely chopped inthe presence of newborn calf serum (Invitrogen Life Technologies). Thechopped tissue was then incubated with DNase (250 �g/ml; Sigma-Aldrich) and the mixture was passed through a 300-�m filter, followed bya 40-�m filter to remove large tissue debris. The cell suspension was thencentrifuged (290 � g, 10 min, 20°C) and the resulting pellet was resus-pended in DCCM-1 medium containing 50 �g/ml DNase. These cells wereincubated in tissue culture flasks for 2 h at 37°C in a humidified incubatorto allow differential adherence of contaminating mononuclear cells.
After 2 h, the nonadherent ATII cells were removed and the cell sus-pension was centrifuged as before. The cell pellet was then resuspended inDCCM-1 containing 10% newborn calf serum and 1% PSG at a concen-tration of 1 � 106 cells/ml. Cells were then seeded at 1 � 106 ATII cells/well. Cells reached confluence by 48 h. These cells have been thoroughlycharacterized using electron microscopy, staining for alkaline phosphatase,and expression of surfactant proteins A and C (15, 19). Thus, on electronmicroscopy the cells are cuboidal in morphology and have surfactant-con-taining lamellar bodies, tight junctions, and microvilli.
Cell density
Monolayers of alveolar macrophages and ATII cells were visualized underlight microscopy and photographed at �40 magnification. For each well,four images of the field of vision were taken. These images could then betransferred to a computer screen for cell counting. Results showed thatdespite the different cell seeding concentrations there was no significantdifference between the cell density on the plate (number per unit area �SD, ATII 161 � 6.8; alveolar macrophage 159 � 5.0; p � 0.05). This wasa consistent feature, even between subjects, of the in vitro model system,because ATII cells were always used at confluence when cell density wasself-limiting and macrophages were always seeded at the same density.This indicated that levels of protein mediators detected by ELISA in the
cell supernatant in picograms per milliliter did not need to be adjusted toaccount for difference in cell numbers, which were directly comparablebetween cell types.
Isolation of monocytes from peripheral blood
One hundred milliliters of venous blood was collected on three separateoccasions from three healthy volunteers and the erythrocytes were removedusing dextran sedimentation. The resulting leukocyte-rich plasma was lay-ered onto 2.5 ml of Nycoprep 1.068 (Robbins Scientific) and centrifuged(600 � g, 15 min, 20°C). The plasma and cells were then collected intoseparate tubes and centrifuged (700 � g, 10 min, 20°C). The monocytepellet was then resuspended in platelet-poor plasma (5% plasma, 95% sa-line) and centrifuged again (78 � g, 10 min, 20°C) to remove platelets.This was repeated four times.
The resulting cells were resuspended in DCCM-1 containing 1% PSGand 10% FCS at 1 � 106/ml and plated in 6-well tissue culture plates at 2 �106/well. After 1.5 h, the wells were washed to remove nonadherent cells.Following this, cells were �95% pure.
Isolation of neutrophils from peripheral blood
Leukocyte-rich plasma was obtained as described above. Following sedi-mentation, the leukocyte-rich plasma was centrifuged (400 � g, 10 min,4°C) and the supernatant was discarded. The cell pellet was resuspended inPBS and centrifuged for a second time as before. Following washing, thecell pellets were ready for separation using a discontinuous Percoll gradient(Sigma-Aldrich).
Using a discontinuous gradient consisting of three layers (81, 70, and55% v/v) the cell pellet was resuspended in 3 ml of 55% v/v (the top layer)which was then overlaid onto the preprepared gradient. The cells were thencentrifuged (750 � g, 20 min, 4°C) and the neutrophils were harvestedfrom the 70%/81% interface and washed in PBS for use in migrationassays.
LPS stimulation of alveolar macrophages and ATII cells
Cells from six sequential patient tissue samples were used for this study.ATII cells and macrophages were cultured in serum free DCCM-1 for the24 h before addition of LPS (Escherichia coli 055:B5; Sigma-Aldrich). Themedium was then removed and the cells were incubated with LPS at con-centrations of 10, 100, or 1000 ng/ml, in serum-free DCCM-1. Each treat-ment was performed in triplicate. The resulting conditioned medium wasaspirated and the secreted cytokines and chemokines were measured byELISA.
Conditioned medium was also generated from ATII epithelial cells andmacrophages from three further subjects, to use in the monocyte and neu-trophil migration assays. Following stimulation for 24 h with 100 ng/mlLPS, the medium was removed and the cells were washed thoroughly toremove any residual LPS. Fresh serum-free medium was then added andthe cells were cultured for a further 24 h. The LPS-free conditioned me-dium from this 24-h time period was then used for leukocyte migrationassays.
Effect of Ab blockade of IL-1� and TNF-� on LPS stimulationof alveolar macrophages and ATII cells
Cells from three further subjects were cultured as described above andserum-starved for 24 h before experiments. The medium was then removedand cells were incubated with 100 ng/ml LPS in the presence or absence ofneutralizing Abs to IL-1� and/or TNF-� (R&D Systems). These were usedat concentrations of 10 and 20 �g/ml, respectively, in accordance withsupplier’s recommendations based on concentration-response neutraliza-tion studies of recombinant IL-1� and TNF-�. As an appropriate control,cells were also exposed to the relevant carrier protein, in this case mouseIgG1 (R&D Systems).
MAPK expression
Cells from a further four subjects were stimulated with LPS at a concen-tration of 100 ng/ml. The cells were harvested at regular intervals for up to4 h. The cells were washed and harvested by scraping into lysis buffer(0.05% deoxycholate, 0.1% Nonidet P-40, 0.1% Triton X-100, and 0.025%w/v SDS; Sigma-Aldrich) containing the phosphatase inhibitors sodiumpyrophosphate (1 mM) and sodium orthovanadate (2 mM) plus the pro-tease inhibitors PMSF (1 mM), aprotinin (25 �g/ml), and leupeptin (10�g/ml; Sigma-Aldrich). After centrifugation at 4500 � g at 4°C for 10 minto remove cellular debris, samples were stored at �70°C until analysis.Phospho-ERK 1/2 and p38 MAPK levels in the cell lysates were deter-mined by immunoblotting.
464 DIFFERENTIAL CYTOKINE RELEASE BY ATII CELLS AND MACROPHAGES
Following the quantification of total protein (Bio-Rad protein assay), thesamples were diluted 1/6 with a 4% loading buffer (containing 2-ME, glyc-erol, SDS, and bromphenol blue in Tris buffer; Invitrogen Life Technolo-gies) and boiled for 5 min. Rainbow m.w. markers (Invitrogen Life Tech-nologies) were used to determine the m.w. of the immunoreactive bands.Forty micrograms of protein was loaded per lane on a 4–12% bis-acryl-amide gel (Invitrogen Life Technologies) and then electrophoresed con-tinuously at 80 mA for �50 min in running buffer (Invitrogen Life Tech-nologies) using a Bio-Rad Mini Protean II Electrophoresis System. Afterseparation, the protein bands were electrophoretically transferred, at 400mA per gel for 1 h to a nitrocellulose membrane (Invitrogen Life Tech-nologies) using a transfer buffer (Invitrogen Life Technologies).
Nitrocellulose blots were agitated with TBST (20 mM Tris, 150 mMNaCl, and 0.1% Tween 20; Sigma-Aldrich) plus 5% milk for 1 h to blocknonspecific binding sites. Nitrocellulose blots were washed three times for10 min with TBST and incubated overnight at 4°C with specific mAbs(Santa Cruz Biotechnology) to total ERK 1/2, to phosphorylated ERK 1/2,to total p38 and to phosphorylated p38 (in TBST plus 5% milk); isotypeAbs (Santa Cruz Biotechnology) were used as control to detect nonspecificbinding. Following washing, the nitrocellulose blots were incubated withHRP-conjugated secondary Abs (in TBST plus 5% milk; Santa Cruz Bio-technology) to the primary Ab. The blots were then washed again as de-scribed above and developed using the ECL detection system (UpstateBiotechnology). Equal sample loading was confirmed by probing for thehousekeeping protein GAPDH (data not shown). The degree of phosphor-ylation was determined by densitometry using Labworks software.
