| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of Immunology and Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, MN 55905
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RI); it is speculated
that this receptor plays a role in eosinophil mediator release in
allergic diseases. However, questions still remain. How much of the
FcRI protein is actually expressed on the cell surface of the
eosinophil? If they are present, are these IgE receptors associated
with effector functions of eosinophils? To address these issues, we
studied blood eosinophils from patients with ragweed hay fever. A high
level of low affinity IgG receptor (Fc
RII, CD32), but no expression
of FcRI, was detectable on the eosinophil surface by standard FACS
analysis. However, after in vitro sensitization with biotinylated
chimeric IgE (cIgE), cell-bound cIgE was detected by PE-conjugated
streptavidin. This cIgE binding was partially inhibited by
anti-FcRI mAb, suggesting that eosinophils do express minimal
amounts of FcRI detectable only by a sensitive method. Indeed, FACS
analysis of whole blood showed that eosinophils express 
0.5% of the
FcRI that basophils express. When stimulated with human IgE or
anti-human IgE, these eosinophils did not exert effector functions;
there was neither production of leukotriene C4 or
superoxide anion nor any detectable degranulation response. In
contrast, eosinophils possessed membrane-bound human IgG and showed
functional responses when stimulated with human IgG or anti-human
IgG. Thus, IgG and/or cytokines, such as IL-5, appear to be more
important for eosinophil activation in allergic diseases than
IgE. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A critical feature of the allergic response is the release of histamine
and other inflammatory mediators from mast cells and basophils after
allergen cross-links specific IgE Ab occupying high affinity IgE
receptors (FcRI) (10). Furthermore, epidemiological studies show a
close correlation between serum IgE levels and the prevalence and
severity of the bronchial asthma (11). Thus, there is converging
evidence to support a role for IgE in the pathophysiology of allergic
diseases in humans. Earlier, eosinophils were considered to express
only low affinity IgE receptor (FcRII) (12); however, recent studies
by Gounni et al. (13) demonstrated that eosinophils express FcRI and
that they release inflammatory mediators and exert cytotoxicity to
helminths through this receptor. Subsequent immunohistochemical and
immunocytochemical studies on patients with allergic diseases showed
that local allergen provocation induces expression of FcRI by
eosinophils infiltrating into the airways (14) and skin (15, 16).
Furthermore, FcRI was detected in blood eosinophils from patients
with various allergic diseases (17). Therefore, it is reasonable to
speculate that FcRI plays a role in the eosinophil secretory process
in allergic diseases.
There are caveats to this conception. Most of the patient studies were
performed by histologic examination using in situ hybridization and
immunocytochemical techniques. Therefore, little is known about the
actual expression of FcRI protein on the eosinophil surface. For
example, in human T cells, the low affinity IgG receptor proteins
(FcRII, CD32) are stored within the cytoplasm but are not expressed
on the cell surface (18), and their surface expression was detected
only by a highly sensitive biotin-avidin flow cytometric technique
(19). Furthermore, the functional significance of the FcRI
expression by eosinophils from patients with allergic diseases is also
unknown. To address these questions, we studied blood eosinophils from
patients with ragweed hay fever during the peak of the allergy season.
We had three goals. First, we examined whether FcRI proteins are
expressed on the surface of the eosinophil, and we compared the
expression levels to known FcRI-bearing cells, such as basophils.
Second, we examined whether FcRI confers IgE binding to eosinophils.
Finally, we examined whether IgE receptors stimulate mediator release
by these eosinophils.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Thirty patients, 18–60 yr of age, with moderate to severe ragweed allergic rhinitis and 14 healthy nonatopic subjects were studied. The patients had a history of ragweed hay fever, a positive skin test to ragweed extract, and an elevated level of IgE Ab for short ragweed (305–2190% of normal, median 1114%). Patients were excluded if they were receiving any forms of glucocorticoid treatment within the preceding 6 mo or immunotherapy within the preceding 12 mo. Each patient kept a diary of daily symptom scores for nasal congestion, pruritus or discharge, ocular pruritus, eye watering, eye redness, and pruritis of the ear or palate to validate their clinical reaction to airborne ragweed allergen. Normal control subjects had no clinical or laboratory evidence of atopy. Blood samples were obtained from patients and control subjects during the peak of hay fever season in Rochester, MN (end of August) and used as described below. The study was approved by the Institutional Review Board of the Mayo Clinic and Mayo Foundation.
Antibodies
Anti-FcRI 
subunit mAb (15-1, mouse IgG1
(mIgG1))3 and biotinylated
chimeric mouse/human IgE monoclonal protein (cIgE) were kind gifts from
Dr. J.-P. Kinet, Beth Israel Deaconess Medical Center, Harvard Medical
School (Boston, MA). Concentration-response experiments to establish
saturating concentrations for cIgE to sensitize human eosinophils in
vitro and for 15-1 to inhibit the sensitization of cIgE were performed
before this study began (our unpublished results); the saturating
concentrations were 25 µg/ml for cIgE and 10 µg/ml for 15-1 per
1 x 106 eosinophils. Anti-CD11b mAb (Bear1, mIgG1),
anti-FcRII (CD23) mAb (9P25, mIgG1), anti-CD25 mAb (B1.49.9,
mIgG2a), anti-CD32 mAb (2E1, mIgG2a), and anti-CD69 mAb
(TP1.55.3, mIgG2b) were purchased from Immunotech (Westbrook, ME).
