* Xenova Group plc, Cambridge, United Kingdom; and
Division of Infection, Inflammation and Repair, University of Southampton, United Kingdom
Correspondence: Herbert Schwarz, Xenova Group, 310 Cambridge Science Park, Cambridge CB4 OWG, UK. E-mail: herbert_schwarz{at}xenova.co.uk
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: ADCC • complement lysis • colitis
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A crucial component of these fusion proteins is the class of the IgG domain. The Fc domain of IgG1, the most commonly used Fc domain in fusion proteins, exerts immune effector function, such as complement lysis and antibody-dependent cell cytotoxicity (ADCC). Although these activities can be important for the function of some of the fusion proteins, they may also lead to undesired side-effects in cases where the ligand is expressed on other cells and tissues.
OX40 is a costimulatory molecule of the TNFR family and is expressed selectively on activated T lymphocytes [4 ]. T cells receive activating and survival signals through OX40 [5 ]. Expression of OX40 ligand is activation-dependent and is found on antigen-presenting cells (APC). Further, OX40 ligand is expressed by activated vascular endothelial cells at sites of inflammation and plays a role in extravasation of OX40-positive T cells [6 7 8 9 ]. Therefore, soluble OX40 decoys can inhibit immune responses by blocking T cell costimulation and by preventing T cell extravasation into inflamed tissue [4 , 10 ].
An OX40-IgG1 fusion protein can inhibit murine splenocyte proliferation and cytokine production in vitro [11 ]. Further, it can successfully ameliorate trinitrobenzene sulfonic acid (TNBS)-induced colitis and spontaneous colitis in interleukin (IL)-2-deficient mice, as well as experimental autoimmune encephalomyelitis (EAE) [11 , 12 ].
Here, we describe the construction of further OX40-IgG fusion proteins with different levels of immune-effector functions. The role of these effector functions and the role of neutralizing OX40 ligand for OX40-IgG-mediated immune inhibition are investigated by in vitro and in vivo experiments.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cells and cell culture
OX40-expressing 556 and OX40L-expressing 5L cells were a gift
from Dr. W. Godfrey (University of Minnesota, Minneapolis). They are
derived from the murine myeloma line SP-2, which was transfected and
selected to stably express human OX40 or OX40L, respectively.
U937 cells were obtained from ECACC (Salisbury, UK) and were grown in RPMI 1640 with glutamine, penicillin (50 iu/ml)/streptomycin (50 µg/ml), and 10% Bioclear fetal bovine serum (FBS; Invitrogen Life Technologies, Inchinan, UK).
Human peripheral blood mononuclear cells (PBMC) were isolated from whole blood of healthy volunteers. Blood (50 ml) was added to 8 ml citrate-phosphate-dextrose solution, made up to 100 ml with phosphate-buffered saline (PBS) and layered onto three tubes with 17 ml lymphocyte separation medium (ICN, Basingstoke, UK). After centrifugation at 900 g for 30 min, the cells from the boundary of each tube were collected and resuspended in 50 ml PBS and spun at 900 g for 10 min. All cells from the three tubes were pooled, washed in 10 ml RPMI 10% fetal calf serum (FCS), and finally resuspended in 25 ml RPMI 10% FCS.
Cloning of the IgG4 domain
The Fc domain of human IgG4 was cloned by reverse
transcriptase-polymerase chain reaction (RT-PCR) using cDNA of PBMC as
template. The upstream primer (SWK92: 5' GATCGGATCCCGAGTCCAAATATGGTCCC)
was designed to anneal to the 5' of the IgG4 hinge region and to
incorporate a 5' BamHI site to allow fusion to the cDNA
encoding the extracellular domain of OX40. The downstream primer
(SWK93: 3' CTAGTCTAGATCATTATTTACCCAGAGACAGGGAG) was designed to
anneal to the 3' of the IgG4 CH3 domain and incorporates two 3' stop
codons followed by a XbaI restriction enzyme site. The
OX40-IgG4 cDNA was inserted into plasmid pOX8 and was transfected into
CHO cells. The secreted OX40-IgG4 fusion protein was purified via
protein A sepharose.