Measurement of cytokines and chemokines by ELISA
DuoSet (R&D Systems) kits consisting of paired Abs were used to measurecytokine and chemokine release from ATII cells and alveolar macrophagesfollowing LPS stimulation. The threshold limit of detection of the assays is15.6 pg/ml for TNF-� and MCP-1, 4.7 pg/ml for IL-6, 3.9 pg/ml for IL-1�,
and 31.25 pg/ml for IL-8 and GRO�. The interassay coefficient of variancewas �5% for all assays conducted.
Leukocyte migration assays
Neutrophils and monocytes were isolated as described above and resus-pended in DCCM-1 medium at a concentration of 1 � 106/ml. A total of150 �l of the cell suspension (either neutrophils or monocytes) was thenplaced into the upper chamber of a Transwell insert (8-�m pore membrane)and 300 �l of conditioned medium was placed directly into the 24-wellplates (i.e., in the lower chamber). Plates were then incubated overnight ina humidified incubator at 37°C.
Following incubation, the cell culture insert was removed and the non-migratory cells in the upper chamber were aspirated and wiped away fromthe membrane. The underside of the membrane was then examined foradherent migratory cells; this was always negative. Migratory leukocytes inthe lower chamber were counted under a light microscope. Migration wasexpressed as cells per field of vision, n � 5 fields/well.
The possible effect of chemokinesis (i.e., nonspecific migration) wasdetermined by preincubating the leukocytes with conditioned medium in atissue culture insert for 15 min, before addition of the same conditionedmedium to the lower chamber. The assay was then conducted as describedabove.
Anti-chemokine Ab blockade of leukocyte migration
The effect of supermaximal levels of blocking mouse mAbs to GRO� (50�g/ml), IL-8 (25 �g/ml), RANTES (5 �g/ml), and MCP-1 (10 �g/ml) wasdetermined by addition of Abs (R&D Systems) to conditioned medium inthe lower chamber of a parallel set of experiments. The supermaximal Ablevels that were used were based on the concentration of chemokine in theconditioned medium as recommended by the supplier, to completely neu-tralize chemokine activity. As the appropriate negative controls, the effectof addition of the relevant IgG isotypes (IgG1 and IgG2A; R&D Systems)to the conditioned media was also investigated.
FIGURE 1. LPS-induced time- anddose-dependent release of IL-1� (A),TNF-� (B), and IL-6 (C) by primaryhuman alveolar macrophages andATII cells. Cells were cultured with in-creasing doses of LPS for up to 24 h.Conditioned media were aspirated ateach time interval and assayed for cy-tokine release by ELISA. Data are ex-pressed as picograms per milliliter ofconditioned media, mean � SE (n � 6subjects). The asterisks denote signifi-cant differences between the untreatedand LPS-treated cells at each timeinterval: �, p � 0.05; ��, p � 0.005;���, p � 0.0005.
Data are presented as mean � SE. A one-way ANOVA was used to ana-lyze the time- and concentration-dependent effects of LPS on cytokine andchemokine release from ATII cells and alveolar macrophages. Paired t testswere used to determine significant differences in LPS-induced mediatorrelease following addition of neutralizing Abs to IL-1� and TNF-�. Thisstatistical test was also used to evaluate the significance of addition ofchemokine neutralizing Abs in migration assays. A p value � 0.05 wasconsidered to be statistically significant.
ResultsCytokine release from primary human alveolar macrophagesand type II epithelial cells
IL-1� (Fig. 1A). Following 1000 ng/ml LPS exposure for 24 h,macrophages released significantly more (over than twice asmuch) IL-1� than ATII cells ( p � 0.0001). Release by macro-phages increased significantly in a concentration-dependent man-ner at each time point, but was most marked at 12 h ( p � 0.0001).In contrast, ATII cell production of IL-1� plateaued at 6 and 12 h,but showed the most marked concentration-dependent response at24 h ( p � 0.02). Macrophages also released IL-1� basally whereasATII cells did not.
TNF-� (Fig. 1B). Similarly, macrophages released significantlymore (approximately twice as much) TNF-� as ATII cells in re-sponse to 1000 ng/ml LPS for 24 h ( p � 0.0001); productionincreased steadily over the 24 h of the study. Although ATII cellrelease of TNF-� increased steadily over 24 h when exposed to1000 ng/ml LPS, lower concentrations of LPS induced maximalTNF-� release by 12 h, in a concentration-dependent manner ( p �
0.0004). ATII cells did not produce TNF-� basally, unlike mac-rophages, and did not release TNF-� in response to 1 ng/ml LPS.
IL-6 (Fig. 1C). ATII cells released significantly more (approxi-mately four times as much) IL-6 as macrophages ( p � 0.0001),and release increased in a time- and concentration-dependent man-ner over 24 h ( p � 0.0002). In macrophages, however, release ofIL-6, while also concentration-dependent ( p � 0.001), began toplateau after 6 h. Both cell types produced IL-6 basally.
Chemokine release from primary human alveolar macrophagesand type II epithelial cells
IL-8 (Fig. 2A). LPS-stimulated ATII epithelial cells released sig-nificantly more (approximately three times as much) IL-8 as mac-rophages (LPS: 1000 ng/ml, 24 h, p � 0.0001). Release of IL-8from ATII cells was concentration dependent ( p � 0.0008) andcontinued to increase over the 24 h, with a sharp increase between6 and 12 h, whereas in macrophages it started to plateau after 6 h.Although the response to LPS by macrophages was much flatter,with very little difference between control and LPS-stimulatedcells, there was a significant concentration-dependent response( p � 0.0001). IL-8 was released basally by both cell types but ingreater amounts by macrophages.
MCP-1 (Fig. 2B). ATII epithelial cells released significantlymore (approximately six times more) MCP-1 than alveolar mac-rophages (LPS 1000 ng/ml, 24 h, p � 0.0001). There was a strik-ing increase in MCP-1 release by ATII cells following LPS expo-sure (�7.5-fold, 1000 ng/ml LPS) over the course of theexperiment. In direct contrast, even following 1000 ng/ml LPS,
FIGURE 2. LPS-induced time- anddose-dependent release of IL-8 (A),MCP-1 (B), and GRO� (C) by pri-mary human alveolar macrophagesand ATII cells. Cells were culturedwith increasing doses of LPS for up to24 h. Conditioned media were aspi-rated at each time interval and as-sayed for cytokine release by ELISA.Data are expressed as picograms permilliliter conditioned media, mean �SE (n � 6 subjects). The asterisks de-note significant differences betweenthe untreated and LPS-treated cells ateach time interval: ��, p � 0.005;���, p � 0.0005.
466 DIFFERENTIAL CYTOKINE RELEASE BY ATII CELLS AND MACROPHAGES
macrophages only released �1000 pg/ml after 24 h. Both celltypes released MCP-1 basally and in response to LPS stimulationin a time and concentration-dependent manner (ATII, p � 0.003;alveolar macrophages, p � 0.002). ATII cell release was steadyover the full time course whereas, for macrophages, there was amarked increase in release between 12 and 24 h.
GRO� (Fig. 2C). ATII cells released significantly more GRO�(approximately four times more) than macrophages (1000 ng/mlLPS, 24 h, p � 0.001). Release from both cell types was time- andconcentration-dependent ( p � 0.0001 for both cell types). Releaseof GRO� from macrophages appeared to be steady and constantover the 24 h, whereas in ATII cells there was a dramatic increasein release after 6 h (�3.5-fold, 1000 ng/ml LPS) which continuedfor up to 24 h. Both cell types released GRO� basally.
