| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
RI-Specific Fusion Proteins Is Dependent on Receptor Number and Modified by IgG1
Departments of
*
Physiology and
Microbiology, Dartmouth Medical School, Lebanon, NH 03756; and
Medarex, Bloomsbury, NJ 08804
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RI (CD64), is constitutively
expressed exclusively on professional APCs. Human FcRI binds
monomeric IgG with high affinity and is, therefore, saturated in vivo.
The binding of IgG to FcRI causes receptor recycling, while Abs that
cross-link FcRI cause rapid down-modulation of surface FcRI.
Because studies performed in the absence of ligand may not be
representative of FcRI modulation in vivo, we investigated the
ability of FcRI-cross-linking Abs and non-cross-linking derivatives
to modulate FcRI in the presence and absence of ligand. In the
absence of ligand mAb H22 and wH22xeGFP, an enhanced green
fluorescent protein (eGFP)-labeled fusion protein of H22, cross-linked
and rapidly down-modulated surface FcRI on the human myeloid cell
line, U937, and its high FcRI-expressing subclone, 10.6. This effect
was dependent on the concentration of fusion protein and the level of
FcRI expression and correlated with internalization of both
wH22xeGFP and FcRI, itself, as assessed by confocal microscopy. A
single-chain Fv version, sFv22xeGFP, which does not cross-link FcRI,
was unable to modulate FcRI in the absence of IgG. However, if
ligand was present, treatment with either monovalent or cross-linking
fusion protein led to intracellular receptor accumulation. These
findings suggest at least two alternate mechanisms of internalization
that are influenced by ligand and demonstrate the physiologic potential
of FcRI to transport a large antigenic load into APCs for
processing. These studies may lead to the development of better
FcRI-targeted vaccines, as well as therapies to down-modulate FcR
involved in autoimmune diseases. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RI (or CD64), is
constitutively expressed exclusively on professional APCs (monocytes,
macrophages, and dendritic cells). Human FcRI binds monomeric human
IgG1 and IgG3, as well as mouse IgG2a and IgG3, with high affinity
(108–109
M-1) and is consequently saturated in vivo
(reviewed in Refs. 1, 2, 3, 4, 5, 6). Several mAbs have been developed
which bind FcRI outside the ligand-binding site and therefore are
able to bind in the presence of IgG (7). When
cross-linked, these mAbs have been shown to trigger FcRI functions
such as phagocytosis, Ab-dependent cellular cytotoxicity, and
superoxide generation (1, 8, 9, 10, 11). Targeting Ags
specifically to FcRI via one of these mAbs has been shown to enhance
Ag presentation, as evidenced by T cell proliferation
(12, 13, 14) as well as cytotoxicity in vitro
(15) and vaccine-like effects in vivo (16, 17). Thus, FcRI-targeted Ags are currently being developed as
vaccines (16, 18, 19, 20, 21). However, the mechanism by which
FcRI-targeted Ags are processed and presented is unknown.
Abs that cross-link FcRI have been shown to down-modulate surface
expression of FcRI and, therefore, have been used in the treatment
of immune thrombocytopenia purpura, an autoimmune disease
characterized by the FcR-mediated destruction of platelets. Treatment
of immune thrombocytopenia purpura patients with either mAb 197
(22) or mAb H22 (23) led to down-modulation
of surface FcRI on peripheral blood monocytes and correlated with
elevated platelet count.
In vitro studies on human monocytes have shown that when FcRI is
cross-linked, receptors are internalized (rather than shed) within
hours of cross-linking. Furthermore, the down-modulation of surface
FcRI by mAb H22 was shown to be concentration-dependent, such that
0.1 µg/ml of H22 was optimal and lower or higher concentrations were
less effective (24). It has been demonstrated that FcRI
recycles into and out of the cell when ligand is bound, and that the
addition of an anti-FcRI Ab which binds outside the
ligand-binding site causes receptors to accumulate inside the cell and
not recycle to the cell surface (25). Because prior
studies performed in the absence of natural ligand (human IgG1 or mouse
IgG2a) may not be representative of FcRI modulation in vivo, we
chose to investigate further the capacity of H22 and FcRI-targeted
fusion proteins to modulate FcRI in the presence and absence of
ligand.
