Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Eosinophils are effector cells involved in the inflammatory responses of allergy and asthma (1). Eosinophils are found in increased numbers in the bronchoalveolar lavage fluid (BALF) of patients with asthma (1) and are recruited to the airway after antigen challenge (2). Persistence of eosinophils within tissue lesions or the airway space is thought to contribute to the capacity of eosinophils to cause airway inflammation in asthma (3), and this process is regulated by many factors. There has been considerable interest that gram-negative endotoxin (lipopolysaccharide [LPS]) may regulate the function of eosinophils in asthma. For example, eosinophils isolated from human peripheral blood survive longer in the presence of LPS in vitro (4). Eosinophil recruitment to nasal fluids increases after challenge with low doses of LPS (5). Recruitment of eosinophils into the pleural cavity of mice is elicited by LPS in an interaction requiring the presence of lymphocytes and macrophages (6, 7). Intradermal injection of high doses of LPS (1 µg/ site) increases eosinophil accumulation in guinea-pig skin (8). Lower doses of LPS (30 ng/site) primes skin lesions in guinea pigs for recruitment of eosinophils in response to chemoattractants (9). However, little is known of the mechanisms by which LPS affects eosinophil functions such as migration and survival.

LPS is a component of the outer membrane of gram-negative bacteria consisting of both lipid and carbohydrate moieties (10). LPS elicits a broad inflammatory response, stimulating the production of several inflammatory mediators from leukocytes (11), such as granulocyte macrophage colony-stimulating factor (GM-CSF) production from monocytes (12). LPS can bind to serum LPS-binding protein (LBP) and soluble or cell-associated CD14, which together can form a high-affinity complex that mediates some of the interactions of LPS with cells (11). Currently, the roles of other LPS receptors, such as members of the Toll family and P₂ nucleotide receptors in LPS signaling, are being elucidated (13). CD14 is expressed on peripheral blood neutrophils and monocytes as a phosphatidylinositol-linked protein, and is found in normal human plasma as a soluble form at 6 µg/ml (11). LBP is found in normal serum at 0.5 to 10 µg/ml (11). Both CD14 and LBP are increased in the BALF of asthmatics 24 h after segmental challenge with ragweed antigen (16, 17).

To define the function of effector molecules in allergies or asthma, eosinophils isolated from peripheral blood via CD16-negative (CD16) selection (to remove peripheral blood neutrophils) have been used (18). Viability of eosinophils isolated in this manner is increased in vitro with the cytokines interleukin (IL)-3, IL-5, or GM-CSF (3), and also with LPS (4). Given the importance of LPS in affecting bronchial responsiveness, the following study examined the effect of LPS on survival of human eosinophils isolated via two different methods: the standard CD16 selection method, and the more stringent criteria of the CD14/CD16-negative (CD14/CD16) selection method. Whereas eosinophils isolated via CD16 selection lived longer in response to LPS, the response was lost when eosinophils were isolated via CD14/CD16 selection. These results indicate that the survival-enhancing effect of LPS requires CD14⁺ cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Low-endotoxin sterile water was purchased from Baxter Healthcare Corp. (Deerfield, IL). Hanks' balanced salt solution (HBSS) containing calcium and magnesium; RPMI 1640 medium; L-glutamine; penicillin; streptomycin; and low-endotoxin, heat-inactivated fetal calf serum (FCS) were purchased from Life Technologies (Grand Island, NY). Anti-CD16- and anti-CD14-conjugated magnetic beads were obtained from Miltenyi Biotech Inc. (Sunnyvale, CA). LPS (Escherichia coli serotype 055:B5), Percoll, fluorescein diacetate, propidium iodide (PI), and polymyxin B were purchased from Sigma Chemical Co. (St. Louis, MO). Lipid X was kindly obtained from Dr. Peter Stuetz, Sandoz Pharmaceutical Co. (Vienna, Austria). This preparation was synthetic and was shown to be free of contaminating by-products (19). Recombinant human IL-5 and GM-CSF were purchased from R&D Systems (Minneapolis, MN). Care was taken not to introduce endotoxin contamination into any reagents.