Ab blockade of cytokine and chemokine release from alveolarmacrophages and ATII cells
Previous studies have implied that many of the proinflammatoryeffects of LPS are mediated indirectly through release of IL-1� andTNF-� which then feed back to stimulate the cells in an autocrinemanner. To investigate this possibility in these experiments, neu-tralizing Abs to IL-1� and TNF-� were added to the media alongwith 100 ng/ml LPS.
Neutralizing Abs to TNF-� inhibited IL-6 release from alveolarmacrophages at 12 h ( p � 0.03) and neutralization of both cyto-kines significantly inhibited release at 12 and 24 h ( p � 0.03 andp � 0.003, respectively). Neutralization of IL-1� alone had noeffect (Fig. 3A), although at 24 h the inhibition of IL-6 releasefollowing IL-1� blockade was significant at the p � 0.1 level ( p �0.09). In ATII cells, neutralization of IL-1� had no effect on IL-6release, whereas neutralization of TNF-� significantly inhibitedrelease at 24 h ( p � 0.006). Neutralization of both cytokines con-comitantly had no increased effect over neutralization of TNF-�alone (Fig. 3B). When supernatants were assayed for MCP-1 re-lease, results showed that neutralization of IL-1� and TNF-� aloneor in combination had no effect on release of MCP-1 from alveolarmacrophages (Fig. 4A). However, ATII cell release of MCP-1 wasinhibited by neutralization of TNF-� at 12 and 24 h ( p � 0.03 andp � 0.01, respectively), while neutralization of IL-1� inhibitedMCP-1 release at 24 h ( p � 0.04). Neutralization of both cyto-kines in tandem inhibited MCP-1 release at 12 and 24 h ( p �0.004 and p � 0.0002, respectively) and appeared to have a smalladditive effect ( p � 0.006; Fig. 4B), bringing MCP-1 levels backdown to those observed at 3 h. Addition of the Ab carrier proteinalone to the media had no effect on mediator release and thus ruled
FIGURE 3. Time course of inhibition of IL-6 release by LPS-exposedalveolar macrophages (A) and ATII cells (B) in the presence and absenceof neutralizing Abs to IL-1� and TNF-�. Cells were exposed to 100 ng/mlLPS alone and in combination with neutralizing Abs to IL-� and TNF-�.IgG alone had no effect on IL-6 release. Data are expressed as mean � SE,n � 3 subjects. The asterisks denote significant differences between neu-tralizing Ab-treated and neutralizing Ab-untreated cells at each time inter-val. Macrophage minus LPS control vs LPS plus anti-TNF-� 12h; LPScontrol vs LPS plus both Abs 12 and 24 h, �, p � 0.03; ATII cells minusLPS control vs LPS plus anti-TNF-� and both Abs 24 h, ��, p � 0.006.
FIGURE 4. Time course of inhibition of MCP-1 release by LPS-ex-posed alveolar macrophages (A) and ATII cells (B) in the presence andabsence of neutralizing Abs to IL-1� and TNF-�. Cells were exposed to100 ng/ml LPS both alone and in combination with neutralizing Abs toIL-� and TNF-�. IgG alone had no effect on MCP-1 release. Data areexpressed as mean � SE, n � 3 subjects. The asterisks denote significantdifferences between neutralizing Ab-treated and neutralizing Ab-untreatedcells at each time interval. ATII cells minus LPS control vs anti-TNF-� at12 and 24 h and vs anti-IL-1� at 24 h, �, p � 0.04; ATII cells minusLPS control vs both Abs at 12 h, ��, p � 0.004; LPS control vs bothAbs at 24 h, ���, p � 0.0002.
out the possibility of any nonspecific inhibitory effects it may have(data not shown).
LPS activation of MAPK
Under resting conditions phospho-ERK-2 was not detected in al-veolar macrophages; however ERK-1 was modestly phosphory-lated relative to the total ERK-1 pool. Exposure of alveolar mac-rophages to LPS (100 ng/ml) resulted in a time-dependentphosphorylation of ERK-1 and ERK-2 (Fig. 5). This effect wasdetectable as early as 5 min after stimulation, was maximal at 30min, and was still detectable at 120 min (pERK-2/tERK 30 min vs120 min; p � 0.036, Fig. 5). p38 MAPK, which was not activatedin resting cells, was also phosphorylated by LPS. Again, this effectwas detected 5 min after stimulation, peaked at 30 min, and thendeclined over the remaining 210 min of the experiment (Fig. 5).Due to the low levels of p38 phosphorylation, densitometry wasnot accurate enough to calculate the phospho-p38/p38 ratios. Incontrast to the findings with macrophages, there was high consti-tutive phosphorylation of ERK-1 and ERK-2 by ATII cells andLPS did not induce a measurable change at this high basal level ofactivation (Fig. 6). Phosphorylated p38 MAPK was difficult todetect in ATII cells and did not alter following stimulation withLPS (Fig. 6).
Leukocyte migration in response to conditioned media fromLPS-stimulated primary human alveolar macrophages and typeII epithelial cells
Chemokinesis. To validate our assay system and assess the con-tribution of chemokinesis to the cell migration observed, neutro-phils and monocytes were preincubated with conditioned mediumfor 30 min before their introduction to the invasion chamber con-taining conditioned medium in the lower chamber.
Following preincubation of monocytes with conditioned me-dium, there was a small but significant increase in migration ( p �0.007; data not shown); however, this was significantly less thanthat of naive monocytes that had not been preincubated with con-ditioned medium ( p � 0.0001). Similarly, in the neutrophil mi-gration assay, following preincubation with conditioned mediumfrom both alveolar macrophages and ATII cells, there was a smallbut significant increase in neutrophil migration ( p � 0.001, p �0.0001, respectively; data not shown). Again, migration of naiveneutrophils toward conditioned medium was significantly greater(�4-fold greater, p � 0.001). This suggests that a small proportionof both monocytes and neutrophils migrate across the membrane in-dependently of chemotactic mediators in conditioned medium. Con-sequently, the data have been corrected by subtraction of nonspecificmigration (i.e., migration following preincubation of cells in condi-tioned media) from that induced by conditioned media (i.e., naive cellmigration toward conditioned media) to provide migration due to fac-tors released into the media by macrophages and ATII cells. Thus,migration of both monocytes and neutrophils toward conditioned me-dium from alveolar macrophages and ATII cells was significantlyincreased ( p � 0.0001 monocytes, p � 0.001 neutrophils).
Ab blockade of leukocyte migration. Following validation of themigration assay system, Abs to chemokines detected in the con-ditioned media (IL-8, MCP-1, GRO�, and RANTES) were addedalone or in combination to the lower chamber in an attempt toinhibit both monocyte and neutrophil migration. To exclude anyeffect of the Ab carrier protein on leukocyte migration, leukocyteswere also exposed to the conditioned medium in the presence andabsence of the relevant nonimmune IgGs. There was no effect of
FIGURE 5. Expression of phosphorylated ERK1/2 and p38 MAPK bymacrophages in response to LPS. Macrophages were exposed to 100 ng/mlLPS for up to 240 min. The immunoblotting of ERK1/2 and p38 MAPK forone subject sample is representative of data from all four subjects. The datafor the p-ERK/total ERK ratio was derived from all four subjects and wasquantified by densitometry, as shown in the line graph.
FIGURE 6. Expression of phosphorylated ERK1/2 and p38 by ATIIcells in response to LPS. ATII cells were exposed to 100 ng/ml LPS for upto 240 min. The immunoblotting of one subject sample is representative ofdata from all four subjects. The data for the p-ERK/total ERK ratio wasderived from all four subjects and was quantified by densitometry, asshown in the line graph.
468 DIFFERENTIAL CYTOKINE RELEASE BY ATII CELLS AND MACROPHAGES
IgG alone as no significant difference in migration was observed(data not shown).