In this work we describe FcRI down-modulation (hereafter referred to
as "modulation") on the human myeloid cell line U937 and its high
FcRI-expressing subclone 10.6, using H22 and wH22xeGFP, a fusion
protein of H22 and enhanced green fluorescent protein
(eGFP).3 Both
modulators (H22 and wH22xeGFP) can bind to FcRI trivalently, because
they contain an intact human IgG1 Fc region in addition to two Fab
regions specific for FcRI. We also examined sFv22xeGFP, an eGFP
molecule genetically fused to the single-chain version of H22, which
binds FcRI monovalently. The goal of these studies was to examine
the efficiency of FcRI internalization and modulation in response to
the two new eGFP fusion proteins and the parent H22. We also explored
the contribution of the relative level of receptor expression and the
presence of ligand on modulation. Our findings suggest that trivalent
anti-FcRI constructs modulate FcRI in the absence of IgG in a
concentration-dependent manner similar to the modulation of monocyte
FcRI reported by Wallace et al. (24), although the
kinetics of modulation are more rapid on 10.6 cells. However, the
presence of IgG leads to greater modulation at higher concentrations of
modulator and permits the internalization of non-cross-linked
anti-FcRI constructs. Two different models are proposed for
receptor modulation in the presence and absence of ligand.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
U937, a human myeloid cell line expressing FcRI, was obtained
from American Type Culture Collection, (Manassas, VA)
(26). 10.6 cells, a high FcRI-expressing subclone of
the parent U937 line, were a kind gift of Dr. P. M. Guyre
(Dartmouth Medical School, Lebanon, NH) (8). Cells were
cultured in RPMI 1640 supplemented with 10% FBS,
L-glutamine, 25 mM HEPES, and 50 µg/ml gentamicin (all
from BioWhittaker, Walkersville, MD) (complete medium) in the presence
or absence of 25 ng/ml IFN- (Genentech, South San Francisco, CA) at
37°C/5% CO2.
Abs and reagents
Anti-FcRI Abs used as modulators included H22 (7, 27); H22 F(ab')2; MDX-H210, a bispecific
Ab comprised of the Fab' of mAb H22 and anti-HER2/neu
(28); and 197, a mAb which has been shown to cross-link
and modulate FcRI (10, 22, 29) (all kind gifts of
Medarex, Bloomsbury, NJ). Also used were two FcRI-targeted eGFP
fusion proteins, sFv22xeGFP and wH22xeGFP, constructed from mAb H22 and
eGFP (Ref. 30 and C. Guyre, M. Barreda, S. Swink, and M.
Fanger, manuscript in preparation). sFv22xeGFP, which binds
FcRI monovalently, is a single-chain Fv fragment of mAb H22 fused to
eGFP. wH22xeGFP, which binds FcRI both by its Fab ends and its Fc
end, thereby cross-linking FcRI, is a fusion protein of the whole
H22 mAb with eGFP fused to the CH3 domains of the H22 H chains
(31). The amount of aggregates in the fusion proteins
varied from 0–4%, and the percentage of aggregates did not correlate
with modulation activity. Saturation curves were generated for each
individual batch of fusion protein that was made, and the 1/2 maximum
binding for individual batches was determined. In all cases, 5 nM was
<1/2 maximum for wH22xeGFP, and 15 nM was < 1/2 maximum for
sFv22xeGFP. The 1/2 maximum binding for H22 was 
4 nM, and 5 nM was
subsaturating. Some modulation assays were performed in the presence or
absence of human gammaglobulin (Sigma-Aldrich, St. Louis, MO) at a
final concentration of 3 mg/ml in medium containing 0.5 mg/ml BSA.
FcRI was detected using a murine mAb, 32.2 (Medarex), which binds
FcRI outside the ligand-binding domain at an epitope distinct from
the H22 binding site (7, 32), followed by a
R-PE-conjugated F(ab')2 goat anti-mouse IgG
(H and L chains, minimal cross-reaction to human, bovine, horse,
rabbit, and swine; Jackson ImmunoResearch Laboratories, West Grove,
PA). For confocal microscopy studies, detection Ab was Cy3-conjugated
F(ab')2 goat anti-mouse IgG plus IgM (Jackson
ImmunoResearch Laboratories). P3, an irrelevant murine IgG1, was used
as an isotype control.
Modulation assay
Cells were treated with varying amounts of modulator ± 3
mg/ml IgG in complete medium from 30 min to 24 h either at 37°C
or on ice. In some studies, normal human serum (NHS; Gemini
Bio-Products, Calabasas, CA) or mouse serum (Sigma-Aldrich) was used in
place of FBS. Following treatment, cells were washed twice with
ice-cold PBA (PBS containing 1% BSA (Sigma-Aldrich) and 0.05% sodium
azide; Fisher Scientific, Fairlawn, NJ) and stained on ice for surface
FcRI in a 96-well round-bottom polypropylene microtiter plate
(Costar, Corning, NY). Briefly, cells were resuspended in 20 µg/ml
mAb 32.2 and incubated on ice for 1 h. Cells were then washed
three times with ice-cold PBA, resuspended in a 1/40 dilution of
R-PE-conjugated goat anti-mouse IgG, and incubated on ice for 45
min. Cells were washed twice with ice-cold PBA and resuspended in 1%
methanol-free formaldehyde (MFF; Polysciences, Warrington, PA) in PBS.
Fixed cells were stored at 4°C for at least 16 h before flow
cytometric analysis was performed.