Site-directed mutagenesis
The mutation of Leu235 to Glu in hIgG4 was achieved
through site-directed mutagenesis using the Transformer Site Directed
Mutagenesis kit (Clontech Laboratories, Basingstoke, UK). The
oligonucleotide primer SWK115 (5': CCAGCACCTGAGTTCGAAGGGGGACCATCAGTC)
was designed to replace Leu235 by a Glu. This primer was
used alongside the oligonucleotide JCF35 (5':
ATGACTTGGTTGAATACTCACCAGTCA), which served as a selection primer by
destroying an essential ScaI site within the plasmid vector.
Oligonucleotides were first phosphorylated, and using wild-type hIgG4
cDNA as template, mutagenesis was performed per the manufacturer’s
protocol. The introduction of the mutation was verified by sequencing.
Expression of OX40-IgG fusion protein
The OX40-IgG1 fusion protein was expressed and purified as
described previously with a few modifications [11
,
13
]. Briefly, the cDNA encoding the extracellular domain
of human OX40 from amino acid 1 to 208 was fused to the 5' end of the
cDNA encoding the constant domain of human IgG1 (hinge region, CH2 and
CH3) and inserted into the mammalian expression vector pEE14 (Lonza
Biologics). The OX40-IgG1fusion protein was expressed in CHO cells and
purified by protein A sepharose. The OX40-IgG4 and OX40-IgG4mut
proteins were expressed and purified accordingly.
Fc receptor binding studies
U937 cells were stimulated overnight with 625 pg/ml human
interferon-
(IFN-; NBS Biologicals 22217, Huntingdon) for maximum
Fc receptor expression. Cells were resuspended in fluorescein-activated
cell sorter (FACS) buffer [PBS, 2% heat-inactivated Bioclear FBS
(Invitrogen Life Technologies), 0.02% sodium azide] and were
incubated with OX40-hIgG protein in serial 1:1 dilutions ranging from
100 µg/ml to approximately 0.1 µg/ml for 1 h on ice in a
volume of 50 µl. Cells were then washed twice in 2 ml FACS buffer.
Binding of the OX40-IgG fusion proteins was detected by an incubation
(40 min on ice) with 20 µl 40 µg/ml biotinylated anti-OX40 antibody
(clone L106, DAKO, Ely, UK) or an isotype-control antibody
(biotinylated mIgG1, BD Pharmingen, Oxford). Cells were washed again
and stained with 20 µl 5 µg/ml phycoerythrin (PE)-streptavidin (BD
Pharmingen) for 40 min on ice. The cells were washed and resuspended at
1 x 106/ml and analyzed using a FACScan and LysisII
software (Becton Dickinson, Mountain View, CA).
Complement-mediated cell lysis
Cells (106 5L and 565) were labeled with 4.6 MBq
51chromium sulfate (Amersham, Little Chalfont, UK) for
60–90 min and were then incubated with 2 µg/ml fusion proteins (200
µl per 5x105 cells) in cold PBS for 30 min at 4°C.
After washing once in GVB++ buffer (150 mM
NaCl, 5 mM sodium 5,5' diethyl barbiturate, 6.5% gelatin, 40 mM
MgCl2, 60 µM CaCl2, pH 7.35), the
cells were resuspended at 2 x 105/ml. One-hundred
microliters was added to each well already containing 100 µl guinea
pig complement serum (Sigma Chemical Co., Poole, UK) at a dilution of
1:100. After 2 h at 37°C, 25 µl supernatants from each well
were added to 150 µl Opti-Phase supermix (Wallac, Milton Keynes, UK)
in a flexible counter plate for evaluation in a MicroBetaPlus
ß-counter. Spontaneous lysis was determined from 5L and 565 cells,
which were incubated without the fusion proteins or PBMC, and maximal
lysis was determined by the addition of 5% Triton X-100.
Antibody-dependent, cell-mediated cytotoxicity
Cells (106 5L or 565) were labeled with 4.6 MBq
51chromium sulfate (Amersham) for 60–90 min. Cells were
washed three times with RPMI 10% FCS. Labeled 5L and 565 cells (100
µl) were plated in 96-well plates at a density of 105
cells/ml in RPMI 10% FCS with varying concentrations of the fusion
proteins (ranging from 20 µg/ml to 652 ng/ml) and incubated at 37°C
for 30 min. PBMC were then added to a final concentration of
106 cells per ml. After 4 h at 37°C, 25 µl
supernatants from each well were added to 150 µl Opti-Phase supermix
(Wallac) in a flexible counter plate for evaluation in a MicroBetaPlus
ß-counter. Spontaneous lysis was determined from 5L and 565 cells
that were incubated without the fusion proteins or PBMC, and maximal
lysis was determined by the addition of 5% Triton X-100.