Monocyte migration. As observed in the preliminary investiga-tion of chemokinesis, conditioned medium from LPS-stimulatedalveolar macrophages (Fig. 7A) and ATII cells (Fig. 7B) caused asignificant increase in monocyte migration. Conditioned mediumfrom ATII cells caused a significantly higher number of monocytesto migrate which correlated with the total amount of chemokinedetected ( p � 0.0002; r2 � 0.98). Addition of RANTES neutral-izing Abs had no significant effect on monocyte migration in re-sponse to conditioned media from either cell type. NeutralizingAbs to IL-8 caused significant inhibition of monocyte migrationtoward conditioned media from both alveolar macrophages andATII cells (29%, p � 0.01; 38.6%, p � 0.006, respectively). In-hibition of GRO� was similar to IL-8; however, inhibition wasslightly greater with ATII cell-conditioned media compared withthat from alveolar macrophages (51.3%, p � 0.0009; 35.5%, p �0.002, respectively). The greatest inhibition of monocyte migrationwas observed with Abs to MCP-1. As observed with the previousAbs, inhibition of monocyte migration in response to ATII cell-conditioned media was greater than the inhibition of migrationtoward alveolar macrophage-conditioned media (82.8%, p �0.0001; 70.1%, p � 0.0001). The addition of all Abs to the con-ditioned media from either ATII cells or alveolar macrophagescaused a small but nonsignificant increase in inhibition over that ofproduced by Abs to the major chemoattractive chemokine.
Neutrophil migration. Conditioned medium from LPS-stimu-lated primary human alveolar macrophages (Fig. 8A) and ATIIcells (Fig. 8B) caused a significant increase in neutrophil migration
as observed in the preliminary chemotaxis assays ( p � 0.0001).Conditioned medium from ATII cells, however, induced signifi-cantly greater numbers of neutrophils to migrate compared withthat of alveolar macrophages ( p � 0.01). The increase in migra-tion observed significantly correlated with the total chemokine lev-els measured in the conditioned media (r2 � 0.92).
In observations similar to that seen in the monocyte migrationassays, addition of neutralizing Abs to RANTES had no significanteffect on neutrophil migration. Neutralizing Abs to MCP-1 signif-icantly inhibited neutrophil migration toward alveolar macrophageand ATII cell-conditioned media (24.4%, p � 0.05; 37.3%, p �0.002, respectively). Abs to IL-8 and GRO� significantly inhibitedneutrophil migration. However, when neutralizing Abs to GRO�were added to macrophage-conditioned media, inhibition wasgreater than when Abs to IL-8 were used (53.8%, p � 0.008;40.9% p � 0.001, respectively). The opposite was true for ATIIcells; Abs to IL-8 caused greater inhibition than those to GRO�(61.3%, p � 0.0001; 52%, p � 0.0001, respectively). Addition ofall Abs to conditioned medium from either macrophages or ATIIcells caused a further small, but significant, decrease in neutrophilmigration over that observed with Abs against the major chemoat-tractive chemokine.
DiscussionThese studies show distinct differences between human alveolarmacrophages and human ATII epithelial cells in their response toLPS stimulation; this impacted on the magnitude of leukocyte mi-gration toward conditioned media from these cells, which differed,
FIGURE 7. Ab blockade of monocyte migration in response to condi-tioned media from LPS-stimulated alveolar macrophages (A) and ATIIcells (B). The effect of supermaximal levels of blocking mouse mAbs toGRO� (50 �g/ml), IL-8 (25 �g/ml), RANTES (5 �g/ml), and MCP-1 (10�g/ml) was determined by addition of Abs to conditioned media in thelower chamber of a parallel set of experiments to the migration assay.The effect of addition of the relevant IgG isotypes (IgG1 and IgG2A) tothe conditioned media was also determined and found to be negligible(data not shown). Data are expressed as cells per field of vision, mean �SE, n � 3 subjects. The asterisks denote significant differences betweenthe Ab blocked migration and maximum migration toward conditionedmedia: �, p � 0.05; ��, p � 0.005; ���, p � 0.0005.
FIGURE 8. Ab blockade of neutrophil migration in response to condi-tioned media from LPS-stimulated alveolar macrophages (A) and ATIIcells (B). The effect of supermaximal levels of blocking mouse mAbs toGRO� (50 �g/ml), IL-8 (25 �g/ml), RANTES (5 �g/ml), and MCP-1 (10�g/ml) was determined by addition of Abs to conditioned media in thelower chamber of a parallel set of experiments. The effect of addition of therelevant IgG isotypes (IgG1 and IgG2A) to the conditioned media was alsodetermined and found to be negligible (data not shown). Data are expressedas cells per field of vision, mean � SE, n � 3 subjects. The asterisks denote1) significant differences between the Ab blocked migration and maximummigration toward conditioned media and 2) the significant difference be-tween the most effective single Ab and the effect of all the Abs combined:�, p � 0.05; ��, p � 0.005; ���, p � 0.0005.
reflecting the profile of chemokines released by each cell type.Thus, under these experimental conditions, ATII cells producedhigher levels of the chemokines and leukocyte migration than al-veolar macrophages, whereas macrophages were a rich source ofIL-1� and TNF-�. Ab blockade of IL-1� and TNF-� showed thatLPS induction of these cytokines, and their autocrine effects, wereimportant for ATII cell, but less so for macrophage, cytokine, andchemokine release. In contrast, phosphorylation of components ofthe MAPK pathways, ERK1/2 and p38, following LPS occurred inmacrophages but not ATII cells, suggesting differential activationof signaling pathways by LPS. This work is unique in its use ofprimary human cells obtained from the same subjects to give a truecomparative study of alveolar macrophages and ATII cells, andclearly and unequivocally demonstrates significant differences be-tween these two cell types in their response to LPS and regulationof cytokine and chemokine release.
LPS triggers intracellular signaling via TLR4 activation. Wehave shown previously that human ATII cells express both TLR4and TLR2 (17), while others have shown expression of these re-ceptors by human alveolar macrophages (20). Alveolar macro-phages from smokers and COPD patients express normal TLR4levels, although TLR2 expression is decreased (20). The MAPKsignal transduction pathways are an important component of TLRsignaling which regulate cellular functions such as gene transcrip-tion, programmed cell death, and inflammation. Here, we haveshown that LPS induces rapid phosphorylation of p38 andERK1/2, by alveolar macrophages together with a marked time-and concentration-related increase in release of IL-1�, TNF-�, andIL-6 and a less marked, but similar, pattern of increased release ofIL-8, MCP-1, and GRO�. Previous investigations show that bothp38 MAPK and ERK1/2 are essential for primary human alveolarmacrophage and THP-1 cell TNF-� synthesis (21, 22). Inhibitionof activation of one or other of these MAPKs leads to partial re-duction in TNF-� levels, while inhibition of both MAPKs com-pletely prevents TNF-� synthesis (22, 23). However, JNK andNF-�B, not studied here, are also important for LPS-induced hu-man alveolar macrophage release of TNF-� (24, 25). LPS induc-tion of chemokine release from human macrophages, includingIL-8, GRO�, MIP-1�, and MCP-1 has previously been demon-strated to involve p38 MAPK (23, 26). When THP-1 cells werestimulated with shed factors from Helicobacter pylori, which con-tained LPS, IL-8 release was found to be dependent upon phos-phorylation of p38 MAPK, ERK1/2, and JNK, as well as NF-�B(27). Its effects were attributed to LPS as they were inhibited bypolymyxin and anti-CD14 Abs. Other compounds, including 8-iso-prostane, cigarette smoke, Bacillus anthracis, secretory phospho-lipase A2, and rhinovirus, induce release of a range of chemokines,including IL-8, MCP-1, MIP-1�, MIP-1�, and IFN-�-inducibleprotein 10 via differential activation of p38 MAPK, ERK1/2, andJNK (25, 26, 28–30).