Flow cytometric analysis
Fixed cells were analyzed by flow cytometry using a FACScan (BD
Biosciences, San Jose, CA). For each experiment compensation was set
using unstained ("autofluorescent") or isotype control, green-only,
and red-only samples to eliminate fluorochrome crossover between FL-1
and FL-2. Ten thousand events were collected for each sample in a given
experiment. FL-2 data represent the red channel (585/42 nm filter),
demonstrating the relative amounts of FcRI on the surface of cells
following treatment. Modulation was calculated from FL-2 data using the
following equation, where MFI represents the geometric mean
fluorescence intensity:
 |
RI expression) is reflected in increased
percent modulation. Confocal microscopy
Cells were subjected to receptor modulation as described above
and then fixed in ice-cold 1% MFF (Polysciences) for a minimum of
1 h. Fixed cells were then permeabilized in PBAS (PBA containing
0.5% saponin; Sigma-Aldrich) and stained as described above using PBAS
as wash buffer and Ab diluent. Following staining, cells were fixed in
1% MFF and stored at 4°C for a minimum of 16 h. Cells were then
pelleted and 3.5 µl of cells were placed on a glass slide (Gold
Seal Products, Portsmouth, NH). Approximately 3.5 µl of Prolong
Antifade reagent (Molecular Probes, Eugene, OR) was added, followed by
an 18-mm glass coverslip (Costar) which was then sealed with
nail varnish. Imaging was performed exactly as described elsewhere
(31) on a Bio-Rad MRC-1024 Confocal Scanning Laser
Microscope system (Bio-Rad, Hercules, CA) using a krypton/argon laser
and LaserSharp version 3.2 software (Bio-Rad).
Molar conversions
To compare data displayed in micrograms per milliliter to molar concentrations, molar equivalents for each fusion protein and Ab used were calculated using the estimated m.w., as determined by Gene Inspector software (Textco, West Lebanon, NH). The molar equivalents of 1 µg/ml of fusion protein are 17.7 nM sFv22xeGFP, 4.96 nM wH22xeGFP, and 6.77 nM H22. The m.w. of mAb 197 was estimated to be equal to H22, because both are IgG.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RI is required for receptor modulation
in the absence of ligand
Modulation assays were performed by treating cells with a
potential modulator at 37°C and then staining cells on ice for
residual surface FcRI using mAb 32.2 followed by R-PE-conjugated
goat anti-mouse IgG. Studies using directly labeled R-PE-32.2
demonstrated results similar to this two-step approach (data not
shown). MFI of treated cells was compared with that of untreated cells
to determine the loss of cell surface receptors due to the modulator.
Fig. 1
illustrates the capacity of fusion
proteins sFv22xeGFP and wH22xeGFP to modulate surface FcRI. Although
previous studies showed that H22 does not block the binding of 32.2, we
tested whether the inclusion of eGFP in our fusion proteins inhibited
the binding of 32.2. Surface expression of FcRI on cells treated
with sFv22xeGFP or wH22xeGFP on ice was not different from cells
treated with PBS, indicating that neither fusion protein blocked the
binding of the detection Ab, 32.2 (Fig. 1A). Cells treated
for 2 h with PBS or monovalent sFv22xeGFP were unable to modulate
FcRI at either temperature, while the multivalent wH22xeGFP reduced
surface expression 50% at 37°C (Fig. 1A). Time course
studies confirmed the inability of sFv22xeGFP to modulate FcRI (Fig. 1B). However, modulation of surface FcRI by wH22xeGFP was
40% within 10 min and appeared to be maximal by 30 min (Fig. 1B).
|
RI mAbs and fusion proteins were superior to saturating
concentrations for maximally modulating FcRI in 4- and 16-h assays
(data not shown). Therefore, to directly compare the modulation
kinetics of wH22xeGFP and other Abs, cells were treated with
subsaturating concentrations of each. Because the binding of one
modulator (mAb 197) to FcRI blocks the binding site of the detection
Ab, mAb 32.2, the standard modulation assay had to be modified. FcRI
was measured indirectly for samples treated with 197 using
PE-conjugated goat anti-mouse IgG, which detects the surface-bound
197 itself. For samples treated with the other fusion proteins and Abs,
FcRI was measured using mAb 32.2, as in the standard modulation
assay. For each construct, modulation was calculated based on cells
treated with the given construct at 37°C to cells treated with that
same construct on ice. Fig. 2, like Fig. 1
, demonstrates that treatment with medium alone or sFv22xeGFP did not
lead to FcRI modulation. However, cells treated with equimolar
amounts of wH22xeGFP, H22, or 197 led to varying degrees of modulation,
with 197 modulating most rapidly, followed by wH22xeGFP. This variation
among modulators at equimolar concentrations may be due to differences
in binding kinetics and/or how well each can cross-link FcRI. Of
note, subsaturating concentrations of wH22xeGFP led to slower FcRI
modulation than did saturating concentrations (Fig. 1).
|
RI
To better understand the nature of receptor modulation, confocal
microscopy was performed to determine whether the reduction in surface
FcRI was due to receptor internalization. IFN--treated U937 cells
were incubated with sFv22xeGFP, wH22xeGFP, or H22 as modulators and
then fixed and examined for intracellular FcRI using mAb 32.2. In
4-h modulation assays, FcRI appeared only on the cell surface if no
modulator was used (data not shown) or if the monovalent sFv22xeGFP was
used as modulator (Fig. 3, top
panel). This lack of internalized receptor correlated with the
lack of FcRI modulation in flow cytometric assays. By contrast, when
wH22xeGFP was used as modulator, intracellular FcRI was observed at
all concentrations that led to receptor modulation in flow assays
(0.1–10 µg/ml) (Fig. 3, middle panel; compare with Fig. 4A). Similarly, when H22 was
used as modulator, internalized FcRI was more prominent at the low
concentrations (0.1–1 µg/ml), which showed maximal receptor
modulation in flow assays (Fig. 3, bottom panel; compare
with Fig. 4). Similar patterns of internalization were observed in 1-h
assays (data not shown). Interestingly, FcRI staining often appeared
polarized to one side of the cell at high concentrations of modulator.