Monoclonal Antibodies

Hybridoma cell lines producing monoclonal antibody (mAb) to CD14 (60bd, immunoglobulin [Ig] G1) (20) and to the 2 integrin (TS1/18, IgG1) (21) were purchased from ATCC (Manassas, VA). mAbs were purified from serum-free (HyQ CCM1; Hyclone USA, Logan, UT) tissue culture supernatant by passage over a column of protein G complexed to agarose (Life Technologies), eluted with 50 mM glycine-HCl, pH 2.5, and immediately buffer-exchanged to 20 mM phosphate and 10 mM ethylenediaminetetraacetic acid, pH 7. Purified mAbs were concentrated using Centriplus spin concentration filters of 100 kD molecular weight cutoff (Amicon, Beverely, MA). mAbs to the cytokines IL-3 (IgG1), IL-5 (IgG1), and GM-CSF (IgG1) were purchased from R&D Systems. Simultest LeukoGATE (phycoerythrin [PE]-labeled anti-CD14 + fluorescein isothiocyanate [FITC]-labeled anti-CD45) was purchased from Becton Dickinson (San Jose, CA). Control, nonspecific mouse IgG1 used in survival assays was purchased from R&D Systems, and nonspecific mouse IgG used as control mAb for flow cytometry was purchased from Sigma. FITC-labeled goat antimouse IgG was obtained from Jackson ImmunoResearch Laboratories (West Grove, PA).

Eosinophil Isolation

Eosinophils were isolated from the peripheral blood of normal donors and patients with allergic rhinitis or mild bronchial asthma using the magnetic bead cell separation system (MACS; Miltenyi Biotech), as described by Hansel and colleagues (18). In brief, heparinized venous blood was centrifuged over a Percoll gradient (1.090 g/ml). Red blood cells were removed from the granulocyte pellet by hypotonic lysis, and the granulocytes were washed twice with HBSS without Ca²⁺ supplemented with 2% FCS. For CD16-negatively selected eosinophils, granulocytes were incubated with 100 µl anti-CD16 mAb-conjugated magnetic beads for 40 min at 4°C, and eosinophils were negatively selected by passing the CD16-labeled granulocytes over a MACS column in a magnetic field. For CD14/CD16 selection, 50 µl anti-CD14 mAb-conjugated magnetic beads were added to the anti-CD16 beads. To compare the two methods in an experiment, the granulocytes were split into two tubes before addition of anti-CD16 or the mixture of anti-CD16 + anti-CD14 mAb-coated magnetic beads. Purified eosinophils were rinsed in HBSS + 2% FCS, and yield and purity were determined before final centrifugation to bring the eosinophils to the proper concentration and medium used in experiments. The purity of eosinophils (Diff-Quik stains; Baxter Healthcare Corp., McGaw Park, IL) was greater than 99% in most instances, and preparations less than 95% were discarded. Viability was greater than 99% as determined by trypan blue dye exclusion.

In some experiments, monocytes were isolated from normal donors as described previously (22) using One-Step Monocytes (Accurate Chemical Co., Westbury, NY) gradient medium. Monocytes were incubated at 1 × 10⁶ cells/ml in RPMI 1640 containing 1% FCS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin with the indicated additions for 24 h at 37°C. Conditioned medium was collected by removing the cells with two sequential centrifugations and was kept at 20°C for further studies.

For experiments involving eosinophil coincubation with mononuclear cells, the mononuclear band from the Percoll gradient was collected and the platelets were removed by several low-speed centrifugations in Ca²⁺-free HBSS supplemented with 2% FCS. Mononuclear cells were kept at 4°C in RPMI 1640 containing 1% FCS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin until being warmed briefly for incubation with eosinophils.

Eosinophil Culture

Ninety-six-well, flat-bottomed tissue culture treated plates (Corning, Corning, NY) were blocked with 100 µl neat FCS for 2 h at 37°C. Wells were washed three times with HBSS, then purified eosinophils (2 × 10⁵/0.1 ml) in RPMI 1640 containing 1% FCS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin were added to each well. In some experiments eosinophils were incubated for 30 min at 40°C with mAbs or inhibitors as noted. The inhibitors or mAbs were not removed during the assay. Eosinophils were warmed briefly to 37°C before the addition of LPS. Eosinophils were incubated at 37°C in a humidified, 5% CO₂ incubator for 3 d, at which point viability was determined.