In direct contrast to our findings in macrophages, there was con-stitutively high phosphorylation ERK1/2 in ATII cells, and therewas no detectable difference in phosphorylation of either ERK orp38 MAPK after LPS treatment. Such fundamental differences be-tween alveolar macrophages and ATII cells in MAPK expressionmay contribute to their differential response to LPS stimulation. Toour knowledge, there are no other published studies of the effectsof LPS and activation of MAPKs on cytokine and chemokine re-lease by primary human ATII cells. High basal levels of activatedERK1/2 have been shown in primary rat alveolar type II cells invitro (31), as well as the adenocarcinoma A549 cell line, which isoften used as a surrogate for alveolar type II cells (32–34). Incontrast to the current study, stimulation of IL-8 release by A549cells following cigarette smoke condensate, Burkholderia cepacia,
and crystalline silica resulted in increased activation of eitherERK, p38, or both MAPKs, respectively (31, 33, 34). This may bedue to differences in cell signaling processes between primaryATII cells and the A549 adenocarcinoma cell line; indeed, we havepreviously demonstrated morphological and biochemical differ-ences (19) between these cell types in vitro. However, the apparentlack of increased phosphorylation of ATII cell p38 MAPK andERK1/2 following LPS stimulation is surprising in the face of thestriking increases of cytokine, and particularly chemokine, release.It is possible that LPS triggers p38 MAPK and/or ERK1/2 activa-tion in primary ATII cells but at such a low level that it cannot bedetected by immunoblotting. Furthermore, other signaling path-ways seem likely to be involved, for example the JNK MAPKpathway, NF-�B, or other downstream pathways. Alternatively,there may be differences in the relative sensitivity of MyD88-sen-sitive and -insensitive pathways; the profile of cytokine release andrapid responses support the involvement of the MyD88-sensitivepathway, but it is unclear whether the MyD88-insensitive pathwayis also involved.
ERK1/2 is important in controlling the cell cycle and this mayinvolve sustained signaling (35–37). Some studies suggest thatERK1/2 and p38 MAPK signaling may have opposing actions tomodify or fine tune the effects of external stimuli/conditions (38,39). Basal, activated, phosphorylated ERK is high in human alve-olar macrophages cultured for up to 6 h after isolation from bron-choalveolar lavage and its constitutive activity is required for pro-longed survival of these cells in the face of adversity in situ (40).In the present study, we also detected constitutive activated ERKexpression by human lung macrophages, even after many hours invitro (over 28 h) although the levels were much lower than totallevels. We hypothesize that constitutively active ERK may alsocontribute to the basal release of low levels of TNF-�, IL-1�, andhigher levels of IL-8 by macrophages. In contrast, as noted earlier,there was no further phosphorylation of p38 MAPK or ERK1/2after LPS stimulation of ATII cells, suggesting either that theseMAPKs play a minor role in LPS-induced cytokine and chemokineproduction by these cells or that only small increases, not detect-able by immunoblotting, are involved. ATII cells are confluent atthe time of investigation (3–4 days; Ref. 19), forming a single cellmonolayer (15, 19), with all of the characteristics of ATII cells insitu. These cells differentiate into alveolar type-I-like epithelialcells in longer term culture (7–8 days following plating). Thus,high basal phosphorylated ERK1/2 may be related to ATII celldifferentiation in this model; low basal levels of phosphorylatedp38 MAPK described in the present study might also be required,as p38 MAPK can counteract the action of ERK (38, 39). Thesepossibilities have not been examined in primary human lung al-veolar epithelial cells.
There was a remarkable difference between cell types in re-sponse to IL-1� and TNF-� Ab blockade, where ATII cell releaseof MCP-1 and IL-6 was extremely sensitive to this treatment. It isimportant to note that, even in the presence of relatively high au-tocrine levels of TNF-� and IL-1�, macrophages produced rela-tively low levels of IL-6 and MCP-1, and use of super maximallevels of anti-TNF-� and anti-IL-1� Ab had little effect on MCP-1,with lesser effects on IL-6 production compared with ATII cellsover a 24 h period. One might argue that this is unimportant withsuch small levels of LPS-induced IL-6 and MCP-1 release. How-ever, the LPS-stimulated release of these mediators by macro-phages is highly significant, from virtually nothing to hundreds ofpicograms per milliliter of media; the fact that there was littlechange in these levels, particularly for MCP-1, even after 24 h ofincubation with blocking Abs to TNF-� and IL-1� suggests thatthere are no substantial autocrine effects of these mediators on
470 DIFFERENTIAL CYTOKINE RELEASE BY ATII CELLS AND MACROPHAGES
macrophages and that LPS induction of these mediators is largelyindependent of either TNFR- or IL-1R-signaling pathways. Alter-natively, macrophages may corelease soluble TNF-� receptor andIL-1R antagonist, in the same way as neutrophils (41), to auto-regulate cell signaling. The macrophage response was in directcontrast to the effects of the same treatment, under identical con-ditions, on ATII cell release of MCP-1 and IL-6, where MCP-1release was completely inhibited by anti-TNF-�, and partially in-hibited (�50%) by anti-IL-1�; IL-6 release was also significantlyinhibited (�60%) by anti-TNF-�. This inhibition occurred even inthe face of release of high levels of MCP-1 and IL-6 followingtreatment of ATII cells with LPS, although there were relativelylower levels of extracellular TNF-� and IL-1� in the media. Thesefindings imply that only low levels of TNF-� and/or IL-1� arerequired to induce synthesis of high levels of MCP-1 and moder-ately high levels of IL-6. Thus, it is probable that both the TNF-�receptor- and the IL-1�R-mediated signaling pathways are impor-tant mechanisms in the prolonged effects of LPS on ATII cellrelease of IL-6 and MCP-1.
Two recent articles (42, 43) using mouse embryonic fibroblastshighlight the significance of LPS-induced TNF-� secretion andpositive TNF-�-feedback mechanisms in the induction of achronic response (with respect to NF-�B expression), as comparedwith negative feedback mechanisms involved in the acute effectsof TNF-� stimulation when used in the absence of LPS, whichhelp explain differential responses to TNF-� and LPS. It is sug-gested that both the MyD88-dependent and -independent pathwaysof TLR4 activation contribute to TNF-� synthesis and secretionand up-regulation of NF-�B activity. Of particular relevance to thepresent study is the evidence of autocrine feedback of LPS-in-duced TNF-� release and subsequent activation of the TNF path-way via the TNFR, to elicit the chronic responses and prolongedchemokine and cytokine expression. In an earlier study of fetal ratalveolar epithelium, IL-1�R antagonism was demonstrated to ame-liorate LPS-induced NF-�B translocation and activation, illustrat-ing a similar positive feedback effect of LPS-induced IL-1� (44).Furthermore, the feedback control of human blood neutrophil LPS-induced IL-8 secretion has been shown to be autoregulated bycorelease of the soluble TNF-� receptor and IL-1R antagonist.When the activity of soluble TNF-� receptor and IL-1R antagonistwere blocked with neutralizing Abs, IL-1� was found to play asignificant role in LPS-induced feedback control of IL-8 produc-tion (41). Thus, these studies support the concept of positive feed-back control of LPS-induced mediator expression by coinductionof TNF-� and IL-1� release; however, it is clear that the exactcell-derived cytokine and signaling pathways vary according to thestimulus and nature of the cell source, e.g., myeloid or nonmyeloidin origin.