Although we do not know the basis for the polarized staining, it was
often observed when modulator was present and may be due to the
assembly of FcRI into membrane rafts.
|
|
Because human serum contains high levels of IgG, the
ligand-binding domain of FcRI is usually occupied in vivo.
Therefore, it seemed important to determine whether ligand affected
the ability of H22 constructs to modulate FcRI. Fig. 4 demonstrates
that ligand altered modulation by H22 and wH22xeGFP. In the presence of
saturating concentrations of human gammaglobulin, both H22 and
wH22xeGFP were able to modulate FcRI at high concentrations,
including those which, in the absence of ligand, led to little or no
modulation (e.g., 1 and 10 µg/ml H22) or lower levels of modulation
(e.g., 10 µg/ml wH22xeGFP) (Fig. 4A). We next compared
modulation in the presence of serum known to contain IgG not capable of
binding FcRI (FBS) (33) to that which contains human
FcRI-binding isotypes (human and mouse serum). FcRI modulation by
0.1 µg/ml H22 in the absence of ligand was significant, but 10-fold
lower or higher concentrations did not cause modulation. However, if
ligand was present (in the form of NHS (Fig. 4B) or mouse
serum (Fig. 4C)), modulation occurred in the presence of
higher concentrations of H22 (1 and 10 µg/ml). These data support the
role of ligand in altering the ability of H22 to modulate FcRI.
Because ligand seemed to modify the ability of Abs that cross-link
FcRI to modulate FcRI, we investigated the effect of ligand on
modulation by anti-FcRI Ab variants that do not cross-link. In
the absence of ligand, H22 was able to modulate FcRI as expected,
but neither MDX-H210 nor H22 F(ab')2, two
FcRI-specific molecules that lack an Fc region, caused modulation
(Fig. 5A). This indicates that
the trivalent H22, but not its monovalent or divalent derivatives could
modulate FcRI. However, when ligand was present in the form of NHS,
all three Abs were able to modulate FcRI (Fig. 5B). These
data suggest that, in addition to altering the modulation of FcRI by
Abs that cross-link the receptor, ligand affected the ability of
non-cross-linking anti-FcRI Abs to modulate FcRI.
|
RI
constructs and FcRI
Flow cytometric analyses revealed that modulation of FcRI by
high concentrations of H22 and wH22xeGFP could be attained if ligand
was present (Fig. 4). To determine whether this modulation correlated
with the internalization of FcRI, confocal microscopy studies were
performed. Cells were treated with 10 µg/ml modulator in the absence
(Fig. 3, right panels) or presence (Fig. 6) of IgG. In the absence of IgG,
wH22xeGFP and FcRI appeared both on the surface and intracellularly
(Fig. 3, middle right panel). However, when IgG was present,
both fusion protein and receptor appeared predominantly inside the cell
(Fig. 6, middle panel). For cells treated with H22 in the
absence of ligand, FcRI appeared mainly on the cell surface (Fig. 3, bottom right panel), while internalized receptor was evident
when IgG was present (Fig. 6, right panel). These data
correlated with modulation data showing that high levels of wH22xeGFP
were capable of modulating FcRI but the addition of ligand enhanced
the modulation, and that H22 was unable to modulate FcRI at this
concentration unless ligand was present (Fig. 4).
|
RI, in the presence (Fig. 6, left panel), but not absence (Fig. 3, top right
panel), of IgG, indicating the ability of monovalent
FcRI-targeted proteins to be internalized with the receptor in the
presence of ligand. This is in accordance with data showing that a
monovalently targeted anti-FcRI Ab, MDX-H210, was able to
modulate FcRI in the presence of ligand (Fig. 5B).
The ability of anti-FcRI Abs to modulate is dependent on
receptor number
To determine whether the concentration dependency of H22 and
wH22xeGFP to modulate FcRI changed with FcRI levels, two
different U937 cell lines were studied: U937 cells obtained from
American Type Culture Collection (U937 cells) and its high
FcRI-expressing subclone (10.6 cells). Both cell lines were tested
with and without prior culture in the presence of IFN-, a cytokine
known to up-regulate FcRI. IFN- has been shown to influence the
processing of Ag (34, 35) but was reported to not alter
the ability of immune complexes to be internalized through FcR
(36). Table I illustrates
the relative receptor expression of each of these cell lines cultured
for 3 days in the presence and absence of IFN-. As expected,
treatment with IFN- up-regulated FcRI expression 4- to 5-fold,
and FcRI expression for each cell line was significantly different
(Table I). In 16-h modulation assays using these cells, treatment with
sFv22xeGFP did not lead to FcRI modulation (Fig. 7). However, when H22 or wH22xeGFP was
used, modulation occurred, and the concentration that led to maximal
modulation increased with receptor number (Fig. 7). Furthermore, in
some cases wH22xeGFP was able to modulate FcRI over a wider range of
concentrations than H22 (e.g., 3 and 10 nM wH22xeGFP vs 3 nM only H22,
Fig. 7C). In addition, wH22xeGFP led to higher levels of
modulation than H22 in most cases (Fig. 7, A, B,
and D). Similar trends in receptor modulation were observed
in 4-h modulation assays (data not shown).