Determination of Eosinophil Viability

Determination of eosinophil viability was performed as described previously (23). In brief, after 72 h of incubation eosinophil viability was measured by staining cells with fluorescein diacetate (FDA) (5 µg/ml) and PI (1 µg/ml). FDA is cleaved by esterases in viable, metabolically active cells to yield fluorescein, thus viable cells are stained green. PI is permeable to cells that have compromised membrane integrity, thus dead cells are stained red. Eosinophils were brought to the bottom of the wells by centrifugation (400 × g × 5 min at 40°C) and were viewed using an inverted microscope with a dual-wavelength filter (Diaphot-TMD, Filter:B-2A; Nikon, Tokyo, Japan). At least 200 cells in three to four wells were counted for each condition tested, and the results are expressed as the mean percentage of viable cells.

Analysis of Eosinophils by Flow Cytometry

Eosinophils (1 × 10⁶) were incubated with the indicated mAbs for 30 min at 4°C in RPMI 1640 containing 1% FCS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin. Unbound mAb and media were rinsed away using phosphate-buffered saline (PBS) containing 1% bovine serum albumin (BSA) and centrifugation. Eosinophils were resuspended in PBS + 1% BSA and incubated with FITC-labeled antimouse secondary antibodies for 40 min on ice in the dark. Cells were rinsed twice, and PI was added immediately before analysis. Alternatively, eosinophils were incubated in the previously described medium with or without LPS for the indicated times, fixed in 1% paraformaldehyde for 30 min at 4°C, and stained with Simultest LeukoGATE, a PE-labeled anti-CD14 + FITC-labeled anti-CD45 mixture, for 30 min at 4°C. Samples were analyzed on a FACStar Plus (Becton Dickinson).

Analysis of GM-CSF Messenger RNA

CD16 eosinophils (3 × 10⁶ cells/sample) in 1 ml RPMI 1640 containing 1% FCS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin were either lysed immediately or incubated at 37°C for up to 24 h in the presence or absence of LPS. Quantitative reverse transcriptase/polymerase chain reaction and Southern blotting for GM-CSF messenger RNA (mRNA) were performed as described previously (23).

Statistics

The results are given as means ± standard error of the mean (SEM) except where otherwise noted. Statistical significance of the differences between various treatments was assessed using Minitab statistical software (Minitab, State College, PA) that performed a one-way analysis of variance with Fisher subcommands for multiple comparisons. A P value of less than 0.05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Viability of Eosinophils Incubated with LPS

Eosinophils were isolated from peripheral blood by centrifugation over a Percoll gradient to yield the granulocyte fraction, followed by CD16 selection to remove neutrophils. LPS increased the viability of the CD16 eosinophils in a dose-dependent manner, with a threshold at concentrations greater than 1 ng/ml (Figure 1). To insure that the effect of LPS was specific, two different inhibitors were tested: polymyxin B, which binds to the lipid A portion of LPS (24, 25); and lipid X, which is a precursor of LPS that acts as a competitive inhibitor (26). Polymyxin B at 50 U/ml was effective at blocking all doses of LPS tested (Figure 1). Lipid X also inhibited survival of CD16 eosinophils induced by 10 ng/ml LPS to levels that were not significantly different from samples incubated without LPS (LPS alone: 39 ± 8% viability; LPS + 50 U/ml polymyxin B: 6 ± 3% viability; LPS + 1 µg/ml lipid X: 6 ± 4% viability; no LPS: 9 ± 4% viability; means ± SEM of three experiments).

View larger version (18K): [in this window] [in a new window]   Figure 1.   Effect of LPS on the survival of CD16 eosinophils. Eosinophils isolated via CD16 selection were incubated with increasing amounts of LPS for 3 d in medium containing either no polymyxin B (squares) or 50 U/ml polymyxin B (circles) as described in MATERIALS AND METHODS. Data represent means ± SEM of three experiments. *Samples that were significantly different (P < 0.05) from 0 ng/ml LPS.