Control mouse myeloma Ig, including mIgG1, mIgG2a, and mIgG2b, were
obtained from ICN Pharmaceuticals (Costa Mesa, CA). PE-conjugated
F(ab')2 fragments of sheep IgG anti-mIgG and
PE-conjugated streptavidin were purchased from Sigma Chemical (St.
Louis, MO) and Becton Dickinson (Mountain View, CA), respectively.
FITC-conjugated goat IgG anti-human IgG, FITC-conjugated goat
anti-human IgE, and FITC-conjugated goat IgG anti-mouse IgM
were from BioSource International (Camarillo, CA).
Reagents for cell functional analyses
We used myeloma IgE purified from serum from a patient with multiple myeloma as described elsewhere (5). Purified human IgE was also purchased from Cortex Biochem (San Leandro, CA). We performed all the experiments with both of the IgE preparations and found that the results were virtually identical. Therefore, only results from our myeloma IgE preparation(s) are shown. Purified human serum IgG was purchased from ICN Pharmaceuticals. F(ab')2 fragments of goat IgG anti-human IgG and F(ab')2 fragments of goat IgG anti-human IgE were purchased from ICN Pharmaceuticals and DiaMed (Windham, ME), respectively. F(ab')2 fragment of goat IgG was obtained from Jackson ImmunoResearch Laboratories, West Chester, PA.
Cell preparation
Eosinophils were isolated from six patients and six normal subjects by a magnetic cell separation system (MACS, Becton Dickinson, San Jose, CA) as described previously with minor modifications (20). Briefly, 60 ml of venous blood anticoagulated with 50 U/ml heparin were diluted with PBS at a 1:1 ratio. Diluted blood was overlaid on isotonic Percoll solution (density, 1.085 g/ml, Sigma) and centrifuged at 1000 x g for 30 min at 4°C. The supernatant and mononuclear cells at the interface were carefully removed, and erythrocytes in sediment were lysed by two cycles of hypotonic water lysis. Isolated granulocytes were washed twice with PIPES buffer (25 mM PIPES, 50 mM NaCl, 5 mM KCl, 25 mM NaOH, 5.4 mM glucose, pH 7.4) containing 1% alpha calf serum (HyClone Laboratories; Logan, UT). An approximately equal volume of anti-CD16-conjugated magnetic beads (Miltenyi Biotec, Auburn, CA) was added to the cell pellet. After 60 min of incubation on ice, cells were loaded onto the separation column positioned in the strong magnetic field of the MACS. Cells were eluted three times with 5 ml of PIPES buffer with defined calf serum. The purity of eosinophils counted by Randolph’s stain was regularly >98%. Purified eosinophils were washed and suspended in reaction medium and then used immediately.
Basophils were obtained as the mononuclear cell layer after density gradient centrifugation from the same donor used for eosinophil isolation. Briefly, heparinized venous blood was centrifuged over Histopaque 1077 (Sigma), following the procedure recommended by the manufacturer. The mononuclear cell layer was collected, washed with PIPES buffer, suspended in reaction medium, and used immediately. The purity of basophils counted by Wright-Giemsa staining of cytospin preparations was between 0.5% and 2%.
Flow cytometric analysis
Eosinophil activation markers.
Expression of CD25, CD69, and CD11b on eosinophils was measured by flow
cytometry as an indicator of eosinophil activation (21, 22, 23). To avoid
purification-induced activation or inadvertent subset selection, total
leukocyte samples were stained with mAb and eosinophils were
electronically gated during FACS analysis. Briefly, 5 parts of
heparinized whole blood from patients and normal subjects were
incubated with 1 part of 6% hetastarch in 0.9% NaCl (Hespan, DuPont
Pharmaceuticals, Wilmington, DE) for 45 min at 37°C to sediment
erythrocytes. The buffy coat was collected, and the remaining
erythrocytes were lysed by erythrocyte lysing solution (Becton
Dickinson). Buffy coat leukocytes were washed with PBS containing 0.1%
NaN3 and 1% BSA (PAB buffer). The cells were resuspended
in PAB buffer at 2 x 107 cells/ml and added to
96-well plates at 100 µl/well. One microgram of anti-CD11b,
anti-CD25, anti-CD69, or isotype-matched control mouse Ig was
added separately to individual wells and incubated for 30 min at 4°C.
After washing, cells were resuspended with 100 µl of PE-conjugated
F(ab')2 fragments of sheep anti-mouse IgG diluted in
PAB buffer at 1:400 and incubated in the dark for 30 min. The samples
were washed and fixed with 1% paraformaldehyde in PAB. FACS analysis
was performed within 2 h using a FACScan flow cytometer (Becton
Dickinson). Established programs were used to ensure consistent
instrument settings (e.g., laser power, compensation, gain, scaling,
etc.) among experiments. Two-dimensional dot-plot analysis of the green
fluorescence intensity and side scatter revealed four distinct
populations of leukocytes, including lymphocytes, monocytes,
neutrophils, and eosinophils. Eosinophils were identified as a cell
population with the strongest green autofluorescence and side scatter
as previously described (24). Appropriateness of the identification of
the eosinophil population was confirmed by positive staining for
anti-CD9 mAb (Immunotech) as well as morphological analysis of
cytospin preparation of sorted cells. The cells were electronically
gated on this eosinophil population, and the expression of activation
markers was measured by examining the intensity of PE fluorescence.