|
|
RI modulation. To
further illustrate this point, the FcRI expression data for each
modulator were plotted. As shown in Fig. 8, the concentration of H22 or wH22xeGFP
that led to maximal modulation (seen as decreased MFI) shifted to the
right as receptor number increased, whereas sFv22xeGFP did not lead to
decreased FcRI expression at any concentration. The addition of
ligand, generally in the form of saturating concentrations of IgG,
consistently led to an increase in levels of FcRI. Interestingly,
high concentrations (
30 nM) of H22, an IgG1 mAb that cross-links and
modulates FcRI at lower concentrations, resulted in increased
receptor expression, even approaching that of cells treated with a
saturating concentration of IgG (Fig. 8). This supports the hypothesis
that high levels of a proven cross-linker can bind FcRI in a
functionally monovalent fashion and, like the binding of ligand, lead
to increased levels of surface FcRI.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RI using anti-FcRI Abs has been
described for cultured human monocytes (24, 37) and
peritoneal exudate macrophages from human FcRI-transgenic mice
(17), as well as in vivo in patients with immune
thrombocytopenia purpura (22, 23). In the present study,
our aim was to further investigate the nature of FcRI modulation by
a humanized mAb, H22, as well as H22 fusion proteins, using the human
myeloid cell lines U937 and its high FcRI-expressing subclone, 10.6,
as a model. Using wH22xeGFP, an H22-eGFP fusion protein which
cross-links FcRI, we found receptor modulation to be rapid, with the
majority of the modulation occurring within the first 10 min of the
assay and a steady-state reached by 30 min. This is similar to what has
been found in studies by others using U937 cells, which showed
internalization to occur within 5 min after cross-linking
receptor-bound IgG (25). In studies by Wallace et al.
(24), FcRI modulation on normal human monocytes was
evident at 2 h and maximal at 8 h when H22 was used as
modulator. The differences in kinetics observed in that study vs ours
could be due to the nature of the cell populations studied, as well as
the fact that freshly isolated monocytes are likely to have residual
bound IgG, which can alter modulation kinetics. Modulation of FcRI
in our studies was not accompanied by modulation of two other surface
molecules, FcRII or class I (Ref. 31 and data not
shown), which do not bind anti-FcRI constructs. This is in
accordance with modulation studies by Wallace et al. (24),
which demonstrated the specificity of H22 for FcRI. Furthermore,
FcRI modulation in our studies was not blocked by Abs to irrelevant
cell surface receptors, FcRII or CD11b (data not shown), suggesting
that those two molecules did not take part in anti-FcRI-mediated
modulation of FcRI.
Reduction of the surface expression of FcRI may be due to
internalization or to other events, such as shedding. We looked at
internalization using confocal microscopy and found that
internalization of both wH22xeGFP and receptor correlated with the
receptor modulation that was observed in 4-h flow assays. Similar
results were obtained in 1-h assays (data not shown). This suggests
that the receptor is not rapidly recycling to the surface after the
targeted fusion protein is delivered into the cell. Shedding of FcRI
is an unlikely cause of receptor modulation, as cells treated for 30
min with wH22xeGFP on ice, which showed no modulation, had similar
levels of green fluorescence as cells treated at 37°C, which had
50% reduction in surface FcRI (data not shown). Because the
wH22xeGFP fluorescence can be detected even if it has been
internalized, similar levels of green fluorescence between control and
modulated samples indicate that little, if any, modulator has been lost
from the cell. This finding is not consistent with receptor
shedding.
Because high concentrations of IgG are present in vivo, we performed
modulation assays in the presence of ligand to get a sense of how these
modulators will behave under normal physiologic conditions. In the
presence of IgG, H22 and wH22xeGFP both modulated FcRI at high
concentrations of modulator, concentrations at which modulation was
lower or did not occur in the absence of ligand. Confocal microscopy
experiments confirmed the internalization of FcRI under conditions
where modulation was observed. Because high concentrations of modulator
may favor monovalent binding to FcRI, cross-linking and subsequent
internalization will not occur. The addition of ligand has been shown
by others to be required for receptor recycling (25).