There is much debate in the literature as to the role of IL-1� andTNF-� in the initiation of the inflammatory response followingendotoxin exposure. Studies in rats showed that following LPSexposure there was a concentration-dependent increase in neutro-phil numbers in the lung which was partly abolished followingtreatment with neutralizing Abs to TNF-� (45). Abs to TNF-�have also proven effective, with reduced circulating IL-1, IL-6, andIL-8, and increased survival, in primate and rabbit models of en-dotoxemia (46, 47), suggesting a key role for TNF-� in the in-flammatory response. In wild-type and knockout (KO) TNFR I andII mice, i.p. LPS caused a rapid increase in pulmonary TNF-�,neutrophils, and chemokines in wild-type and KO TNFRII, but notKO TNFRI, mice, suggesting that LPS induction of TNF and pos-itive feedback via the TNFI receptor is an important mechanism inthe pulmonary response (48). Studies of mice with a knockout forthe p55 TNFR showed that these mice had increased resistance to
LPS due to the lack of TNF-� signaling in the absence of the p55TNFR (49). Similar results have also been found in animal modelsof septic shock using an IL-1R antagonist (50). Although thesestudies in animal models highlight the role of TNF-� and IL-1� inLPS-induced pulmonary inflammation, they do not identify whichcells are responsible.
The differences in macrophage and ATII cell signaling underidentical experimental conditions might be explained by differ-ences in absolute cell numbers. However, in this study, there is nodifference in the absolute numbers of cells per unit area of tissueculture well. Indeed, by 24 h, macrophages release over twice asmuch TNF-� and IL-1�, but at the same time release only about athird of the IL-8, MCP-1, and GRO�, compared with that of ATIIcells, clearly illustrating that this is not a cell number phenomenon.From the evidence in the present investigation, we hypothesizethat, in vivo, LPS-stimulated macrophage release of TNF-� andIL-1�, which in turn, or in concert with the stimulating factor,activates alveolar epithelial cells to release chemokines and triggerleukocyte migration. In healthy nonsmokers, the ratio of ATII cellsto macrophages is �5:1, illustrating the significance of LPS-stim-ulated and cytokine-stimulated chemokine release by ATII cells(51). Subsequent to augmented macrophage recruitment, furtherrelease of TNF-� and IL-1� and stimulation of ATII cells maythen amplify the inflammatory response due to increased ATII cellchemokine secretion. Of relevance to this hypothesis is the recentin vivo study by Jeyaseelan et al. (52). LPS-induced inflammationin mice resulted in increased CXCL5, a mouse chemokine homol-ogous to human epithelial neutrophil-activating peptide 78 andgranulocyte chemotactic peptide 2, being crucial in the inflamma-tory response; ATII cells were the major source of the CXCL5 andLPS induction of CXCL5 operated via the signaling pathways p38and JNK MAPKs. The present unique comparative study of rele-vant human alveolar macrophages and ATII cells from the samesubjects also strongly supports the concept that the resident ATIIcell is likely to be an important modulator of the inflammatoryresponse, not only via direct interaction with inflammatory medi-ators, but also via cross-talk with myeloid and other inflammatoryand resident cells.
The pattern of LPS-induced chemokine secretion by each celltype was vastly different and of particular note was IL-8 release.Macrophages released high levels of IL-8 basally which only in-creased by �50% following 1000 ng/ml LPS. In contrast, ATIIcells released low levels basally which increased 700% following1000 ng/ml LPS exposure for 24 h. This is a novel finding whichpartly reflects the differential responses to LPS discussed earlier.Previous studies using peripheral blood monocytes have shownthat IL-8 can work in an autocrine manner via the CXCR1 receptorto cause further release of IL-8 (53). This mechanism may bepresent in ATII cells, but is down-regulated in alveolar macro-phages. Because alveolar macrophages used in this study releasehigh basal levels of IL-8, it is likely that autoregulation of IL-8synthesis is imperative, possibly involving corelease of solubleTNF and IL-1 receptors and/or agonists (41) to prevent inappro-priate inflammation. In contrast to IL-8, GRO� was released inboth a time- and concentration-dependent manner by both celltypes and release did not plateau; nevertheless, ATII cells secretedsignificantly higher levels. GRO� exhibits broadly similar effectson leukocytes as IL-8 and binds the same receptors. This may beof clinical importance in diseases such as COPD because GRO� iselevated in COPD sputum and has been shown to elicit increasedmigration of monocytes from COPD subjects compared withhealthy nonsmokers (54).
Conditioned media from both cell types stimulated leukocytemigration and the magnitude of the response was related to the
profile and levels of chemokines. This study confirms previouswork (55) showing that human ATII cells are a major source ofMCP-1, which can be induced in vitro by IL-1� and TNF-� (55,56). mRNA studies of human peripheral lung have shown thatMCP-1 mRNA levels significantly correlate with the number ofintraepithelial macrophages observed (57). Our study adds to thesefindings and suggests that epithelial cell-derived MCP-1 is likelyto be a critical factor in monocyte recruitment to the lung (58). Onemight hypothesize that in the peripheral lung, where the alveolarepithelium is in close contact with the pulmonary microvascula-ture, epithelial-derived MCP-1 mediates effects within the capil-lary network to further increase recruitment of monocytes. Inter-estingly, the present study suggests that GRO� and IL-8 alsoinduce monocyte migration, but not to the same magnitude asMCP-1. Although classically thought of as neutrophil chemokines,other studies have shown the potential importance of GRO� andIL-8 in regulating monocyte migration, adhesion to (54), and mi-gration through (59) the microvascular cell wall.
In the leukocyte migration model used for these experiments,significantly more neutrophils than monocytes migrated toward theconditioned media dependent upon the concentration of GRO� orIL-8, which bind to the same cell receptors. Blockade of MCP-1,traditionally thought to be a monocyte-specific chemokine, causedless, but significant, inhibition of neutrophil migration. Although arecent rodent study showed that neutrophil sensitivity to, and mi-gration in response to, MCP-1 was increased by up to 100-foldunder inflammatory conditions (60), the neutrophils used in thepresent investigation were from healthy humans, suggesting thatMCP-1 is a chemoattractant for normal human neutrophils. How-ever, we cannot exclude the possibility that exposure of healthyneutrophils to conditioned medium from LPS-treated cells mayelicit an inflammatory response due to the presence of other factorsthat induce sensitivity to MCP-1.
In conclusion, this unique investigation of primary human alve-olar macrophages and ATII cells from the same subjects showsthat they are likely to act in concert to regulate the inflammatoryresponse, via differential release of primary cytokines by macro-phages and chemokines by ATII cells. Of particular relevance isthe demonstration that ATII cells are likely to play a central role inthe initiation of LPS-induced inflammation, via activation of TLR4and positive feedback control of chemokine release by TNF-� andIL-1�. In addition, we hypothesize that, in vivo, LPS-stimulatedrelease of IL-1� and TNF-� by macrophages exacerbates chemo-kine release from ATII cells. Thus, therapeutic strategies shouldtake account of the cellular targets as well as the mechanisms ofmediator action.
DisclosuresThe authors have no financial conflict of interest.
References1. Zalacain, R., V. Sobradillo, J. Amilibia, J. Barron, V. Achotegui, J. I. Pijoan, and
J. L. Llorente. 1999. Predisposing factors to bacterial colonization in chronicobstructive pulmonary disease. Eur. Respir. J. 13: 343–348.
2. Burns, J. L., R. L. Gibson, S. McNamara, D. Yim, J. Emerson, M. Rosenfeld,P. Hiatt, K. McCoy, R. Castile, A. L. Smith, and B. W. Ramsey. 2001. Longi-tudinal assessment of Pseudomonas aeruginosa in young children with cysticfibrosis. J. Infect. Dis. 183: 444–452.
3. Polsky, B., J. W. Gold, E. Whimbey, J. Dryjanski, A. E. Brown, G. Schiffman,and D. Armstrong. 1986. Bacterial pneumonia in patients with the acquired im-munodeficiency syndrome. Ann. Intern. Med. 104: 38–41.
4. Afessa, B., and B. Green. 2000. Bacterial pneumonia in hospitalized patients withHIV infection: the pulmonary complications, ICU support, and prognostic factorsof hospitalized patients with HIV (PIP) study. Chest 117: 1017–1022.