Fluorescence intensity was determined on 50,000 total cells from each
sample using logarithmic amplification and presented as the differences
in the mean fluorescence channel (
MFC) between the test mAb (e.g.,
anti-CD11b) and isotype-matched control mouse Ig by Becton
Dickinson Lysis II software.
Receptor expression
Expression of receptors for IgE and IgG, including FcRI,
FcRII (CD23), and FcRII (CD32), by eosinophils was examined by a
standard FACS technique using isolated eosinophils. To release putative
in vivo bound IgE molecules from the surface of the eosinophil,
eosinophils were subjected to acid stripping as described previously
(25). Briefly, isolated eosinophils were washed with cold 0.15 M NaCl.
The pellets were treated with lactic acid buffer (130 mM NaCl, 5 mM
KCl, 10 mM lactic acid, pH 3.9) for 3.5 min. Cells were washed with PAB
buffer and suspended in the same buffer at 5 x 106
cells/ml. Aliquots (100 µl) of cell suspension were added to 15-ml
polypropylene conical tubes, mixed with saturating amounts of
anti-FcRI mAb (15-1), anti-FcRII mAb (9P25), or
anti-FcRII mAb (2E1), or the same amount of control mouse Ig.
After an incubation for 30 min on ice, the cell pellets were washed
twice with cold PAB buffer and mixed with 100 µl of 1:400 dilution of
PE-conjugated F(ab')2 fragments of sheep anti-mouse
IgG. After an additional incubation on ice for 30 min in the dark, the
cell pellets were washed twice with PAB buffer and fixed with 1%
paraformaldehyde. Fluorescence analysis was conducted in a FACScan flow
cytometer. Fluorescence intensity was determined on 10,000 cells from
each sample using logarithmic amplification and presented as the
differences in the mean fluorescence channel (MFC) between the test
mAb (e.g., 15-1) and isotype-matched control mouse Ig by Becton
Dickinson Lysis II software.
Ig binding
Binding of IgE or IgG to eosinophils and other leukocyte
populations was examined before and after in vitro passive
sensitization with these Ig. Isolated eosinophils or buffy coat
leukocytes were obtained as described above. One portion of cells was
stained immediately for the analysis of cell-bound IgE or IgG in vivo
(see below). The other portion of cells was passively sensitized in
vitro with IgE or IgG. First, cells were treated with lactic acid
buffer to remove in vivo bound IgE molecules as described above.
Eosinophils and other leukocytes were then washed with sensitization
buffer (RPMI 1640 medium supplemented with 25 mM HEPES, 1% BSA, 0.1%
NaN3, pH 7.3) and resuspended in the medium at 5 x
106 cells/ml and 2 x 107 cells/ml,
respectively. Aliquots (100 µl) of each cell suspension were added to
15-ml polypropylene conical tubes and mixed with human IgE (50
µg/ml), human IgG (50 µg/ml), or biotinylated cIgE (25 µg/ml). In
some experiments, to block the binding of IgE to FcRI, cells were
pretreated with a saturating concentration of anti-FcRI mAb
(15-1, 10 µg/ml) for 30 min on ice before in vitro sensitization with
cIgE. Cells were incubated for 2 h at 4°C with gentle mixing and
washed twice with PAB buffer. To detect IgE or IgG bound on the cell
surface, cells were incubated with saturating amounts of
FITC-conjugated goat anti-human IgE or FITC-conjugated goat
anti-human IgG, or the same amounts of control (FITC-conjugated
goat anti-mouse IgM). Alternatively, cell-bound biotinylated cIgE
was visualized by PE-conjugated streptavidin. After incubation on ice
for 30 min in the dark, cell pellets were washed twice with PAB buffer
and fixed with 1% paraformaldehyde. Fluorescence analysis was
conducted in a FACScan flow cytometer. To analyze IgE, cIgE, or IgG
bound to isolated eosinophils, fluorescence intensity was determined on
10,000 cells from each sample using logarithmic amplification and
presented as the MFC between the test Ab (e.g., anti-human IgE)
and control Ab by Becton Dickinson Lysis II software. To analyze
binding of cIgE to individual cell populations of buffy coat
leukocytes, cells were electronically gated based on their unique green
autofluorescence and side scatter characteristics. Two-dimensional
dot-plot analysis of the green fluorescence intensity and side scatter
of buffy coat leukocytes showed four distinct populations, including
lymphocytes, monocytes, neutrophils, and eosinophils (see
Results for detail). Basophils fell within the population of
lymphocytes. However, when basophils were stained with biotinylated
cIgE and PE-conjugated streptavidin, they formed one unique population
in green fluorescence/side scatter dot-plot analysis due to their
extremely strong PE fluorescence. Thus, individual cell populations
were easily identified by their unique green autofluorescence and side
scatter characteristics; appropriateness of the identification of each
population had been confirmed by positive staining for appropriate Ab
(CD9 for eosinophils and basophils, CD16 for neutrophils, and CD14 for
monocytes) as well as morphological analysis of cytospin preparations
of sorted cells. The cells were electronically gated, and the binding
of cIgE was measured by examining the intensity of PE fluorescence.
Fluorescence intensity was determined on 50,000 total cells from each
sample using logarithmic amplification, which was converted to the
linear equivalent by Becton Dickinson Lysis II software. Data are
presented as the MFI between test samples (e.g., biotinylated cIgE
plus PE-conjugated streptavidin) and control (PE-conjugated
streptavidin alone).