Therefore, we speculate that the addition of ligand in our system
allows for the trafficking of FcRI into the cell, and the bound
modulator prevents recycling of the receptor back to the surface,
leading to a net decrease in surface FcRI expression. Alternatively,
ligand binding may enhance the ability of the modulator to bind to the
receptor, thereby increasing the efficiency of FcRI internalization
and modulation. Ligand was also shown to affect the ability of
monovalently and divalently targeted fusion proteins to modulate
FcRI. In the absence of ligand, neither MDX-H210, a bispecific Ab
which binds FcRI via a Fab' of H22, nor H22
F(ab')2, was able to modulate FcRI. However,
when ligand was present, FcRI modulation was observed. Moreover,
confocal analysis revealed that although FcRI distribution for
control cells treated with IgG was similar to control cells in the
absence of ligand, suggesting there was no internalization of receptor
due to immune aggregates that might be present in the IgG (data not
shown), monovalently targeted sFv22xeGFP was internalized in the
presence, but not absence, of IgG. Therefore, it is possible that even
non-cross-linking Abs can sufficiently interrupt ligand-induced
receptor recycling, such that receptors are internalized, but not
re-expressed, on the cell surface.
The influence of ligand on FcRI function has been explored in many
ways. Studies by Pfefferkorn et al. (38, 39) have
shown that signaling events triggered by cross-linking FcRI can be
altered by ligand. In both U937 cells and human monocytes, FcRI
cross-linking triggered significantly greater respiratory burst
(O2-) in the presence of
ligand. This effect was demonstrated with multiple ligands shown to
bind human FcRI, including human, mouse, rabbit, and rat isotypes,
and did not occur with isotypes which fail to bind FcRI
(38). In addition, when FcRI was cross-linked in the
presence of human IgG1, increased phosphorylation of FcRI-associated
-chain was observed (39). In other studies by Harrison
et al. (25), the Fab of anti-FcRI mAb 32.2 was able
to internalize via receptor recycling in U937 cells in the presence,
but not absence, of IgG. However, in their study primaquine was
required to visualize internalized FcRI, as the Fab and the receptor
recycled back out to the surface in the absence of this drug. The
authors proposed that bound ligand permits FcRI to recycle by
causing dissociation of FcRI from a cytoskeletal actin-binding
protein which, in the absence of IgG, prevents the recycling of FcRI
to the surface (25, 40). Our observation that sFv22xeGFP
can be internalized in the presence of ligand is consistent with this
model, although we were able to visualize the internalized eGFP without
the use of primaquine.
We observed that in the absence of IgG, H22 modulated FcRI only at
certain optimal concentrations, but not at 10-fold higher or lower
concentrations. These patterns were observed in both 4- and 16-h
assays, although modulation percentages were generally slightly higher
in overnight assays. We considered the following model: For a given
cell line with a certain number of FcRI molecules, if few H22
molecules are present per cell (leading to a relatively low molar ratio
of H22:FcRI), extensive trivalent cross-linking (and therefore
FcRI modulation) does not occur. Alternatively, cross-linking may
occur by the few H22 molecules present, but to a low enough level that
the overall cell surface expression does not change significantly. In
contrast, if an excess of H22 molecules is present (leading to a
relatively high molar ratio of H22:FcRI), monovalent binding is
favored and receptor cross-linking does not occur. However, there
exists an optimal concentration of H22 which binds trivalently, causing
cross-linking and receptor modulation. If our hypothesis is correct,
the concentration of modulator required to maximally modulate FcRI
should increase when a cell line with higher levels of FcRI is used.
Therefore, we compared two different U937 cell lines, U937 cells and
the 10.6 high FcRI-expressing subclone, cultured in the presence or
absence of IFN-, a cytokine known to up-regulate FcRI expression.
We reasoned that if very few receptors exist on the surface of a cell,
the probability of forming a trivalent complex will be low because the
receptors are spatially too distant for efficient cross-linking within
the 
16-h treatment times used in these assays. With greater receptor
numbers, either as a function of a high-FcRI-expressing clone or
IFN- treatment, a higher concentration of modulator would be
required to maintain the optimal modulator:receptor molar ratio for
maximal FcRI modulation. Indeed, in the absence of ligand, we found
that a higher concentration of modulator (H22 or wH22xeGFP) was
required to optimally reduce surface FcRI levels on cells expressing
a higher receptor number. Incidentally, we found that the concentration
of H22 required to maximally reduce surface expression of FcRI was
similar to that described by Wallace et al. (24) for human
monocytes, which express similar levels of surface FcRI (reviewed in
Ref. 1). Although we used IFN- treatment to up-regulate
surface FcRI on U937 cells, it is possible that this cytokine has
other effects that may influence modulation, such as altering the
FcRI 
- and -chain association (41). The effects
of anti-FcRI Abs and fusion proteins on receptor modulation
described for these cells may therefore be more characteristic of an
IFN--stimulated monocyte.
We observed that treatment with wH22xeGFP led to greater modulation
than H22 in most cases and could modulate at concentrations up to
10-fold higher than H22. The optimal concentration for modulation was
slightly higher (half-log) for wH22xeGFP than for H22. This
difference may be attributable to a lower affinity of wH22xeGFP for
FcRI, which has been observed in other studies (data not
shown). Interestingly, high concentrations of H22 frequently led to
"negative modulation." We believe this increased level of surface
FcRI is due to the functionally monovalent binding of H22, which
mimics the natural ligand in its ability to increase receptor
expression. This phenomenon may be due to the ability of ligand to
prevent receptor shedding or enhance the binding of detection Abs. For
instance, in the absence of ligand, some FcRI molecules may localize
in packed microdomains on the cell surface, which are less susceptible
to the detection Ab, whereas IgG could cause dispersion of the
receptors, making them more readily detectable.