5. Boyton, R. J., and P. J. Openshaw. 2002. Pulmonary defences to acute respiratoryinfection. Br. Med. Bull. 61: 1–12.
6. Delclaux, C., and E. Azoulay. 2003. Inflammatory response to infectious pulmo-nary injury. Eur. Respir. J. 22: 10S–14S.
7. Kasama, T., Y. Miwa, T. Isozaki, T. Odai, M. Adachi, and S. L. Kunkel. 2005.Neutrophil-derived cytokines: potential therapeutic targets in inflammation. Curr.Drug Targets Inflamm. Allergy 4: 273–279.
8. Daniels, A. U., F. H. Barnes, S. J. Charlebois, and R. A. Smith. 2000. Macro-phage cytokine response to particles and lipopolysaccharide in vitro. J Biomed.Mater. Res. 49: 469–478.
9. Nathan, C. F., and J. B. Hibbs, Jr. 1991. Role of nitric oxide synthesis in mac-rophage antimicrobial activity. Curr. Opin. Immunol. 3: 65–70.
10. Zhang, P., W. R. Summer, G. J. Bagby, and S. Nelson. 2000. Innate immunityand pulmonary host defense. Immunol. Rev. 173: 39–51.
11. Wilmott, R. W., G. Khurana-Hershey, and J. M. Stark. 2000. Current concepts onpulmonary host defense mechanisms in children. Curr. Opin. Pediatr. 12:187–193.
12. Tetley, T. D. 2002. Macrophages and the pathogenesis of COPD. Chest. 121:156S–159S.
13. McRitchie, D. I., N. Isowa, J. D. Edelson, A. M. Xavier, L. Cai, H. Y. Man,Y. T. Wang, S. H. Keshavjee, A. S. Slutsky, and M. Liu. 2000. Production oftumour necrosis factor � by primary cultured rat alveolar epithelial cells. Cyto-kine 12: 644–654.
14. Koyama, S., E. Sato, H. Nomura, K. Kubo, M. Miura, T. Yamashita, S. Nagai,and T. Izumi. 2000. The potential of various lipopolysaccharides to release IL-8and G-CSF. Am. J. Physiol. 278: L658–L666.
15. Witherden, I. R., E. J. Vanden Bon, P. Goldstraw, C. Ratcliffe, U. Pastorino, andT. D. Tetley. 2004. Primary human alveolar type II epithelial cell chemokinerelease: effects of cigarette smoke and neutrophil elastase. Am. J. Respir. CellMol. Biol. 30: 500–509.
16. Thorley, A. J., P. Goldstraw, A. Young, and T. D. Tetley. 2005. Primary humanalveolar type II epithelial cell CCL20 (macrophage inflammatory protein-3�)-induced dendritic cell migration. Am. J. Respir. Cell Mol. Biol. 32: 262–267.
17. Armstrong, L., A. R. L. Medford, K. M. Uppington, J. Robertson, I. R.Witherden, T. D. Tetley, and A. B. Millar. 2004. Expression of functional Toll-like receptor-2 and -4 on alveolar epithelial cells. Am. J. Respir. Cell Mol. Biol.31: 241–245.
18. Payne, J. P., S. J. Kemp, A. Dewar, P. Goldstraw, M. Kendall, L. C. Chen, andT. D. Tetley. 2004. Effects of airborne World Trade Center dust on cytokinerelease by primary human lung cells in vitro. J. Occup. Environ. Med. 46:420–427.
19. Witherden, I. R., and T. D. Tetley. 2001. Isolation and culture of human alveolartype II pneumocytes. In Human Airway Inflammation; Sampling Techniques andAnalytical Protocols. D. F. Rogers and L. E. Donnelly, eds. Humana Press, To-towa, pp. 137–146.
20. Droemann, D., T. Goldmann, T. Tiedje, P. Zabel, K. Dalhoff, and B. Schaaf.2005. Toll-like receptor 2 expression is decreased on alveolar macrophages incigarette smokers and COPD patients. Respir. Res. 6: 68–75.
21. Carter, A. B., M. M. Monick, and G. W. Hunninghake. 1999. Both Erk and p38kinases are necessary for cytokine gene transcription. Am. J. Respir. Cell Mol.Biol. 20: 751–758.
22. Rutault, K., C. A. Hazzalin, and L. C. Mahadevan. 2001. Combinations of ERKand p38 MAPK inhibitors ablate tumor necrosis factor-� (TNF-�) mRNA in-duction: evidence for selective destabilization of TNF-� transcripts. J. Biol.Chem. 276: 6666–6674.
23. Birrell, M. A., S. Wong, K. McCluskie, M. C. Catley, E. L. Hardaker,S. Haj-Yahia, and M. G. Belvisi. 2006. Second-generation inhibitors demonstratethe involvement of p38 mitogen-activated protein kinase in post-transcriptionalmodulation of inflammatory mediator production in human and rodent airways.J. Pharmacol. Exp. Ther. 316: 1318–1327.
24. Monick, M. M., and G. W. Hunninghake. 2003. Second messenger pathways inpulmonary host defense. Annu. Rev. Physiol. 65: 643–667.
25. Chakrabarty, K., W. Wu, J. L. Booth, E. S. Duggan, K. M. Coggeshall, andJ. P. Metcalf. 2006. Bacillus anthracis spores stimulate cytokine and chemokineinnate immune responses in human alveolar macrophages through multiple mi-togen-activated protein kinase pathways. Infect. Immun. 74: 4430–4438.
26. Koch, A., M. Giembycz, R. G. Stirling, S. Lim, I. Adcock, K. Wassermann,E. Erdmann, and K. F. Chung. 2004. Effect of smoking on MAP kinase-inducedmodulation of IL-8 in human alveolar macrophages. Eur. Respir. J. 23: 805–812.
27. Bhattacharyya, A., S. Pathak, S. Datta, S. Chattopadhyay, J. Basu, and M. Kundu.2002. Mitogen-activated protein kinases and nuclear factor-�B regulate Helico-bacter pylori-mediated interleukin-8 release from macrophages. Biochem. J. 368:121–129.
28. Scholz, H., A. Yndestad, J. K. Damas, T. Waehre, S. Tonstad, P. Aukrust, andB. Halvorsen. 2003. 8-Isoprostane increases expression of interleukin-8 in humanmacrophages through activation of mitogen-activated protein kinases. Cardio-vasc. Res. 59: 945–954.
29. Granata, F., A. Frattini, S. Loffredo, P. A. Del, S. Sozzani, G. Marone, andM. Triggiani. 2006. Signaling events involved in cytokine and chemokine pro-duction induced by secretory phospholipase A2 in human lung macrophages. Eur.J. Immunol. 36: 1938–1950.
30. Hall, D. J., M. E. Bates, L. Guar, M. Cronan, N. Korpi, and P. J. Bertics. 2005.The role of p38 MAPK in rhinovirus-induced monocyte chemoattractant pro-tein-1 production by monocytic-lineage cells. J. Immunol. 174: 8056–8063.
31. Ovrevik, J., M. Lag, P. Schwarze, and M. Refsnes. 2004. p38 and Src-ERK1/2pathways regulate crystalline silica-induced chemokine release in pulmonary ep-ithelial cells. Toxicol. Sci. 81: 480–490.
32. Chen, Y. L., S. Z. Lin, W. L. Chang, Y. L. Cheng, and H. J. Harn. 2005. Re-quirement for ERK activation in acetone extract identified from Bupleurum scor-zonerifolium induced A549 tumor cell apoptosis and keratin 8 phosphorylation.Life Sci. 76: 2409–2420.
472 DIFFERENTIAL CYTOKINE RELEASE BY ATII CELLS AND MACROPHAGES
33. Hellermann, G. R., S. B. Nagy, X. Kong, R. F. Lockey, and S. S. Mohapatra.2002. Mechanism of cigarette smoke condensate-induced acute inflammatory re-sponse in human bronchial epithelial cells. Respir. Res. 3: 22–29.