Cell proliferation
Proliferation of cells was determined in 96-well plates. Cells
were pulsed with 0.5 µCi 3H-thymidine per well for
16 h and were harvested and evaluated on a microplate
scintillation counter. Measurements were performed in triplicates, and
results are represented as means ± standard deviations.
Measurement of cytokine release
IL-5 and IFN- concentrations were determined by enzyme-linked
immunosorbent assay (ELISA) using the protocol provided by the
manufacturer (BioSource International, Nivelles, Belgium).
Quantification of cell death
The numbers of dead cells were determined by trypan blue
staining. Live and dead cells in four fields of a hemacytometer were
counted, and the percentages of dead cells were calculated. Means and
standard deviations of four such countings are depicted.
Determination of OX40 binding to OX40L
OX40L-IgG1 at 10 µg/ml was incubated with OX40-IgG1,
OX40-IgG4m, or hIgG4 at varying concentrations for 30 min at 37°C.
Cells (565) expressing OX40 were added to this mix for 30 min at
37°C. After washing, the cells were labeled with biotinylated,
anti-human OX40L antibody (clone H2.33) for 30 min at 4°C followed by
PE-conjugated streptavidin. The cells were run through a Becton
Dickinson FACScan to determine fluorescence of the cells.
Animals
Female BALB/c mice (8- to 10-weeks old) were obtained from A.
Tuck & Sons (Southend-on-Sea, UK). All mice were housed under standard
conditions with free access to food and water.
Induction of colitis
BALB/c mice were weighed before procedure. TNBS (Fluka,
Gillingham, UK) was prepared in a 50% ethanol solution diluted to give
a final concentration of 2 mg TNBS in 75 µl total volume. Mice were
lightly anaesthetized using 200 µl of a 1/10 aqueous dilution of
Hypnorm (Janssen-Cilag, High Wycombe, UK). Colitis was induced by
intrarectal administration of 75 µl TNBS solution using a plastic
catheter. Control mice received 50% aqueous ethanol only. The general
condition and body weight of mice were checked daily.
Treatment with fusion proteins
TNBS colitic mice and ethanol-treated controls were injected
intraperitoneally with hOX40-IgG (100 µg) on days 4–6 after
induction of colitis.
RNA extraction and quantitative RT-PCR
Constructs encoding standard RNAs (pMCQ1, pMCQ2, pMCQ3,
and pMCQ4), kindly provided by Dr. M. F. Kagnoff (Department of
Medicine, University of California, San Diego) [14
],
were used for quantitative, competitive RT-PCR. To generate standard
RNA, plasmids were linearized with SalI (pMCQ1) or
NotI (pMCQ2, 3, 4) and were transcribed in vitro using T7
RNA polymerase under conditions recommended by the supplier (Promega,
Southampton, UK).
Gut tissue was snap frozen in liquid nitrogen and stored at -70°C. Cellular RNA was isolated by homogenizing tissue in TRIzol (Life Technologies, Paisley, UK) and incubating at room temperature for 5 min. RNA was extracted using chloroform (Sigma Chemical Co.), followed by centrifugation for 15 min at 12,000 g at 4°C. The aqueous phase was precipitated with an equal volume of isopropanol (Sigma Chemical Co.), followed by centrifugation for 15 min at 12,000 g at 4°C. The pellet was washed with 70% ethanol and resuspended in 50 µl water. Total RNA was determined by spectrometric analysis.
Serial, tenfold dilutions of standard RNA (1 pg–1 fg) were coreverse
transcribed with total cellular RNA (2 µg) at 42°C for 50 min in a
20 µl reaction volume containing 50 mM Tris, pH 8.3, 75 mM KCl, 3 mM
MgCl2, 3 mM dithiothreitol, 10 mM dNTP mix, and 0.5 µg
oligo(dT) (Pharmacia Biotech, Herts, UK), using 100 U RT (Superscript
II RNase H-, Life Technologies). The reaction was
terminated by heat inactivation at 70°C for 10 min. PCR amplification
was conducted routinely in 50 µl reaction volumes {10 mM Tris, pH
9, 50 mM KCl, 1.5 mM MgCl2, 200 µM dNTPs, 10 pmol 5' and
3' primers, as described elsewhere (see ref. [14
]), and
1 U Taq polymerase (Pharmacia Biotech)}. Forty amplification cycles
of 45 s denaturation at 94°C, 45 s annealing at 58°C, and
75 s extension at 72°C were used. Primers used for TNF-
were:
sense 5'-ATGAGCACAGAAAGCATGATC; antisense 5'-TACAGGCTTGTCACTCGAATT.