Role of Cytokines in LPS Eosinophil Survival

Survival of CD16 eosinophils mediated by LPS is blocked by mAbs to GM-CSF, and elevated levels of GM-CSF protein are found in supernatants of CD16 eosinophils incubated with LPS (4). The mRNA for GM-CSF was increased in our CD16 eosinophil preparations after 3 h of incubation in the presence of LPS (data not shown). No detectable levels of GM-CSF message were seen in unincubated CD16 eosinophils or CD16 eosinophils incubated without LPS, and no difference was observed in the levels of expression of glyceraldehyde-3-phosphate dehydrogenase message in similarly treated CD16 eosinophils (data not shown). Functional blocking mAbs were added before incubation with LPS to determine the roles of IL-3, IL-5, or GM-CSF in LPS-induced survival. The concentrations of mAbs used in these experiments completely blocked survival of eosinophils incubated with the appropriate purified, recombinant cytokine (data not shown). Survival of CD16 eosinophils incubated with anti-IL-3, anti-IL-5, or anti-IL-3 + anti-IL-5 was not different from eosinophils incubated without mAb (LPS alone) or with control mAb (Figure 2). Incubation with anti-GM-CSF attenuated LPS-induced survival by 63% (Figure 2). Addition of anti-IL-3 or anti-IL-5 with anti-GM-CSF did not significantly further decrease viability induced by LPS. However, addition of both anti-IL-3 and anti-IL-5 with anti-GM-CSF, to block all three cytokines, further attenuated eosinophil viability, inhibiting survival induced by LPS by 81% (Figure 2).

View larger version (31K): [in this window] [in a new window]   Figure 2.   Role of cytokines on eosinophil survival mediated by LPS. CD16 eosinophils were treated for 30 min at 4°C with 10 µg/ml mAbs against IL-3, IL-5, GM-CSF, or an isotype-matched control mAb before incubation with medium containing no addition (No LPS) or medium containing 10 ng/ml LPS. Data are expressed as means ± SEM of four experiments. *Samples containing anti-GM-CSF were significantly different (P < 0.05) from samples containing no mAb; samples containing anti-IL-3 + anti- IL-5 + anti-GM-CSF were significantly different (P < 0.05) from samples containing anti-GM-CSF alone.

Receptor Involvement in LPS-Mediated Survival of CD16 Eosinophils

To determine the system or binding protein(s) responsible for LPS interactions with CD16 eosinophils, function-blocking mAbs were incubated with CD16 eosinophils before incubation with LPS. CD11b/CD18 has been proposed to be a receptor for LPS on eosinophils (4). Blocking CD18 with mAb TS1/18 did not inhibit LPS-induced eosinophil survival (Figure 3). A 4-fold increase in the concentration of the mAb used did not significantly change the level of survival of CD16 eosinophils in response to LPS (data not shown). CD14 can mediate LPS interactions with other leukocytes, in conjunction with LPS-binding proteins (29). Addition of mAb 60bd, which blocks the binding site for LPS on CD14 (20), significantly inhibited LPS-mediated survival of eosinophils (Figure 3). Increasing the concentration of 60bd 4-fold did not significantly decrease CD16 survival in response to LPS (up to 88% inhibition of LPS-mediated survival). When the mAbs against CD18 and CD14 were combined, no further attenuation of survival beyond using 60bd alone was seen (up to 90% inhibition of LPS-mediated survival).

View larger version (26K): [in this window] [in a new window]   Figure 3.   Receptor involvement in LPS-mediated survival. CD16 eosinophils were incubated with 5 µg/ml anti- CD14 mAb 60bd, 5 µg/ml anti-CD18 mAb TS1/18, 5 µg/ml anti-CD14 mAb 60bd + 5 µg/ml anti-CD18 mAb TS1/ 18, or without mAb for 30 min at 4°C before incubation with or without 10 ng/ml LPS for 3 d as described in MATERIALS AND METHODS. Data represent means ± SEM of three experiments. *P < 0.05 versus No mAb and not significantly different from No LPS; P < 0.05 versus No mAb and not significantly different from anti-CD14 alone or No LPS.

Determination of Expression of CD14 on CD16 Eosinophils

CD18 has been proposed to be the receptor for LPS on eosinophils, and CD14 expression has not been shown on eosinophils previously (11). It was surprising, therefore, that a mAb to CD14, but not to CD18, inhibited eosinophil survival. Flow cytometry of CD16 eosinophils with the anti-CD14 mAb 60bd revealed that the majority of the cells were similar to background (mouse IgG), and no obvious population of positively staining cells was detected (Figure 4A). Expression of CD14 on a monocyte-like cell line, Mono Mac 6, is increased by incubation with LPS (30). Exposure to LPS for at least 24 h was needed to increase CD16 eosinophil viability (current authors' unpublished observations). To determine whether CD14 expression was increased in eosinophils upon incubation with LPS, CD16 eosinophils were incubated overnight in the presence or absence of LPS. For these experiments, the Simultest LeukoGATE reagent containing a cocktail of anti-CD45 and anti-CD14 mAbs was used to insure that the lack of detectable staining with the anti-CD14 mAb 60bd was not due to a failure of the mAb. No appreciable difference in CD14 expression was observed after incubation with LPS (Table 1).