34. Reddi, K., S. Phagoo, K. Anderson, and D. Warburton. 2003. Burkholderia ce-pacia-induced IL-8 gene expression in an alveolar epithelial cell line: signalingthrough CD14 and mitogen-activated protein kinase. Pediatr. Res. 54: 297–305.
35. Johnson, G. L., and R. Lapadat. 2002. Mitogen-activated protein kinase pathwaysmediated by ERK, JNK, and p38 protein kinases. Science 298: 1911–1912.
36. Murphy, L. O., and J. Blenis. 2006. MAPK signal specificity: the right place atthe right time. Trends Biochem. Sci. 31: 268–275.
37. Sumbayev, V. V., and I. M. Yasinska. 2006. Role of MAP kinase-dependentapoptotic pathway in innate immune responses and viral infection. Scand. J. Im-munol. 63: 391–400.
38. Kim, S. J., J. W. Ju, C. D. Oh, Y. M. Yoon, W. K. Song, J. H. Kim, Y. J. Yoo,O. S. Bang, S. S. Kang, and J. S. Chun. 2002. ERK-1/2 and p38 kinase oppositelyregulate nitric oxide-induced apoptosis of chondrocytes in association with p53,caspase-3, and differentiation status. J. Biol. Chem. 277: 1332–1339.
39. Shan, X., A. Hu, H. Veler, S. Fatma, J. S. Grunstein, S. Chuang, andM. M. Grunstein. 2006. Regulation of Toll-like receptor 4-induced proasthmaticchanges in airway smooth muscle function by opposing actions of ERK1/2 andp38 MAPK signaling. Am. J. Physiol. 291: L324–L333.
40. Monick, M. M., R. K. Mallampalli, M. Bradford, D. McCoy, T. J. Gross,D. M. Flaherty, L. S. Powers, K. Cameron, S. Kelly, A. H. Merrill, Jr., andG. W. Hunninghake. 2004. Cooperative prosurvival activity by ERK and Akt inhuman alveolar macrophages is dependent on high levels of acid ceramidaseactivity. J. Immunol. 173: 123–135.
41. Hattar, K., L. Fink, K. Fietzner, B. Himmel, F. Grimminger, W. Seeger, andU. Sibelius. 2001. Cell density regulates neutrophil IL-8 synthesis: role of IL-1receptor antagonist and soluble TNF receptors. J. Immunol. 166: 6287–6293.
42. Covert, M. W., T. H. Leung, J. E. Gaston, and D. Baltimore. 2005. Achievingstability of lipopolysaccharide-induced NF-�B activation. Science 309:1854–1857.
43. Werner, S. L., D. Barken, and A. Hoffmann. 2005. Stimulus specificity of geneexpression programs determined by temporal control of IKK activity. Science309: 1857–1861.
44. Haddad, J. J., R. Lauterbach, N. E. Saade, B. Safieh-Garabedian, and S. C. Land.2001. �-Melanocyte-related tripeptide, Lys-d-Pro-Val, ameliorates endotoxin-in-duced nuclear factor �B translocation and activation: evidence for involvement ofan interleukin-1� 193–195 receptor antagonism in the alveolar epithelium. Bio-chem. J. 355: 29–38.
45. Kips, J. C., J. Tavernier, and R. A. Pauwels. 1992. Tumor necrosis factor causesbronchial hyperresponsiveness in rats. Am. Rev. Respir. Dis. 145: 332–336.
46. Fiedler, V. B., I. Loof, E. Sander, V. Voehringer, C. Galanos, and M. A. Fournel.1992. Monoclonal antibody to tumor necrosis factor-� prevents lethal endotoxinsepsis in adult rhesus monkeys. J. Lab. Clin. Med. 120: 574–588.
47. Mathison, J. C., E. Wolfson, and R. J. Ulevitch. 1988. Participation of tumornecrosis factor in the mediation of gram negative bacterial lipopolysaccharide-induced injury in rabbits. J. Clin. Invest. 81: 1925–1937.
48. Calkins, C. M., J. K. Heimbach, D. D. Bensard, Y. Song, C. D. Raeburn,X. Meng, and R. C. McIntyre, Jr. 2001. TNF receptor I mediates chemokineproduction and neutrophil accumulation in the lung following systemic lipopoly-saccharide. J. Surg. Res 101: 232–237.
49. Pfeffer, K., T. Matsuyama, T. M. Kundig, A. Wakeham, K. Kishihara,A. Shahinian, K. Wiegmann, P. S. Ohashi, M. Kronke, and T. W. Mak. 1993.Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to en-dotoxic shock, yet succumb to L. monocytogenes infection. Cell 73: 457–467.
50. Ohlsson, K., P. Bjork, M. Bergenfeldt, R. Hageman, and R. C. Thompson. 1990.Interleukin-1 receptor antagonist reduces mortality from endotoxin shock. Nature348: 550–552.
51. Stone, K. C., R. R. Mercer, P. Gehr, B. Stockstill, and J. D. Crapo. 1992. Allo-metric relationships of cell numbers and size in the mammalian lung.Am. J. Respir. Cell Mol. Biol. 6: 235–243.
52. Jeyaseelan, S., R. Manzer, S. K. Young, M. Yamamoto, S. Akira, R. J. Mason,and G. S. Worthen. 2005. Toll-IL-1 receptor domain-containing adaptor proteinis critical for early lung immune responses against Escherichia coli lipopolysac-charide and viable Escherichia coli. J. Immunol. 175: 7484–7495.
53. Browning, D. D., W. C. Diehl, M. H. Hsu, I. U. Schraufstatter, and R. D. Ye.2000. Autocrine regulation of interleukin-8 production in human monocytes.Am. J. Physiol. 279: L1129–L1136.
54. Traves, S. L., S. J. Smith, P. J. Barnes, and L. E. Donnelly. 2004. Specific CXCbut not CC chemokines cause elevated monocyte migration in COPD: a role forCXCR2. J. Leukocyte Biol. 76: 441–450.
55. Eghtesad, M., H. E. Jackson, and A. C. Cunningham. 2001. Primary humanalveolar epithelial cells can elicit the transendothelial migration of CD14� mono-cytes and CD3� lymphocytes. Immunology 102: 157–164.
56. Rosseau, S., J. Selhorst, K. Wiechmann, K. Leissner, U. Maus, K. Mayer,F. Grimminger, W. Seeger, and J. Lohmeyer. 2000. Monocyte migration throughthe alveolar epithelial barrier: adhesion molecule mechanisms and impact of che-mokines. J. Immunol. 164: 427–435.
57. de Boer, W. I., J. K. Sont, A. van Schadewijk, J. Stolk, J. H. van Krieken, andP. S. Hiemstra. 2000. Monocyte chemoattractant protein 1, interleukin 8, andchronic airways inflammation in COPD. J. Pathol. 190: 619–626.
58. Gerszten, R. E., E. A. Garcia-Zepeda, Y. C. Lim, M. Yoshida, H. A. Ding,M. A. Gimbrone, Jr., A. D. Luster, F. W. Luscinskas, and A. Rosenzweig. 1999.MCP-1 and IL-8 trigger firm adhesion of monocytes to vascular endotheliumunder flow conditions. Nature 398: 718–723.
59. Smith, D. F., E. Galkina, K. Ley, and Y. Huo. 2005. GRO family chemokines arespecialized for monocyte arrest from flow. Am. J. Physiol. 289: H1976–H1984.
60. Johnston, B., A. R. Burns, M. Suematsu, T. B. Issekutz, R. C. Woodman, andP. Kubes. 1999. Chronic inflammation upregulates chemokine receptors and in-duces neutrophil migration to monocyte chemoattractant protein-1. J. Clin. In-vest. 103: 1269–1276.