An important finding in the current studies is that, despite apparently
lower affinity, wH22xeGFP generally led to faster and greater levels of
FcRI modulation than H22, suggesting that the addition of an Ag to
H22 can alter the binding and cross-linking characteristics of the mAb.
It is therefore important to assess each FcRI-targeted Ag for its
ability to modulate and internalize, because the binding and
cross-linking properties of targeted Ags may play a role in the
strength of an immune response. For example, Ags which cause more
complete modulation would be expected to result in a larger antigenic
load and increased Ag presentation per mmol of Ag. Our studies suggest
that wH22xeGFP-mediated FcRI modulation may be similar to that of
197, a mouse IgG2a mAb against FcRI shown to be more effective in
modulating FcRI than H22 in the absence of ligand (27).
One possible explanation for the observed differences between H22- and
wH22xeGFP-mediated modulation and internalization is differences in the
accessibility of the H22-binding site of FcRI to each Ab, once the
Fc portion is bound. For example, at a given concentration, each H22
molecule may favor binding to two FcRI molecules: the Fc end and one
Fab end to one receptor and the other Fab end to a second receptor.
However, the addition of eGFP moieties to the C-terminal ends of the H
chains may contribute enough steric hindrance that binding of wH22xeGFP
to the ligand-binding domain and the H22-binding domain of the same FcR
is limited or impossible. Therefore, wH22xeGFP may favor the binding of
three FcR: one by the Fc end and two by the Fab ends. The latter
conformation would favor more extensive and rapid cross-linking of
FcRI. Indeed, using R-PE-labeled polyclonal anti-human IgG as
detection, we have observed higher binding with H22 than with equimolar
concentrations of wH22xeGFP (data not shown), supporting such a
model.
Two mechanisms of FcRI internalization have been proposed by
Harrison et al. (25): 1) binding of monovalent ligand
causes rapid internalization and recycling of FcRI back to the cell
surface via an endosomal pathway; and 2) binding of an FcRI
cross-linker in the presence of ligand causes rapid internalization and
retention of the receptor inside the cell, most likely diverting it to
a lysosomal pathway. Our receptor modulation and internalization data
with monovalent sFv22xeGFP and with multivalent H22 and wH22xeGFP are
consistent with these respective models. Furthermore, we have shown
that exogenous IgG was not required for the internalization and
modulation that results from cross-linking FcRI using H22 and
wH22xeGFP, most likely because these constructs themselves will bind
inside as well as outside the ligand-binding domain of FcRI.
In summary, we have demonstrated that an anti-FcRI Ab (H22) and
an FcRI-targeted fusion protein (wH22xeGFP) capable of cross-linking
FcRI internalized rapidly, along with the receptor, leading to
modulation of surface FcRI. The specific nature of internalization
and modulation was dependent on the level of FcRI expression on the
cell and the concentration of modulator, and was altered by the
presence of ligand. A monovalent version of the targeted fusion
protein, sFv22xeGFP, was unable to modulate surface FcRI but could
internalize if ligand was present. These data, taken together with
those of others, suggest that FcRI-targeted fusion proteins that
cross-link FcRI are internalized and enter an intracellular pathway
different from those which do not cross-link FcRI. This may reflect
the in vivo Ag sampling that occurs by FcR cross-linking and
FcR-mediated endocytosis, respectively. Our studies suggest that under
physiologic conditions, FcRI has the capacity to mediate the uptake
of a large antigenic load, carried into the cell by 40–80% of surface
FcRI. Therefore, a diverse array of Ags can be sampled, processed,
and presented, potentially leading to an effective immune stimulation.
The therapeutic implications of this work are twofold: 1) the
optimization of FcR modulation may lead to better therapeutic
strategies for FcR-mediated autoimmune diseases such as immune
thrombocytopenia purpura; and 2) the successful delivery of Ag into the
appropriate processing pathways of APCs via FcRI targeting may lead
to more effective Ag presentation and subsequent immune response to
foreign invaders or cancer. We have recently demonstrated the
intracellular colocalization of FcRI-targeted Ag with MHC class I
(31) and believe that the pathway for internalization of
FcRI-targeted eGFP will be the same as for those FcRI-targeted
fusion proteins being developed as vaccines for cancer, AIDS, and other
diseases (14, 16, 17, 19, 20, 21). Moreover, it has been shown
that FcRI-targeted Ag can be presented in the context of class I
(15) as well as class II (13, 14). These
fusion proteins may serve as useful tools in helping to elucidate how
FcRI-targeted Ags get processed inside the cell such that
presentation occurs via both class I and class II MHC.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Michael W. Fanger, Department of Microbiology, Dartmouth Medical School, 1 Medical Center Drive, Lebanon, NH 03756. E-mail address: Michael.W.Fanger{at}Dartmouth.edu
3 Abbreviations used in this paper: eGFP, enhanced green fluorescent protein; MFF, methanol-free formaldehyde; NHS, normal human serum; MFI, mean fluorescence intensity.