Leukotriene (LT) C4 production
Eosinophil production of LTC4 was performed in 96-well tissue culture plates as described previously with minor modifications (26). Isolated eosinophils were stimulated by cross-linking cell surface IgE or IgG receptors by human IgE, human IgG, anti-human IgE or anti-human IgG, immobilized onto the wells of tissue culture plates. Wells of 96-well flat-bottom Costar 3596 tissue culture plates (Costar, Cambridge, MA) were coated overnight at 4°C with 50 µl of serial dilutions of human IgE or human IgG, or 50 µg/ml of F(ab')2 fragments of goat anti-human IgG, goat anti-human IgE or goat IgG (control). After aspiration of the solution, 50 µl of 2.5% human serum albumin (HSA, Sigma) was added to each well to block the nonspecific protein binding sites. After incubation for 2 h at 37°C and 5% CO2, wells were washed twice with 0.9% NaCl before use. Freshly isolated eosinophils were washed with HBSS supplemented with 10 mM HEPES, 20 mM L-serine and 5 mM glutathione and resuspended in the same medium at 5 x 105 cells/ml. Aliquots of cell suspensions (200 µl) were added to the wells and incubated for 1 h at 37°C and 5% CO2. The supernatants from each well were collected and frozen at -70°C or assayed immediately. Concentrations of LTC4 in the sample supernatants were measured by ELISA using an LTC4 kit (Cayman Chemical, Ann Arbor, MI) following the procedure recommended by the manufacturer. The sensitivity of the assay was 4.8 pg/ml. All experiments were conducted in duplicate.
Superoxide anion production
Eosinophil superoxide production was induced by Ig or anti-Ig immobilized onto tissue culture plates and measured by reduction of cytochrome c as previously described (27, 28). Briefly, IgE, IgG, anti-IgE, or anti-IgG were immobilized onto polystyrene 96-well flat-bottom tissue culture plates as described above. Freshly isolated eosinophils were washed and resuspended in HBSS with 10 mM HEPES and 100 µM cytochrome c at 5 x 105 cells/ml. Cell suspension (100 µl) was dispensed onto the wells, followed by 100 µl of medium alone. Immediately after addition of stimuli, the reaction wells were measured for absorbance at 550 nm in a microplate autoreader (Thermomax, Molecular Devices, Menlo Park, CA), followed by repeated readings. Between absorbance measurements, the plate was incubated at 37°C. Superoxide anion generation was calculated with an extinction coefficient of 21.1 x 103cm-1 M-1 for reduced cytochrome c at 550 nm and was expressed as nanomols of superoxide produced per 105 cells.
Eosinophil and basophil degranulation
Eosinophils and basophils were stimulated with anti-human IgE or anti-human IgG, either immobilized onto the wells of tissue culture plates or in solution, or FMLP (Calbiochem-Novabiochem, San Diego, CA) in solution. Polystyrene 96-well flat-bottom tissue culture plates were coated with 50 µg/ml F(ab')2 fragments of goat anti-human IgE, goat anti-human IgG, or goat IgG (control); blocked with 2.5% HSA; and washed with saline as described above. Plates used for anti-IgE, anti-IgG, or FMLP in solution were blocked with HSA. Freshly isolated eosinophils or basophil preparations were suspended in RPMI 1640 supplemented with 10 mM HEPES and 0.1% HSA at 5 x 105 cells/ml and 1 x 106 cells/ml, respectively. Aliquots of cell suspensions (100 µl) were added to the wells and stimulated with serial dilutions of F(ab')2 fragments of goat anti-human IgE, goat anti-human IgG or goat IgG (control), or 1 µM FMLP (100 µl). Cells incubated in the wells coated with anti-Ig received 100 µl of medium alone. Cells were incubated for 60 min (basophil preparations) or 180 min (eosinophils) at 37°C and 5% CO2. After incubation, supernatants were collected and stored at -20°C until assayed. To quantitate basophil degranulation, the concentration of histamine in the sample supernatants was measured by histamine enzyme immunoassay (EIA) kit (Immunotech) following the procedure recommended by the manufacturer. To quantitate eosinophil degranulation, the concentration of eosinophil-derived neurotoxin (EDN) in the sample supernatants was measured by RIA. The RIA is a double-Ab competition assay in which radioiodinated EDN, rabbit anti-EDN, and burro anti-rabbit IgG are used, as reported elsewhere (5). Total cellular histamine and EDN contents were measured simultaneously using supernatants from cells lysed with 0.5% Nonidet P-40 detergent. The sensitivities of histamine EIA and EDN RIA were 0.2 nM and 2.0 ng/ml, respectively. All assays were done in duplicate.
Statistical analysis
Data are presented as mean ± SEM from the numbers of experiments indicated. Statistical significance was assessed using the Mann-Whitney U test or paired Student’s t test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Generally, in Rochester, MN, ragweed hay fever season starts in
early August, peaks at the end of August, and continues until the end
of September. In the current study, patients’ symptoms increased in
parallel to the pollen counts and peaked during the peak pollination
period (late August through early September). Similarly, patients’
peripheral blood eosinophil counts increased significantly during the
peak hay fever season from the values before the season (284 ±
41/µl at peak season vs 141 ± 19/µl before the season (early
July); n = 30, p < 0.01). Initial FACS
analyses of peripheral blood eosinophils validated in vivo
"activation" of the patients’ eosinophils. These cells expressed
significantly higher levels of several eosinophil cell surface
activation markers, including CD25, CD69, and CD11b (21, 22, 23), compared
with eosinophils from normal subjects (Fig. 1
). The increased expression of CD11b was
especially pronounced. In addition, when stimulated in vitro with 1
µM FMLP, patients’ eosinophils released significantly larger amounts
of the eosinophil granule protein, EDN, than normal subjects’
eosinophils (27.3 ± 5.6 and 7.8 ± 1.1% of total EDN for
patients and normal subjects, respectively; n = 6,
p < 0.05). These findings suggest that blood
eosinophils from patients with hay fever are "activated" compared
with those from normal subjects, presumably due to exposure to
cytokines and other inflammatory mediators in vivo.
|
We next examined whether patients’ eosinophils can bind IgE or
IgG on their surfaces and whether they express receptors for these Ig.