View larger version (21K): [in this window] [in a new window]   Figure 4.   Expression of CD14 on eosinophils. Eosinophils isolated using anti-CD16 beads (CD16) or anti-CD16 beads + anti-CD14 beads (CD14/CD16) were incubated with 5 µg/ ml of nonspecific mouse IgG, anti-CD14 mAb 60bd, or anti-CD18 mAb TS1/18 and examined by flow cytometry. Data are shown as histograms after flow cytometric analysis as described in MATERIALS AND METHODS. (A) Eosinophils isolated using anti-CD16 beads; (B) eosinophils isolated using anti-CD16 + anti-CD14 beads. The mean fluorescence intensity of samples are: IgG, 3.39; anti- CD14, 4.12; and anti-CD18, 29.2 in A; and IgG, 3.96; anti-CD14, 5.89; and anti-CD18, 41.0 in B. The flow cytometric analysis of eosinophils prepared in this manner was repeated using eosinophils from three other individuals, with similar results.

Effect of Additional Serum on the Survival-Enhancing Effect of LPS

LBP and soluble CD14 are both present in serum (11), which was present at 1% in the culture medium. To determine whether serum concentration had an effect on responsiveness of CD16 eosinophils to LPS, survival was tested in medium containing various amounts of serum. CD16 eosinophils exhibited a 3-fold increase in viability when incubated with LPS in the absence of serum (Figure 5). In 1% FCS-containing medium, survival was near maximal levels (5-fold increase in viability). Addition of serum up to 10% did not significantly increase the response to LPS above medium containing 1% serum (Figure 5).

View larger version (17K): [in this window] [in a new window]   Figure 5.   Effect of serum concentration on the survival of CD16 eosinophils incubated with or without LPS. Eosinophils isolated via CD16 selection were incubated with increasing amounts of serum in the medium for 3 d in the presence (squares) or absence (diamonds) of 10 ng/ml LPS as described in MATERIALS AND METHODS. Data represent means ± SEM of two experiments. Samples containing LPS were significantly greater (P < 0.05) than samples incubated without LPS. All concentrations containing 1 to 10% serum were not significantly different from each other.

Depletion of CD14⁺ Cells on the Survival of Eosinophils Incubated with LPS

Isolation of eosinophils by CD14/CD16 selection yielded populations that were similar to CD16-negatively selected eosinophils when enumerated by manual counting of cytospin preparations (Table 2). When eosinophil preparations were analyzed by flow cytometry for staining with the anti-CD14 mAb 60bd, the majority of eosinophils isolated by either criteria exhibited a mean fluorescence intensity (MFI) similar to that seen with control mouse IgG (Figure 4). As a positive control, both populations of eosinophils stained strongly with the anti-CD18 mAb TS1/18 (Figure 4). The mAb 60bd was confirmed in its ability to stain cells by flow cytometry in this system by staining mononuclear cells from the same donors (data not shown).

CD14/CD16 eosinophils tended to show lower viability when incubated for 3 d without additions to the 1% FCS-containing medium than did CD16 eosinophils, although the difference between the two populations was not significant (Figure 6). The response of CD14/CD16 eosinophils to LPS was considerably less than that of CD16 eosinophils, with values that were not significantly different from those of CD14/CD16 eosinophils incubated in the absence of LPS (Figure 6). However, survival of both CD16 and CD14/CD16 cells was greater when IL-5 was added. The response to LPS in the CD16 population was 87% of the response to IL-5, and the response to LPS in the CD14/CD16 population was only 23% of the response to IL-5.

View larger version (33K): [in this window] [in a new window]   Figure 6.   Effect of CD14+ cell depletion on the survival of eosinophils incubated with LPS. Eosinophils isolated using anti-CD16 beads (CD16) or anti-CD16 beads + anti-CD14 beads (CD14/CD16) were incubated with medium containing no LPS (No Addition), 10 ng/ml LPS, or 1 ng/ ml IL-5 for 3 d before determination of viability as described in MATERIALS AND METHODS. Data represent means ± SEM of three experiments. *P < 0.05 versus appropriate No Addition.