After amplification, PCR products were analyzed on 1% agarose gels, and bands were visualized by ethidium bromide staining. Band intensities were quantified by densitometry (Seescan, Cambridge, UK). The sensitivity of this technique enables the detection of >103 mRNA transcripts per µg total RNA.
Immunhistochemistry
Three-step avidin-peroxidase staining was performed on 5 µm
frozen sections as described previously using the monoclonal anti-CD4
antibody YTS 191 (American Type Culture Collection, Manassas, VA)
[15
]. Biotin-conjugated rabbit anti-rat IgG (DAKO) and
goat anti-hamster IgG (Vector Laboratories, Peterborough, UK) were used
at 1:50 dilutions in Tris-buffered saline (TBS), pH 7.6, containing 4%
(v/v) normal mouse serum (Harlan Seralab, Oxon, UK). Avidin peroxidase
(Sigma Chemical Co.) was used at dilutions of 1:200 in TBS. Peroxidase
activity was detected with 3,3'-diaminobenzidine-tetra-hydrochloride
(Sigma Chemical Co.) in 0.5 mg/ml Tris-HCl, pH 7.6, containing 0.01%
H2O2. The density of the cells in the lamina
propria was determined by image analysis as described previously
[15
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Binding affinities of OX40-IgG fusion proteins
The affinities of the IgG domains of the three OX40-IgG fusion
proteins to the Fc receptor for IgG (FcR) were tested by incubation
with the monocytic cell line U937. The cells were activated overnight
with 625 ng/ml IFN- to increase expression of FcR. OX40 ligand is
not expressed as a result of this stimulation (not shown). Binding of
the OX40 fusion proteins to Fc receptor was analyzed by flow cytometry
after staining with an OX40-specific antibody. The strongest binding
was found for OX40-IgG1 (Fig. 1 A
). OX40-IgG4 bound about half as strongly to the U937 cells, and
no binding was obtained with OX40-IgG4mut (Fig. 1A)
. Binding of
OX40-IgG fusion protein was concentration-dependent and reached
saturation at 3 µg/ml hOX40-hIgG1 (Fig. 1B)
.
 View larger version (27K): [in a new window] |
Figure 1. Analysis of Fc receptor binding. (A) U937 cells were activated
overnight with 625 pg/ml IFN-. Cells (106 U937) per
condition were incubated with serial 1:1 dilutions of OX40-IgG fusion
proteins. Binding of the fusion proteins to Fc receptor was analyzed by
flow cytometry after staining with an OX40-specific antibody. This
experiment was repeated twice with identical results. (B) Flow
cytometry prolifles of hOX40-hIgG1 of (A). The highest concentration
used was 100 µg/ml, which was titrated down in 10 1:1-dilution steps
to 95 ng/ml. The arrows indicate the last FACS prolife for each
dilution step. (top left panel) Controls: 2nd ab, secondary antibodies,
which is biotinylated anti-OX40 (murine IgG1); isotype control,
biotinylated murine IgG1.
|
 View larger version (27K): [in a new window] |
Figure 2. Immune effector functions of OX40-IgG fusion proteins. (A) Complement
lysis: OX40L-expressing 5L cells (upper panel) and OX40-expressing 565
cells (lower panel) were labeled with 51chromium sulfate.
Then they were incubated with indicated concentrations of fusion
proteins in cold PBS for 30 min at 4°C. After washing the cells, 100
µl guinea pig complement serum was added at a dilution of 1:200.
Released 51Cr was used to calculate the degree of cell
lysis. Lysis at each point was determined in triplicates. Depicted are
means ± standard deviations. (B) ADCC: OX40L-expressing 5L cells
(upper panel) and OX40-expressing 565 cells (lower panel) were labeled
with 51chromium sulfate and incubated with indicated
concentrations of fusion proteins in the presence of PBMC. Released
51Cr was used to calculate the degree of cell lysis. Lysis
at each point was determined in triplicates. Depicted are means ±
standard deviations. (A and B) Experiments were performed at least
three times with comparable results.
|
These experiments convincingly demonstrate that the replacement of the Fc domain of IgG1 with the one of IgG4 and the exchange of Leu235 by Glu completely eliminated the complement lysis activity and ADCC of OX40-IgG1.