FACS analysis of cell-bound IgE with FITC-conjugated anti-IgE Ab
showed that none or minimal amounts of IgE were present on the surface
of freshly isolated eosinophils (Figs. 2
and 3A). When these
eosinophils were incubated with IgE in vitro for 2 h, the amounts
of bound IgE significantly increased (p <
0.01, Fig. 3A), suggesting that these eosinophils have the
capacity to bind IgE. Freshly isolated eosinophils bound detectable
amounts of IgG on their surface (Figs. 2 and 3A), and the
levels increased further by in vitro incubation with IgG (Fig. 3A, p < 0.01). Thus, patients’ eosinophils
were able to bind both IgE and IgG in vitro although the amounts of IgE
bound in vivo were negligible.
|
|
RI or FcRII were detected by
15-1 or 9P25 mAb, respectively, in any of the six patients (Figs. 2
and 3B). In contrast, FcRII was easily detectable on the same
eosinophils (Figs. 2 and 3B), consistent with the IgG
binding capacity of eosinophils. The experiments were repeated without
removing IgE from eosinophils; neither FcRI nor FcRII could be
detected on these nonstripped eosinophils (data not shown).
Analysis of FcRI expression by biotinylated cIgE
Perhaps we were unable to detect FcRI on eosinophils because of
the sensitivity limitations of the standard FACS technique. Therefore,
to enhance the signals for FACS analysis, we removed any IgE bound to
the cells by lactic acid treatment, then incubated them in vitro with
biotinylated cIgE, and visualized them with PE-conjugated streptavidin.
Fig. 4A shows that the binding
of cIgE to eosinophils was clearly detectable. Furthermore, the cIgE
binding was inhibited by pretreatment of cells with anti-FcRI
mAb (15-1) (Fig. 4B); the isotype-matched control Ig for
anti-FcRI did not affect cIgE binding to eosinophils. Similar
analyses of eosinophils from six patients showed that this observation
is reproducible and that the binding of cIgE was significantly blocked
by anti-FcRI mAb (Fig. 5A, p <
0.01). The inhibitory effects of anti-FcRI mAb were variable
among the patients, suggesting a wide variance in the expression levels
of FcRI. Interestingly, eosinophils from normal subjects also bound
cIgE, and the binding was also inhibited partially but significantly by
anti-FcRI mAb (Fig. 5A, p < 0.05).
The amounts of bound cIgE levels as well as the amounts of cIgE-binding
inhibited by anti-FcRI mAb tended to be higher in eosinophils
from patients compared with eosinophils from normal subjects. Recent
studies on mast cells and basophils show that the density of cell
surface FcRI is regulated by blood IgE levels (29, 30). In
eosinophils, we also found that the expression of FcRI, as estimated
by anti-FcRI-inhibitable binding of cIgE, showed a strong linear
correlation with serum levels of total IgE (Fig. 5B,
r2 = 0.941).
|
|
RI expression
Although we could detect FcRI expression by the
cIgE-biotin-streptavidin technique (Figs. 4 and 5), the failure to
detect by the standard FACS technique (Figs. 2 and 3) suggested that
the density of FcRI expressed on the surface of the eosinophil is
extremely low. Basophils are known to express a high density of FcRI
(31) and blood monocytes from patients with allergy also express
FcRI (32). Therefore, we compared the expression levels of FcRI
by eosinophils to other leukocytes. To this end, total leukocytes were
sensitized with biotinylated cIgE and visualized by PE-conjugated
streptavidin and flow cytometry. The individual cell populations were
electronically gated on their unique green autofluorescence and side
scatter characteristics. This strategy avoided unanticipated cell
activation during the isolation procedures, avoided inadvertent subset
selection, and enabled simultaneous analyses of different cell
populations from the same sample. As shown in Fig. 6, A and B, green
fluorescence/side scatter dot-plot of leukocytes incubated with
PE-streptavidin alone without sensitization with cIgE revealed four
distinct cell populations, including lymphocytes plus basophils,
monocytes, neutrophils, and eosinophils. When cells were sensitized
with biotinylated cIgE and stained with PE-streptavidin, we found that
cIgE binds to all the cell populations to various degrees (Fig. 6C). Notably, basophils were intensely stained, forming a
unique population distinct from the lymphocyte cluster in PE
fluorescence reading (Fig. 6C) and even in the green
fluorescence reading (Fig. 6D). The distribution of cell
clusters in green fluorescence/side scatter dot-plot did not change
otherwise (Fig. 6D). Interestingly, when leukocytes were
pretreated with anti-FcRI mAb, the binding of cIgE to the
basophil cluster was completely inhibited (Fig. 6E). In contrast, the
inhibitory effects of anti-FcRI mAb on cIgE binding to
eosinophil and neutrophil clusters were not apparent in the dot-plot.