Effect of CD14⁺ Cells on CD14/CD16 Eosinophil Response to LPS

The differences in responses of the eosinophils isolated by these methods could be due to the loss of a trace population of CD14-expressing cells. Addition of mononuclear cells from the same individual (2 × 10⁴ cells, representing 1% of the total cell population) increased the response of CD14/CD16 eosinophils to LPS by 2.2 ± 0.7-fold. Monocytes express CD14 and secrete cytokines, such as GM-CSF, when stimulated with high concentrations (1 µg/ml) of LPS (12). To test whether monocytes could secrete survival-enhancing factors in response to only 10 ng/ml LPS, conditioned media were collected after 24 h incubation of monocytes in the presence or absence of 10 ng/ml LPS. After collection, the conditioned media were treated with polymyxin B to block residual effects of LPS and were incubated with CD14/CD16 eosinophils (Table 3). Conditioned medium from unstimulated monocytes increased survival of CD14/CD16 eosinophils by 24% (Table 3). Conditioned medium from monocytes stimulated with 10 ng/ml LPS further enhanced CD14/CD16 eosinophil survival an additional 44%. The levels of survival obtained with these conditioned media were similar to levels of survival seen in CD14/CD16 eosinophil samples incubated with GM-CSF. Conditioned medium from monocytes whose stimulation with LPS was blocked using polymyxin B was not as effective at stimulating CD14/CD16 eosinophil survival as medium from either unstimulated monocytes or LPS-stimulated monocytes (Table 3).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we were able to reproduce the results of Takanaski and associates (4) and show that eosinophils isolated via conventional CD16 selection strongly respond to LPS by an increase in survival. Experiments using neutralizing mAbs, as in previous work (4), indicated that the increased viability of the CD16 eosinophils was mediated in large part by GM-CSF. We extended these observations to explore the mechanism of LPS interaction with eosinophils and found that cells expressing CD14 are necessary for the increased survival of eosinophils incubated with LPS.

The action of LPS on eosinophil survival is a specific effect of LPS, because two types of inhibitors blocked the increase in survival of CD16 eosinophils by LPS. Polymyxin B is an antibiotic that binds to the lipid A portion of LPS, preventing LPS from binding to LPS protein complexes which mediate its effects (24, 25). Lipid X is a precursor form of LPS that appears to compete for binding with LPS, but because it does not contain the Lipid A moiety it cannot function as an endotoxin and thus serves as a selective antagonist for LPS (26). The primary cytokine rendering this interaction appears to be GM-CSF. Takanaski and coworkers (4) showed increased levels of GM-CSF protein in supernatants of CD16 eosinophils, and we found an increase in message for GM-CSF in CD16 eosinophils incubated with LPS (data not shown). As shown in Figure 2, the antibody cocktail blocking IL-3, IL-5, and GM-CSF decreased eosinophil survival slightly more than did the mAb to GM-CSF alone. Considering that even the antibody cocktail did not completely block CD16 eosinophil survival, it is possible that other factors may be involved to account for the residual (roughly 20%) LPS-mediated survival.

In contrast to what has been previously concluded (4), CD18 does not seem to be the principal mediator of the LPS response of CD16 eosinophils. The function-blocking mAb to CD18 (TS1/18) did not alter the survival of CD16 eosinophils in response to LPS. An insufficient quantity of mAb during the course of the assay could not account for this result because increasing the concentration 4-fold did not alter eosinophil viability. It is possible that mAb TS1/18 may not efficiently block the interaction between LPS and CD18. Therefore we cannot rule out a role for CD18 in the survival-enhancing effect of LPS on CD16 eosinophils. Recent studies (reviewed in Reference 29) suggest that LPS interaction with CD18 is not sufficient to mediate signals but instead may act by clustering LPS on the cell surface, allowing for interaction with other LPS-binding proteins such as CD14, Toll receptors, and/or nucleotide receptors that transduce the signal (13).

The enhancement of CD16 eosinophil survival in response to LPS was significantly blocked by mAbs to CD14, indicating a substantial role for CD14 in this interaction. However, no detectable CD14 was observed on either CD16 or CD14/CD16 eosinophils by flow cytometry. It is unlikely that the lack of detection on these cells was due to faulty mAb specificity because two different anti-CD14 mAbs were used in these studies (60bd and LeukoGATE), and because mononuclear cells stained positively. In addition, the expression of CD14 by CD16 eosinophils was not observed after exposure to LPS (Table 1), as can be the case in monocyte-like cell lines (30).