Inhibition of immune functions by OX40-IgG fusion proteins in vitro
The comparison of OX40-IgG1 and OX40-IgG4mut in functional
experiments should provide insight into the role of cell death in
hOX40-hIgG-mediated immunomodulation. OX40-IgG1 reduced proliferation
of anti-CD3-activated PBMC to about a quarter compared with the isotype
control antibody (Fig. 3 A
, upper panel). OX40-IgG4mut only reduced proliferation slightly
at its highest concentration of 50 µg/ml and had no effect at 12.5
and 25 µg/ml (Fig. 3A
, upper panel).
 View larger version (32K): [in a new window] |
Figure 3. OX40-IgG fusion proteins inhibit in vitro-immune functions. (A) PBMC
(2x105) were activated with 3 ng/ml anti-CD3 (OKT3).
OX40-IgG fusion proteins and isotype control antibodies were added at
indicated concentrations. Proliferation was determined at day four by
3H-thymidine incorporation (upper panel), and the numbers
of live and dead cells were evaluated by trypan blue staining at day
four (lower panel) and are depicted as percentage of dead cells. (B)
PBMC (5x105) in 0.5 ml medium were activated with IL-2 (10
U/ml) and phytohemagglutinin (PHA; 50 ng/ml). OX40-IgG fusion protein
and isotype-control antibodies were added at indicated concentrations.
Supernatants were harvested after three days, and concentrations of
IL-5 were determined by ELISA. (C) PBMC (5x105) in 0.5 ml
were activated with IL-2 (10 U/ml) and indicated concentrations of PHA.
OX40 fusion protein and isotype control antibodies were added at 25
µg/ml. Supernatants were harvested after three days, and
concentrations of IL-5 (upper panel) and IFN- (lower panel) were
determined by ELISA. (A and B) Experiments were performed at least
three times with comparable results. (D) Indicated concentrations of
OX40-IgG fusion proteins or IgG4 control antibody were incubated with
10 µg/ml hOX40L-hIgG1 for 30 min at 37°C. Afterwards, the solutions
were added to 5 x 105 OX40-expressing 565 cells.
OX40L-IgG1 bound to 565 cells was detected by staining with
biotinylated anti-OX40L antibody followed by streptavidin-PE and flow
cytometry.
|
The OX40-IgG1 and OX40-IgG4mut proteins were also compared with their
effects on mitogen-induced cytokine release by human PBMC. Both
proteins inhibited the release of IL-5 in mitogen-activated PBMC
dose-dependently, with 5–10 µg/ml fusion resulting in a complete
inhibition (Fig. 3B)
. This result was not surprising in the case of
OX40-IgG1, which can induce cell death. However, inhibition of IL-5 by
OX40-IgG4mut implied a more genuine regulation of cytokine secretion
distinct from merely knocking-out cells. This could be confirmed by
demonstrating that OX40-IgG4mut enhanced levels of IFN- (Fig. 3C)
.
To exclude that the observed differences in activity were a result of different affinities to OX40 ligand, we tested the respective abilities of OX40-IgG1 and OX40-IgG4m to neutralize OX40 ligand in a flow cytometry-based competition assay. OX40-IgG and OX40-IgG4m had the same potency in neutralizing the OX40 ligand binding to the 565 cells (Fig. 3D) . This demonstrated that the observed differences between the two proteins are a result of their different IgG domains.
Inhibition of colitis by OX40-IgG fusion proteins
The potencies of OX40-IgG1 and OX40-IgG4mut fusion proteins to
inhibit immune responses in vivo were evaluated in murine TNBS-induced
colitis. In this widely used model of Crohn’s disease, much of the gut
injury is because of local, cell-mediated immune reactions. The mice
were treated on days 4–6 with 100 µg fusion protein or an
isotype-control protein, respectively, and were killed on day 7. Both
OX40-IgG fusion proteins reduced the number of infiltrating
CD4-positive T cells into the lamina propria (Fig. 4A
4B
4C
4D
4E
). TNF- mRNA expression was also reduced by the two
OX40-IgG fusion proteins by the same degree to about a quarter compared
with the isotype control (Fig. 4F)
.