|
RI mAb (15-1) are shown in Fig. 7. Binding of cIgE to leukocytes was
observed in the following order: basophils (5062 ± 221) >>
neutrophils (346 ± 23) > eosinophils (124 ± 15) 
monocytes (90 ± 9) (means ± SEM of MFI,
n = 27). Pretreatment of cells with anti-FcRI
mAb completely inhibited the cIgE binding to basophils (99.0 ±
0.1% of inhibition, n = 27), suggesting that FcRI
plays a major role in cIgE binding to basophils. The binding of cIgE to
monocytes was inhibited moderately but significantly by
anti-FcRI mAb in 24 of 27 patients (45.0 ± 5.5% of
inhibition, p < 0.01), suggesting roles for FcRI
and for other binding site(s) in cIgE binding to monocytes. In
contrast, in neutrophils anti-FcRI mAb minimally affected the
binding of cIgE (5.1 ± 5.3% of inhibition), and the inhibitory
effects of anti-FcRI mAb varied considerably among individuals
as evidenced by a wide range of (W/O-With) values. The effects of
anti-FcRI mAb on eosinophils were between those on monocytes and
neutrophils, namely significant (p < 0.01) but
small and variable inhibitory effects (18.2 ± 3.8% of
inhibition). The binding of cIgE to leukocytes due to FcRI was
observed in the following order: basophils (5015 ± 223) >>
monocytes (40 ± 12) > eosinophils (25 ± 7) > neutrophils
(13 ± 13) (means ± SEM of the MFI with or without
anti-FcRI mAb). Thus, although the expression of FcRI by
eosinophils is detectable, the level is strikingly less than that by
basophils and less than that by monocytes.
|
In the next series of experiments, we examined the functional
significance of IgE receptors for mediator release by eosinophils. The
engagement of IgE receptors on basophils and mast cells by ligation of
bound IgE with anti-IgE Ab is commonly used to trigger various
effector functions of these cell types, including degranulation and
production of lipid mediators and cytokines (reviewed in Refs. 33 and
34). Therefore, we investigated whether anti-IgE Ab induces
degranulation of eosinophils and compared the responses of basophils
and eosinophils using cells isolated from the same donors. When
basophils were stimulated with 1 µg/ml anti-IgE Ab in solution,
they released a large amount of histamine (Fig. 8A, 85.7 ± 4% of total
histamine, n = 6). The effect of 10 µg/ml
anti-IgE Ab in solution was weaker, but substantial amounts of
histamine were released (51.3 ± 10.5% of total,
n = 6). In contrast, 1 µg/ml anti-IgE Ab in
solution failed to induce EDN release from eosinophils (Fig. 8B), and higher concentrations of anti-IgE Ab (550
µg/ml) did not induce EDN release from eosinophils (data not shown).
Furthermore, immobilized anti-IgE Ab induced histamine release from
basophils but not EDN release from eosinophils. In contrast,
immobilized anti-IgG Ab induced both basophil histamine release and
eosinophil EDN release.
|
The experiments described above suggest that a degranulation
response is provoked by anti-IgE Ab in basophils, but not in
eosinophils. Perhaps the amounts of IgE mounted on eosinophils are too
small to trigger sufficient signals for cellular function when
cross-linked by anti-IgE Ab. Furthermore, if the affinity of the
IgE binding sites is low, the sites may be partially occupied and
partially unoccupied in freshly isolated eosinophils (see Figs. 2 and 3). Therefore, to characterize completely the eosinophil responses to
Ig, we stimulated eosinophils with Ig itself or anti-Ig Ab
immobilized onto tissue culture plates and examined the production of
an eicosanoid, LTC4. As shown in Fig. 9, ligation of unoccupied IgG receptors
by immobilized IgG provoked LTC4 production in a
concentration-dependent manner. In contrast, ligation of unoccupied IgE
receptors (or binding sites) with immobilized IgE failed to induce
LTC4 production. Similarly, ligation of cell-bound IgG by
anti-IgG Ab stimulated eosinophil LTC4 production;
however, anti-IgE Ab did not induce significant production of
LTC4. Furthermore, as shown in Fig. 10A, eosinophils produced
superoxide anion in a time-dependent manner when stimulated with
immobilized IgG; immobilized IgE stimulated only minimal superoxide
production. In addition, anti-IgG Ab, but not anti-IgE Ab,
induced eosinophil superoxide production (Fig. 10B).
Altogether, these findings suggest that eosinophil degranulation and
production of LTC4 and superoxide anion are only minimally
induced by IgE-dependent stimuli, such as IgE or anti-IgE Ab,
whereas these functions are induced vigorously by IgG-dependent
stimuli.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RI on mast cells and basophils triggers production
and release of inflammatory mediators from these cells (reviewed in
Refs. 33 and 34). Because eosinophils from hypereosinophilic patients
may express FcRI (13), it has been suggested that FcRI plays an
important role in mediator release by eosinophils in allergic disease
(35). However, we have made four major observations suggesting that
this is unlikely. First, by standard FACS analysis, eosinophils from
patients with hay fever during the peak of the season express no or
minimal levels of FcRI or FcRII on their surfaces (Figs. 2 and 3); however, these eosinophils do express eosinophil activation markers
and FcRII. Second, eosinophils are able to bind IgE (Fig. 3);
however, in contrast to basophils, the contribution of FcRI toward
eosinophil IgE binding is minor (Fig. 7). Third, IgE-dependent stimuli
fail to provoke the release of inflammatory mediators by eosinophils
(Figs. 8, 9, and 10). Fourth, in contrast, eosinophils from the same
donors express FcRII (Fig. 2) and bind IgG (Fig. 3), and their
effector functions are triggered by IgG-dependent stimuli (Figs. 8, 9, and 10). Therefore, eosinophils from patients with ragweed hay fever
express negligible FcRI or FcRII. Unlike mast cells and
basophils, factors other than IgE, such as IgG and cytokines, likely
play roles in activation of eosinophils in allergic diseases.