One possible source of CD14 could be the serum used in the incubation medium. The medium used to culture eosinophils in this study contained 1% FCS, which could contain between 60 and 100 ng/ml of LBP and soluble CD14 (11). In association with LBP, LPS binds CD14 either expressed on the surface of leukocytes or in solution and can activate cells that are CD14-deficient (11). CD16 eosinophils showed a substantially increased survival when incubated with LPS in medium that contained no serum, and thus no serum source of soluble CD14 protein or LBP (Figure 3). It seems unlikely, therefore, that soluble CD14 is necessary for the increase in survival induced by LPS. There was an increase in viability seen when additional serum was present, indicating that LPS-binding proteins that may be present in the serum could enhance the effect seen in these experiments. Considering that a similar increase was not seen in samples incubated without LPS, the effect of the additional serum content appeared to be specific for LPS and not a more general effect of increased growth factors or nutrients. Soluble CD14 and LBP are found in the airway after allergen challenge (16, 17), however there are no studies to date examining whether soluble CD14 and LBP can interact directly with eosinophils to mediate eosinophil responses. Although eosinophils are known to express nucleotide receptors that may participate in LPS signaling (31), it is not known whether eosinophils express members of the Toll family of proteins. Future studies are needed to address these possibilities.

Given that there was no detectable level of CD14 on freshly isolated eosinophils, it is doubtful that residual CD14 from the donor remained in association with the eosinophils after the purification. As described in MATERIALS AND METHODS, several stepsincluding several rinses were performed in the purification of eosinophils from peripheral blood. However, depletion of CD14-expressing cells by negative selection with anti-CD14-coated beads decreased the response of eosinophils (referred to as CD14/ CD16) to LPS (Figure 6). Eosinophils isolated by the CD14/CD16 criteria were viable and remained responsive to GM-CSF (Table 3) and IL-5 (Figure 6), indicating that the purification method did not alter the general responsiveness of these cells. The difference in response to LPS between cells obtained by the different isolation criteria suggests that the lack of response from the CD14/ CD16 eosinophils was due to a loss of a trace population of cells expressing CD14 on their cell surface. Survival of CD16 eosinophils mediated by LPS is decreased in the presence of IL-10 (4), a cytokine known to inhibit the production of GM-CSF, IL-1, tumor necrosis factor-, and granulocyte colony-stimulating factor from monocytes stimulated with LPS (12). Thus, it is possible that trace amounts of monocytes present in CD16 eosinophil populations contribute to the eosinophil survival response to LPS. Addition of mononuclear cells to constitute 1% of the total cell population increased the response of CD16/ CD14 eosinophils to LPS by roughly 2-fold. Conditioned medium from LPS-treated monocytes increased the survival of CD16/CD14 eosinophils, showing that monocytes stimulated with low doses of LPS have the capacity to secrete viability-enhancing factors. The factor(s) secreted by monocytes after incubation with LPS were not examined in this study. The possibility that other CD14⁺ cells may account for the increase in eosinophil survival cannot be excluded on the basis of these data. It is feasible that a small population of eosinophils express a level of CD14 too low to be detected by flow cytometry. Neutrophils were most likely not the causative cells in these studies, because in most instances the inclusion of CD14 selection did not decrease the percentage of neutrophils in the population (Table 2). Future studies using cell sorting of the CD16 eosinophil population for CD14⁺ cells are needed to confirm the identity of the LPS-responding cells.

The idea of an indirect effect of LPS on eosinophil function is not without precedence. Influx of eosinophils into the pleura of mice after treatment with LPS requires the presence of lymphocytes and resident macrophages, suggesting a key role for lymphocytes and/or macrophages in eosinophil recruitment into the pleural cavity (6, 7). Similarly, accumulation of eosinophils into nasal airway spaces (5) in response to LPS may also require cells that express CD14, given that LPS is not a chemoattractant for eosinophils (32).

The role of eosinophils in asthma and the exacerbation of asthma by endotoxin has long been appreciated. Soluble CD14 and LBP have been found in the airways of asthmatics after antigen challenge (16, 17), reinforcing the hypothesis that LPS responses occur in the lung. LPS-responsive cells are found in increased numbers in the airway after allergen challenge, and data herein suggest that activation of such cells could contribute to the viability of eosinophils in the airway. Thus, development of pharmacologic therapies to treat the exacerbation of asthma by endotoxin should take into consideration the interaction of LPS-responsive cells with eosinophils.