 View larger version (45K): [in a new window] |
Figure 4. Treatment with OX40-IgG fusion proteins reduces inflammation. Colitis
was induced in BALB/c mice by TNBS, and mice were treated with 100 µg
hIgG, OX40-IgG1, or OX40-IgG4mut on days 4–6. Mice were killed at day
7, and gut tissue was stained for CD4-positive T cell infiltration
(arrows) into the lamina propria, (A) ethanol control, (B) TNBS mice
treated with hIgG, (C) TNBS mice treated with hOX40-hIgG1, and (D) TNBS
mice treated with hOX40-hIgG4mut. Original magnification, x200. Serial
section stained in parallel in the absence of primary antibody showed
no staining. (E) Quantitative evaluation of infiltrated CD4-positive T
cells: mean ± standard error of mean. (F) TNF- mRNA transcript
levels in gut tissue of mice (A–D) were determined by quantitative
RT-PCR. Each group represents three to six mice, and the experiment was
repeated twice with identical results.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Three separate mechanisms that are not mutually exclusive could account for the observed reduction in immune responses by OX40-IgG1. It could bind and neutralize the activity of OX40 ligand on the APC; it could kill APC by ADCC and/or complement lysis; and it could bind and neutralize the activity of OX40 ligand on inflamed endothelial cells. The first two mechanisms would block T cell costimulation, and the third mechanism would block T cell extravasation. Which of the three mechanisms was operational and to which extent could not be determined from the previous experiments.
As a result of the intact IgG1 domain OX40-IgG1, this fusion protein was expected to have the identical immune effector functions as a human antibody of the IgG1 isotype. Indeed, induction of ADCC and complement lysis by OX40-IgG1 could be demonstrated in OX40 ligand-expressing cells (Fig. 2) .
To assess the contribution of cell killing to the potency of OX40-IgG1, an isoform was constructed that lacked immune effector functions. The ability of OX40-IgG1 to induce complement lysis was eliminated by replacing the Fc domain of human IgG1 with that of human IgG4, which does not bind C1q, the first protein in the complement cascade. However, as the IgG4 domain retains some ADCC activity, although less than the IgG1 domain, an OX40-IgG4 protein could still kill OX40 ligand-expressing cells. Therefore, ADCC activity of hOX40-hIgG4 needed to be eliminated as well. The strategy to achieve this was based on earlier data that follow.
ADCC is exerted by monocytes and natural killer cells when they
recognize opsonized cells via their Fc receptors. Human IgG4 does not
bind FcRII and FcRIII, but it does bind FcRI. A crucial
subdomain for this interaction has been mapped to amino acids 234–238
of IgG4 [16
]. Murine IgG2a is the homologue of human
IgG1, and both molecules have a high affinity for FcRI
(10-8). Their FcR binding domains also share
the following identical sequence: Leu234-Leu-Gly-Gly-Pro
(Fig. 5
). Leu235 was found to be essential for FcRI binding
and subsequent ADCC induction by murine IgG2a [17
].
Murine IgG2b is a variant of mIgG2a. Its FcRI binding site varies by
a single amino acid substitution (Leu235
Glu) from that
of murine IgG2a, resulting in a low affinity for FcRI
(10-6). Targeted mutation of
Glu235 to Leu in murine IgG2b not only changes its sequence
to that of murine IgG2a but also increases its affinity for FcRI to
that of murine IgG2a [17
, 18
].
 View larger version (19K): [in a new window] |
Figure 5. Sequence of FcRI-binding sites Depicted are the sequence motifs
(amino acids 234–238) of human and murine IgG, which are mainly
responsible for binding to FcRI. Also listed are the respective
affinities of the Igs to FcRI and their capacity to induce ADCC and
CDC. +++, ++, +, and -, High, medium, low, and no activity,
respectively; n.d., not determined; CDC, complement-dependent
cytotoxicity.
|
RI
(10-7) and the sequence
Phe234-Leu-Gly-Gly-Pro. It retains the Leu at position 235,
which is essential for FcRI binding as is outlined above. Therefore,
we concluded that mutation of Leu235 to a Glu in human IgG4
(OX40-IgG4mut) should completely abolish FcRI binding and ADCC. This
hypothesis could be confirmed nicely by FcR binding and ADCC assays
(Figs. 1
and 2B)
.
U937 cells not only express FcRI, but also FcRII and FcRIII.