The expression of IgE receptors by eosinophils has been a controversial
issue. Murine eosinophils do not express FcRI or FcRII (36).
Although human eosinophils were considered to express only the low
affinity IgE receptor (FcRII) (12, 37, 38, 39), a more recent study by
the same investigators demonstrated that eosinophils from
hypereosinophilic patients express FcRI (13). This latter report
triggered clinical studies to examine eosinophil expression of FcRI.
For example, the transcription of mRNA for -, ß-, and/or
-chains of FcRI in eosinophils was detected during
allergen-induced late phase cutaneous reactions in patients with
allergy (16). In addition, local allergen provocation induced
expression of FcRI protein by eosinophils infiltrating into the
airways (14) and skin (15, 16). However, other investigators showed
that the expression of FcRI protein was not detected on the surface
of the eosinophil (40, 41). In addition, we showed that eosinophils
from patients with hay fever express very low numbers of FcRI on
their surface, which can be detected by the biotin-streptavidin
technique but not by the standard FACS technique. Although the reasons
for the discrepancies in our observations and some of the previous
studies are unknown, we speculate that synthesis of receptor proteins
and surface expression of the receptor proteins are regulated under
different mechanisms. In other words, the detection of receptor
proteins by immunohistochemistry may not necessarily represent the
expression of the proteins on cell surface. For example, human T cells
contain proteins for low affinity IgG receptor (FcRII, CD32) within
the cytoplasm, but these receptors are not expressed on the cell
surface (18). A similar scenario is known for FcRIII (CD16) in
eosinophils (42). In addition, the earlier studies often used
eosinophils from patients with hematologic or immunologic disorders,
including idiopathic hypereosinophilic syndrome and lymphomas. Perhaps
disease heterogeneity accounts for the receptor expression
heterogeneity.
The strikingly different responses to IgE between basophils and
eosinophils are apparent in this study. From the MFI of FACS analysis,
50 times more cIgE was bound to basophils than to eosinophils (Fig. 7). Furthermore, >99% of cIgE binding to basophils was inhibited by
anti-FcRI mAb pretreatment, suggesting that FcRI is the
predominant IgE receptor in basophils. In contrast, anti-FcRI
mAb only partially (18%) inhibited cIgE binding to eosinophils. Thus,
we estimate that eosinophils from these ragweed hay fever patients
express 0.5% of the FcRI compared with basophils. In addition,
the basophil degranulation response was triggered by ligation of
FcRI, either by anti-IgE Ab or by immobilized IgE; these same
stimuli failed to induce degranulation or release of mediators from
eosinophils. Interestingly, treatment of patients with atopy with
humanized anti-IgE Ab resulted in 96% decrease in the
expression of FcRI and 40% decrease in anti-IgE-induced
histamine release response by basophils (30). Because eosinophils
possess only 0.5% of the FcRI possessed by basophils, this
paucity of FcRI expression by eosinophils may explain the
heterogeneous responses of eosinophils and basophils to FcRI
ligation. Perhaps eosinophils lose or basophils acquire their capacity
to express high levels of FcRI during their differentiation from
common progenitors (43).
Then, the question remains as to which receptors or binding sites
contribute to the binding of myeloma IgE (Figs. 2 and 3) and cIgE
(Figs. 4 through 7) by eosinophils. The results of our experiments and
a review of the literature provide some insights. As shown in Figs. 2 and 3, none to minimal expression of FcRI or FcRII (CD23) was
detectable using 15-1 or 9P25 mAb, respectively, and the standard FACS
technique. These findings are also consistent with previous reports,
demonstrating negligible binding of anti-FcRI mAb (clones 22E7
and 15-1) (17) and a panel of mAb against CD23 (44, 45) to eosinophils.
Therefore, the expression levels of high affinity IgE receptor
(FcRI) and low affinity IgE receptor (FcRII) on eosinophils are
likely too low to play major roles in IgE binding. Furthermore, the
minimal or lack of binding of IgE on eosinophils in vivo (Figs. 2 and 3) suggests that IgE binding to eosinophils is of low affinity. Besides
these IgE Fc receptors, an IgE-binding molecule belonging to an S-type
lectin family, called Mac-2/BP, can also bind IgE through the
carbohydrate recognition domain (46). Mac-2/BP is expressed by
various cell types, including neutrophils and eosinophils (47, 48).
Indeed, Ab against Mac-2/BP strongly inhibited IgE binding and
IgE-dependent activation of human neutrophils (47). Therefore, by
analogy to neutrophils, Mac-2/BP is a good candidate molecule for
IgE-binding sites on eosinophils. Our findings, showing large amounts
of IgE binding to both eosinophils and neutrophils independent of
FcRI (Fig. 7), are consistent with this speculation. Alternatively,
IgE may bind nonspecifically through its carbohydrate moiety to
lectin-like binding sites on the surface of the eosinophil.