Received for publication March 26, 2001. Accepted for publication September 24, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
receptors in cancer and infectious disease. J. Leukocyte Biol. 55:816.[Abstract]
-RI are able to trigger receptor function. J. Immunol. 143:1650.[Abstract]
RI in IFN--treated U937 cells. Mol. Immunol. 27:263.[Medline]
RI: oxidase deactivation precedes internalization of cross-linked receptors. J. Immunol. 143:2640.[Abstract]
receptor-mediated phagocytosis by human blood dendritic cells. J. Immunol. 157:541.[Abstract]
RI (CD64) expressed on human blood dendritic cells. J. Immunol. 158:3090.[Abstract]
receptor (monocyte/macrophage)-specific immunogens. J. Immunol. 149:3477.[Abstract]
RI-targeted fusion proteins result in efficient presentation by human monocytes of antigenic and antagonist T cell epitopes. J. Clin. Invest. 98:2001.[Medline]
RI on myeloid cells is presented in association with MHC class I. J. Immunol. Methods 248:183.[Medline]
RI/CD64 triggers enhanced antibody responses in transgenic mice. J. Clin. Invest. 97:331.[Medline]
R-directed immunotherapies. J. G. J. van de Winkel, and P. M. Horgarth, eds. The Immunoglobulin Receptors and Their Physiological and Pathological Roles in Immunity 291. Kluwer Academic Publishers, UK.
RI) infusion in a patient with immune thrombocytopenia purpura (ITP) results in down-modulation of FcRI on circulating monocytes. Br. J. Haematol. 92:718.[Medline]
RI), blocks phagocytosis, and modulates receptor expression. J. Leukocyte Biol. 62:469.[Abstract]
RI-mediated endocytosis. J. Biol. Chem. 269:24396.-Ig receptor type I (FcRI) monoclonal antibody. J. Immunol. 155:4996.[Abstract]
receptor type I-expressing effector cells. Cancer Res. 57:4008. receptor using a monomeric antibody with unique properties. Blood 77:607.RI-targeted antigen with class I MHC: implications for antigen processing. J. Immunol. 166:2469.RI receptor, and evidence for restricted binding of ruminant IgG to U937 cells. Mol. Immunol. 29:1407.[Medline]
-interferon and expression of MHC genes regulate peptide hydrolysis by proteasomes. Nature 365:264.[Medline]
. Annu. Rev. Immunol. 15:749.[Medline]
on antigen processing in human monocytes. Eur. J. Immunol. 24:3124.[Medline]
RI blockade and modulation for immunotherapy. Cancer Immunol. Immunother. 45:137.[Medline]
receptor (FcRI) signaling. J. Biol. Chem. 270:8164.-chain tyrosine phosphorylation subverted by a protein-tyrosine phosphatase inhibitor. J. Biol. Chem. 271:11099.RI). Cell 67:275.[Medline]
receptor I- and Fc receptor I-mediated tumor cytotoxicity by monocyte-derived macrophages. J. Immunol. 164:5746.This article has been cited by other articles:
|
|
 |
|
  P. V. Beum, D. A. Mack, A. W. Pawluczkowycz, M. A. Lindorfer, and R. P. Taylor Binding of Rituximab, Trastuzumab, Cetuximab, or mAb T101 to Cancer Cells Promotes Trogocytosis Mediated by THP-1 Cells and Monocytes J. Immunol., December 1, 2008; 181(11): 8120 - 8132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. V. Beum, M. A. Lindorfer, and R. P. Taylor Within Peripheral Blood Mononuclear Cells, Antibody-Dependent Cellular Cytotoxicity of Rituximab-Opsonized Daudi cells Is Promoted by NK Cells and Inhibited by Monocytes due to Shaving J. Immunol., August 15, 2008; 181(4): 2916 - 2924. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. Rawool, C. Bitsaktsis, Y. Li, D. R. Gosselin, Y. Lin, N. V. Kurkure, D. W. Metzger, and E. J. Gosselin Utilization of Fc Receptors as a Mucosal Vaccine Strategy against an Intracellular Bacterium, Francisella tularensis J. Immunol., April 15, 2008; 180(8): 5548 - 5557. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Bevaart, H. H. Van Ojik, A. W. Sun, T. H. Sulahian, J. H. W. Leusen, G. J. Weiner, J. G. J. van de Winkel, and M. J. Van Vugt CpG oligodeoxynucleotides enhance Fc{gamma}RI-mediated cross presentation by dendritic cells Int. Immunol., August 1, 2004; 16(8): 1091 - 1098. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) modulated 38% of surface Fc
) nor 15 nM sFv22xeGFP (•) led to receptor
modulation. However, treatment with 5 nM of mAb 197 (
), wH22xeGFP
(
) led to 72, 61, and 29% modulation, respectively,
by 4 h. Mean values of duplicate samples ± SD are
plotted.
) and
wH22xeGFP (
), while modulation was low or absent in the absence of
IgG (filled symbols). B, The presence of fresh human
serum (