It is therefore not possible to reliably compare binding affinities
between hOX40-hIgG1 and hOX40-hIgG4 using these cells, as IgG4 only
binds to FcRI, whereas IgG1 binds to FcRII and FcRIII as well.
However, the experiment in Figure 1
nicely demonstrates that
substitution of Leu235 by Glu in IgG4 eliminates the
FcRI binding of IgG4. Also, as IgG4 induces ADCC via FcRI, these
data correlate well with the capacity of hOX40-hIgG4 and hOX40-hIgG4mut
to induce ADCC.
OX40-IgG1 was more potent than OX40-IgG4mut in inhibiting immune
reactions in vitro. OX40-IgG1 profoundly inhibited proliferation of
PBMC and the release of IL-5 and IFN-. OX40-IgG4mut had no or little
effect on proliferation of PBMC. It inhibited IL-5 release but
increased levels of IFN-. The stronger effects of OX40-IgG1 are
likely a result of the elimination of APC by ADCC, as evidenced by the
higher number of dead cells in the OX40-IgG1-treated cultures. As
OX40-IgG1 differs from OX40-IgG4mut only by its ability to induce cell
death, this activity has to be the basis of its higher potency.
OX40-IgG4mut, which neutralizes OX40L, does not seem to be a general inhibitor of immune reactions in vitro. It had no effect on proliferation and did not inhibit cytokine release in general. Rather, by inhibiting an OX40L-mediated TH2 shift, it changed the cytokine profile toward a TH1 response.
While OX40-IgG1 seems to be more potent in inhibiting immune reactions in vitro, OX40-IgG1 and OX40-IgG4mut fusion proteins reduced colitis in mice to the same extent. A likely explanation for this difference is the fact that in vitro, the OX40-IgG fusion proteins are restricted to interfere with APC-mediated costimulation of T cells. However, in vivo, OX40 fusion proteins can act additionally on OX40L on the inflamed vascular endothelium, killing the endothelial cells or neutralizing OX40L on their surface. Both mechanisms would decrease the extravasation signal for OX40-positive T cells. The equal potencies of OX40-IgG1 and OX40-IgG4mut indicate that blocking T cell extravasation is sufficient to quench autoimmune reactions and may be the major mechanism for the OX40-IgG activity in vivo.
As the possibility cannot be excluded that the OX40-IgG1 fusion protein may kill endothelial cells, potentially entailing side effects for an OX40-IgG1-treated autoimmune patient, the OX40-IgG4mut protein may be the safer alternative for human therapy.
These data also demonstrate that Ig effector functions are active in fusion proteins in the same way as in antibodies. This should allow the design of fusion proteins with desired effector functions. Fusion proteins, which target cancer cells or autoreactive immune cells, may gain in potency if they have the capability to induce cell death in their target tissues. This increase in potency has to be balanced with possible side-effects if the ligand of the fusion protein is expressed on tissues other than the target cells.
Received November 12, 2001; revised April 12, 2002; accepted April 15, 2002.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. E. Burrell, G. Lu, X. C. Li, and D. K. Bishop OX40 Costimulation Prevents Allograft Acceptance Induced by CD40-CD40L Blockade J. Immunol., January 1, 2009; 182(1): 379 - 390. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Roosnek, P. Schneider, and B. Huard Heparan sulfate proteoglycans, Fc receptors, and DC suppression Blood, August 1, 2008; 112(3): 915 - 916. [Full Text] [PDF]
|
|
|
|
 |
|
 A. D. Weinberg, D. E. Evans, C. Thalhofer, T. Shi, and R. A. Prell The generation of T cell memory: a review describing the molecular and cellular events following OX40 (CD134) engagement J. Leukoc. Biol., June 1, 2004; 75(6): 962 - 972. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. R. Humphreys, G. Walzl, L. Edwards, A. Rae, S. Hill, and T. Hussell A Critical Role for OX40 in T Cell-mediated Immunopathology during Lung Viral Infection J. Exp. Med., October 20, 2003; 198(8): 1237 - 1242. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Wilson, C. P. Rossi, S. Carboni, C. Fremaux, D. Perrin, C. Soto, M. Kosco-Vilbois, and A. Scheer A Homogeneous 384-Well High-Throughput Binding Assay for a TNF Receptor Using Alphascreen Technology J Biomol Screen, October 1, 2003; 8(5): 522 - 532. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