Two other questions remain. Because previous histological studies
(14, 15, 16) could not differentiate between the proteins within the cells
and those on the cell surface, the first question involves the surface
expression of FcR by tissue eosinophils. As shown in Fig. 5B, the levels of FcRI expression by eosinophils show a
positive correlation with serum IgE levels, suggesting that the
expression of FcRI by eosinophils is regulated by IgE-dependent
mechanisms similar to those in mast cells (29) and basophils (30). In
addition, IL-4 enhances the expression of mRNA and protein for the
-chain of FcRI (40). It is well known that IL-4 is expressed in
the local tissues of patients with allergy (49, 50); a more recent
report suggests that IgE may be produced by B cells in the airway
tissues in patients with allergy (51). Therefore, the tissue
microenvironment may be optimal for FcRI expression, suggesting that
the surface expression of FcRI by tissue eosinophils may be higher
than the surface expression in peripheral blood eosinophils. On the
other hand, even in the optimal tissue environment, the ability of
eosinophils to express FcRI on their surfaces may be limited. For
example, eosinophils obtained from bronchoalveolar lavage fluids from
patients with allergy do not express detectable levels of FcRI on
their surfaces (41). Eosinophils cultured for up to 11 days with
myeloma IgE or IL-4, conditions known to up-regulate FcRI on
basophils, failed to induce any detectable surface FcRI (41).
Furthermore, Terada et al. (40) found that after culture with IL-4,
eosinophil production of FcRI -subunit protein was increased, but
expression of FcRI was not detectable. In preliminary studies, we
were not able to detect FcRI by standard FACS analysis on
eosinophils cultured for 3 days with human myeloma IgE in the presence
of various cytokines, such as IL-5, IL-3, IL-4, TNF- and fibroblast
supernatants (data not shown).
The second question concerns the interaction of IgE with other
eosinophil agonists. We found that anti-IgE Ab or immobilized IgE
by themselves failed to provoke eosinophil mediator release (Figs. 8 through 10). However, a previous report also suggests that IgE mediates
eosinophil killing of schistosomula of Schistosoma mansoni
when cells are stimulated with lipid mediators, such as
platelet-activating factor or LTB4, without the apparent
expression of FcR (52). Therefore, IgE may collaborate with other
eosinophil agonists to provoke some functions of the cells, such as
cellular cytotoxicity. In preliminary studies, IgE did not enhance EDN
release from eosinophils stimulated with platelet-activating factor or
IL-5 (H. Kita and M. Muraki, unpublished observation). Furthermore, it
is also possible that FcRI or IgE may be involved in other functions
than secretory process of eosinophils, such as Ag presentation, as
proposed in other cell types (53).
Remaining questions aside, our study clearly shows that the surface
expression of FcRI by blood eosinophils from patients with hay fever
is minimal and that, in contrast to IgG, IgE does not mediate release
of inflammatory mediators from these eosinophils. Therefore, previous
observations regarding the expression of FcRI and FcRII by blood
eosinophils from patients with eosinophilia may not be directly
applicable to patients with allergy. Although our study may be limited
because only patients with hay fever were studied, our observations
raise questions about the current conception that eosinophil IgE
receptors play a role in the pathophysiology of allergic diseases. The
observations in IgE knockout mice, which demonstrated eosinophilic
inflammation and increased bronchial hyperreactivity in the absence of
IgE (54), may in part be applicable to humans; i.e., factors other than
IgE may act as trigger(s) for mediator release by eosinophils in
allergic inflammation.
|
Acknowledgments |
|---|
RI mAb (15-1) and biotinylated cIgE; Mr. James E.
Tarara, Ms. Teresa J. Halsey, and Ms. Holly B. Lamb for
their excellent technical assistance in flow cytometry; Ms. Cheryl
Adolphson for editorial assistance; Ms. Linda Arneson for secretarial
assistance; and Ms. Judith Blomgren and Ms. Kay Bachman for recruiting
and obtaining blood samples from patients. |
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Hirohito Kita, Department of Immunology, Mayo Clinic, Rochester, MN 55905. E-mail address:
3 Abbreviations used in this paper: mIgG, mouse IgG1; cIgE, chimeric mouse/human IgE; EDN, eosinophil-derived neurotoxin; HSA, human serum albumin; LT, leukotriene; PAB, PBS containing 0.1% NaN3 and 1% BSA; MFC, differences in the mean fluorescence channel; EIA, enzyme immunoassay.
Received for publication November 10, 1998. Accepted for publication February 10, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RI) receptor- chain mRNA and protein-bearing eosinophils in human allergen-induced atopic asthma. Am. J. Respir. Crit. Care Med. 158:233.RI+ eosinophils in human atopic skin. Eur. J. Immunol. 27:1236.[Medline]
RI)-bearing eosinophils, mast cells, macrophages and Langerhans’ cells in allergen-induced late-phase cutaneous reactions in atopic subjects. Immunology 93:281.[Medline]
RI) on peripheral blood basophils, monocytes, and eosinophils in atopic and nonatopic subjects: relationship to total serum IgE concentrations. J. Allergy Clin. Immunol. 99:699.[Medline]
RII) in normal human peripheral blood mononuclear cells. Immunology 86:525.[Medline]
RII/CDw32) by human circulating T and B lymphocytes. J. Immunol. 150:5175.[Abstract]
, normal subjects; •, data from patients with hay
fever.
